SymbolicSummation for Combinatorialand...
Transcript of SymbolicSummation for Combinatorialand...
![Page 1: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/1.jpg)
Strobl, June 10, 2015
AofA’2015
Symbolic Summation for
Combinatorial and Related Problems
Carsten Schneider
SFB F050 Algorithmic and Enumerative CombinatoricsResearch Institute for Symbolic Computation
Johannes Kepler University Linz
RISC, J. Kepler University Linz
![Page 2: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/2.jpg)
Some of the available summation tools:Abramov, S.A.: On the summation of rational functions. Zh. vychisl. mat. Fiz. 11, 1071–1074 (1971)Abramov, S.A.: The rational component of the solution of a first-order linear recurrence relation with a rationalright-hand side. U.S.S.R. Comput. Maths. Math. Phys. 15, 216–221 (1975). Transl. from Zh. vychisl. mat. mat. fiz.15, pp. 1035–1039, 1975Abramov, S.A.: Rational solutions of linear differential and difference equations with polynomial coefficients. U.S.S.R.Comput. Math. Math. Phys. 29(6), 7–12 (1989)Abramov, S.A., Petkovsek, M.: D’Alembertian solutions of linear differential and difference equations. In: J. von zurGathen (ed.) Proc. ISSAC’94, pp. 169–174. ACM Press (1994)Abramov, S.A., Petkovsek, M.: Rational normal forms and minimal decompositions of hypergeometric terms. J. SymbolicComput. 33(5), 521–543 (2002)Apagodu, M., Zeilberger, D., 2006. Multi-variable Zeilberger and Almkvist–Zeilberger algorithms and the sharpening ofWilf–Zeilberger theory. Advances in Applied Math. 37, 139–152.Bauer, A., Petkovsek, M.: Multibasic and mixed hypergeometric Gosper-type algorithms. J. Symbolic Comput. 28(4–5),711–736 (1999)Bronstein, M.: On solutions of linear ordinary difference equations in their coefficient field. J. Symbolic Comput. 29(6),841–877 (2000)Chen, S., Jaroschek, M., Kauers, M., Singer, M.F.: Desingularization Explains Order-Degree Curves for Ore Operators. In:M. Kauers (ed.) Proc. of ISSAC’13, pp. 157–164 (2013)Chen, S., Kauers, M.: Order-Degree Curves for Hypergeometric Creative Telescoping. In: J. van der Hoeven, M. vanHoeij (eds.) Proceedings of ISSAC 2012, pp. 122–129 (2012)Chyzak, F.: An extension of Zeilberger’s fast algorithm to general holonomic functions. Discrete Math. 217, 115–134(2000)Fasenmyer, M. C., November 1945. Some generalized hypergeometric polynomials. Ph.D. thesis, University of Michigan.Gosper, R.W.: Decision procedures for indefinite hypergeometric summation. Proc. Nat. Acad. Sci. U.S.A. 75, 40–42(1978)Hendriks, P.A., Singer, M.F.: Solving difference equations in finite terms. J. Symbolic Comput. 27(3), 239–259 (1999)Karr, M.: Summation in finite terms. J. ACM 28, 305–350 (1981)Karr, M.: Theory of summation in finite terms. J. Symbolic Comput. 1, 303–315 (1985)M. Kauers and P. Paule. The concrete tetrahedron. Texts and Monographs in Symbolic Computation.SpringerWienNewYork, Vienna, 2011. Symbolic sums, recurrence equations, generating functions, asymptotic estimates.
...RISC, J. Kepler University Linz
![Page 3: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/3.jpg)
Some of the available summation tools:...
Koornwinder, T.H.: On Zeilberger’s algorithm and its q-analogue. J. Comp. Appl. Math. 48, 91–111 (1993)
Koutschan, C.: Creative telescoping for holonomic functions. In: C. Schneider, J. Blumlein (eds.) Computer Algebra inQuantum Field Theory: Integration, Summation and Special Functions, Texts and Monographs in Symbolic Computation,pp. 171–194. Springer (2013). ArXiv:1307.4554 [cs.SC]
Paule, P.: Greatest factorial factorization and symbolic summation. J. Symbolic Comput. 20(3), 235–268 (1995)
Paule, P.: Contiguous relations and creative telescoping. unpublished manuscript p. 33 pages (2001)
Paule, P., Riese, A.: A Mathematica q-analogue of Zeilberger’s algorithm based on an algebraically motivated aproach toq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and Related Topics, vol. 14,pp. 179–210. AMS (1997)
Paule, P., Schorn, M.: A Mathematica version of Zeilberger’s algorithm for proving binomial coefficient identities.J. Symbolic Comput. 20(5-6), 673–698 (1995)
Petkovsek, M.: Hypergeometric solutions of linear recurrences with polynomial coefficients. J. Symbolic Comput. 14(2-3),243–264 (1992)
Petkovsek, M., Wilf, H.S., Zeilberger, D.: A = B. A. K. Peters, Wellesley, MA (1996)
Petkovsek, M., Zakrajsek, H.: Solving linear recurrence equations with polynomial coefficients. In: C. Schneider,J. Blumlein (eds.) Computer Algebra in Quantum Field Theory: Integration, Summation and Special Functions, Textsand Monographs in Symbolic Computation, pp. 259–284. Springer (2013)
Pirastu, R., Strehl, V.: Rational summation and Gosper-Petkovsek representation. J. Symbolic Comput. 20(5-6),617–635 (1995)
Wegschaider, K., May 1997. Computer generated proofs of binomial multi-sum identities. Master’s thesis, RISC,Johannes Kepler University.
Wilf, H. S., Zeilberger, D., 1992. An algorithmic proof theory for hypergeometric (ordinary and “q”) multisum/integralidentities. Invent. Math. 108 (3), 575–633.
Zeilberger, D., 1990. A holonomic systems approach to special functions identities. J. Comput. Appl. Math. 32, 321–368.
Zeilberger, D.: The method of creative telescoping. J. Symbolic Comput. 11, 195–204 (1991)
RISC, J. Kepler University Linz
![Page 4: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/4.jpg)
Some of the available summation tools:...
Koornwinder, T.H.: On Zeilberger’s algorithm and its q-analogue. J. Comp. Appl. Math. 48, 91–111 (1993)
Koutschan, C.: Creative telescoping for holonomic functions. In: C. Schneider, J. Blumlein (eds.) Computer Algebra inQuantum Field Theory: Integration, Summation and Special Functions, Texts and Monographs in Symbolic Computation,pp. 171–194. Springer (2013). ArXiv:1307.4554 [cs.SC]
Paule, P.: Greatest factorial factorization and symbolic summation. J. Symbolic Comput. 20(3), 235–268 (1995)
Paule, P.: Contiguous relations and creative telescoping. unpublished manuscript p. 33 pages (2001)
Paule, P., Riese, A.: A Mathematica q-analogue of Zeilberger’s algorithm based on an algebraically motivated aproach toq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and Related Topics, vol. 14,pp. 179–210. AMS (1997)
Paule, P., Schorn, M.: A Mathematica version of Zeilberger’s algorithm for proving binomial coefficient identities.J. Symbolic Comput. 20(5-6), 673–698 (1995)
Petkovsek, M.: Hypergeometric solutions of linear recurrences with polynomial coefficients. J. Symbolic Comput. 14(2-3),243–264 (1992)
Petkovsek, M., Wilf, H.S., Zeilberger, D.: A = B. A. K. Peters, Wellesley, MA (1996)
Petkovsek, M., Zakrajsek, H.: Solving linear recurrence equations with polynomial coefficients. In: C. Schneider,J. Blumlein (eds.) Computer Algebra in Quantum Field Theory: Integration, Summation and Special Functions, Textsand Monographs in Symbolic Computation, pp. 259–284. Springer (2013)
Pirastu, R., Strehl, V.: Rational summation and Gosper-Petkovsek representation. J. Symbolic Comput. 20(5-6),617–635 (1995)
Wegschaider, K., May 1997. Computer generated proofs of binomial multi-sum identities. Master’s thesis, RISC,Johannes Kepler University.
Wilf, H. S., Zeilberger, D., 1992. An algorithmic proof theory for hypergeometric (ordinary and “q”) multisum/integralidentities. Invent. Math. 108 (3), 575–633.
Zeilberger, D., 1990. A holonomic systems approach to special functions identities. J. Comput. Appl. Math. 32, 321–368.
Zeilberger, D.: The method of creative telescoping. J. Symbolic Comput. 11, 195–204 (1991)
Here I will restrict to the setting of difference rings/fields.RISC, J. Kepler University Linz
![Page 5: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/5.jpg)
A bet (at my cost)
You’ve Got Mail (7/2004)
From: Doron Zeilberger
To: Robin Pemantle, Herbert Wilf
CC:Carsten Schneider
Robin and Herb,
I am willing to bet that Carsten Schneider’s SIGMA package
for handling sums with harmonic numbers (among others)
can do it in a jiffy. I am Cc-ing this to Carsten.
Carsten: please do it, and Cc- the answer to me.
-Doron
RISC, J. Kepler University Linz
![Page 6: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/6.jpg)
A bet (at my cost)
The problem
From: Robin Pemantle [University of Pennsylvania]
To: herb wilf; doron zeilberger
Herb, Doron,
I have a sum that, when I evaluate numerically, looks suspiciously
like it comes out to exactly 1.
Is there a way I can automatically decide this?
The sum may be written in many ways, but one is:
∞∑
n,k=1
Hk(Hn+1 − 1)
kn(n+ 1)(k + n); Hk :=
k∑
i=1
1
i
[Arose in the analysis of the simplex algorithm on the Klee-Minty cube(J. Balogh, R. Pemantle)]
RISC, J. Kepler University Linz
![Page 7: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/7.jpg)
A bet (at my cost)
S =
∞∑
n=1
Hn+1 − 1
n(n+ 1)
∞∑
k=1
Hk
k(k + n)
where Hk =k∑
i=1
1
i.
RISC, J. Kepler University Linz
![Page 8: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/8.jpg)
A bet (at my cost)
GIVENA’(n) :=
a∑
k=1
Hk
k(k + n)︸ ︷︷ ︸
=: f(n, k)
.
RISC, J. Kepler University Linz
![Page 9: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/9.jpg)
A bet (at my cost)
TelescopingGIVEN
A’(n) :=
a∑
k=1
Hk
k(k + n)︸ ︷︷ ︸
=: f(n, k)
.
FIND g(n, k):
g(n, k + 1)− g(n, k) = f(n, k)
for all n, k ≥ 1.
RISC, J. Kepler University Linz
![Page 10: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/10.jpg)
A bet (at my cost)
TelescopingGIVEN
A’(n) :=
a∑
k=1
Hk
k(k + n)︸ ︷︷ ︸
=: f(n, k)
.
FIND g(n, k):
g(n, k + 1)− g(n, k) = f(n, k)
for all n, k ≥ 1.
g(n, a+ 1)− g(n, 1) =
a∑
k=1
f(n, k)
RISC, J. Kepler University Linz
![Page 11: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/11.jpg)
A bet (at my cost)
TelescopingGIVEN
A’(n) :=
a∑
k=1
Hk
k(k + n)︸ ︷︷ ︸
=: f(n, k)
.
FIND g(n, k):
g(n, k + 1)− g(n, k) = f(n, k)
for all n, k ≥ 1.
no solution©◦ ◦⌢
RISC, J. Kepler University Linz
![Page 12: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/12.jpg)
A bet (at my cost)
Zeilberger’s creative telescoping paradigmGIVEN
A’(n) :=
a∑
k=1
Hk
k(k + n)︸ ︷︷ ︸
=: f(n, k)
.
FIND g(n, k) and c0(n), c1(n):
g(n, k + 1)− g(n, k) = c0(n)f(n, k) + c1(n) f(n+ 1, k)
for all n, k ≥ 1.
no solution©◦ ◦⌢
RISC, J. Kepler University Linz
![Page 13: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/13.jpg)
A bet (at my cost)
Zeilberger’s creative telescoping paradigmGIVEN
A’(n) :=
a∑
k=1
Hk
k(k + n)︸ ︷︷ ︸
=: f(n, k)
.
FIND g(n, k) and c0(n), c1(n), c2(n):
g(n, k + 1)− g(n, k) = c0(n)f(n, k) + c1(n) f(n+ 1, k) + c2(n) f(n+ 2, k)
for all n, k ≥ 1.
solution©◦ ◦⌣
RISC, J. Kepler University Linz
![Page 14: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/14.jpg)
A bet (at my cost)
Zeilberger’s creative telescoping paradigmGIVEN
A’(n) :=
a∑
k=1
Hk
k(k + n)︸ ︷︷ ︸
=: f(n, k)
.
FIND g(n, k) and c0(n), c1(n), c2(n):
g(n, k + 1)− g(n, k) = c0(n)f(n, k) + c1(n) f(n+ 1, k) + c2(n) f(n+ 2, k)
for all n, k ≥ 1.
Sigma computes: c0(n) = n2, c1(n) = −(n+ 1)(2n+ 1), c2(n) = (n+ 1)(n+ 2)
andg(n, k) := −
kHk + n+ k
(n+ k)(n+ k + 1),
g(n, k + 1) := −(1 + n)Hk + n+ k + 2
(n+ k + 1)(n+ k + 2).
RISC, J. Kepler University Linz
![Page 15: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/15.jpg)
A bet (at my cost)
Zeilberger’s creative telescoping paradigmGIVEN
A’(n) :=
a∑
k=1
Hk
k(k + n)︸ ︷︷ ︸
=: f(n, k)
.
FIND g(n, k) and c0(n), c1(n), c2(n):
g(n, k + 1)− g(n, k) = c0(n)f(n, k) + c1(n) f(n+ 1, k) + c2(n) f(n+ 2, k)
for all n, k ≥ 1.
Summing this equation over k from 1 to a gives:
g(n, a+1)− g(n, 1) =a∑
k=1
[
c0(n)f(n, k)+c1(n)f(n+1, k)+c2(n)f(n+2, k)]
RISC, J. Kepler University Linz
![Page 16: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/16.jpg)
A bet (at my cost)
Zeilberger’s creative telescoping paradigmGIVEN
A’(n) :=
a∑
k=1
Hk
k(k + n)︸ ︷︷ ︸
=: f(n, k)
.
FIND g(n, k) and c0(n), c1(n), c2(n):
g(n, k + 1)− g(n, k) = c0(n)f(n, k) + c1(n) f(n+ 1, k) + c2(n) f(n+ 2, k)
for all n, k ≥ 1.
Summing this equation over k from 1 to a gives:
g(n, a+1)− g(n, 1) =a∑
k=1
c0(n)f(n, k)+a∑
k=1
c1(n)f(n+1, k)+a∑
k=1
c2(n)f(n+2, k)
RISC, J. Kepler University Linz
![Page 17: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/17.jpg)
A bet (at my cost)
Zeilberger’s creative telescoping paradigmGIVEN
A’(n) :=
a∑
k=1
Hk
k(k + n)︸ ︷︷ ︸
=: f(n, k)
.
FIND g(n, k) and c0(n), c1(n), c2(n):
g(n, k + 1)− g(n, k) = c0(n)f(n, k) + c1(n) f(n+ 1, k) + c2(n) f(n+ 2, k)
for all n, k ≥ 1.
Summing this equation over k from 1 to a gives:
g(n, a+1)− g(n, 1) = c0(n)a∑
k=1
f(n, k)+c1(n)a∑
k=1
f(n+1, k)+c2(n)a∑
k=1
f(n+2, k)
RISC, J. Kepler University Linz
![Page 18: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/18.jpg)
A bet (at my cost)
Zeilberger’s creative telescoping paradigmGIVEN
A’(n) :=
a∑
k=1
Hk
k(k + n)︸ ︷︷ ︸
=: f(n, k)
.
FIND g(n, k) and c0(n), c1(n), c2(n):
g(n, k + 1)− g(n, k) = c0(n)f(n, k) + c1(n) f(n+ 1, k) + c2(n) f(n+ 2, k)
for all n, k ≥ 1.
Summing this equation over k from 1 to a gives:
g(n, a+1)− g(n, 1) = c0(n)A’(n) + c1(n)A’(n+ 1) + c2(n)A’(n+ 2)
RISC, J. Kepler University Linz
![Page 19: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/19.jpg)
A bet (at my cost)
Zeilberger’s creative telescoping paradigmGIVEN
A’(n) :=
a∑
k=1
Hk
k(k + n)︸ ︷︷ ︸
=: f(n, k)
.
FIND g(n, k) and c0(n), c1(n), c2(n):
g(n, k + 1)− g(n, k) = c0(n)f(n, k) + c1(n) f(n+ 1, k) + c2(n) f(n+ 2, k)
for all n, k ≥ 1.
Summing this equation over k from 1 to a gives:
g(n, a+1)− g(n, 1) = c0(n)A’(n) + c1(n)A’(n+ 1) + c2(n)A’(n+ 2)
|| ||a
(n+1)(a+n+1) n2A’(n)− (n+1)(2n+1)A’(n+ 1) + (n+1)(n+2)A’(n+2)
− (a+1)Ha
(a+n+1)(a+n+2)
RISC, J. Kepler University Linz
![Page 20: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/20.jpg)
A bet (at my cost)
Summation principles (in difference field/ring setting)
n2A(n)− (n+ 1)(2n + 1)A(n + 1) + (n+ 1)(n + 2)A(n + 2) =
1
n+ 1
Recurrence finder
A(n) =∞∑
k=1
Hk
k(k + n)
RISC, J. Kepler University Linz
![Page 21: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/21.jpg)
A bet (at my cost)
Summation principles (in difference field/ring setting)
n2A(n)− (n+ 1)(2n + 1)A(n + 1) + (n+ 1)(n + 2)A(n + 2) =
1
n+ 1
Recurrence solver
A(n) =∞∑
k=1
Hk
k(k + n)
{c1nHn − 1
n2+ c2
1
n
+nH2
n − 2Hn + nH(2)n
2n2
∣∣∣∣∣c1, c2 ∈ R}
where
H(2)n =
n∑
i=1
1
i2
RISC, J. Kepler University Linz
![Page 22: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/22.jpg)
A bet (at my cost)
Summation principles (in difference field/ring setting)
n2A(n)− (n+ 1)(2n + 1)A(n + 1) + (n+ 1)(n + 2)A(n + 2) =
1
n+ 1
Recurrence solver
A(n) =∞∑
k=1
Hk
k(k + n)∈
{c1nHn − 1
n2+ c2
1
n
+nH2
n − 2Hn + nH(2)n
2n2
∣∣∣∣∣c1, c2 ∈ R}
where
H(2)n =
n∑
i=1
1
i2
RISC, J. Kepler University Linz
![Page 23: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/23.jpg)
A bet (at my cost)
Summation principles (in difference field/ring setting)
n2A(n)− (n+ 1)(2n + 1)A(n + 1) + (n+ 1)(n + 2)A(n + 2) =
1
n+ 1
Summation package Sigma(based on difference field algorithms/theory
see, e.g., Karr 1981, Bronstein 2000, Schneider 2001 –)
A(n) =
∞∑
k=1
Hk
k(k + n)=
0nHn − 1
n2+ ζ2
1
n
+nH2
n − 2Hn + nH(2)n
2n2
where
H(2)n =
n∑
i=1
1
i2ζz =
∞∑
i=1
1
iz(= ζ(z))
RISC, J. Kepler University Linz
![Page 24: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/24.jpg)
A bet (at my cost)
In[1]:= << Sigma.m
Sigma - A summation package by Carsten Schneider c© RISC-Linz
In[2]:= mySum =a∑
k=1
Hk
k(k + n)
RISC, J. Kepler University Linz
![Page 25: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/25.jpg)
A bet (at my cost)
In[1]:= << Sigma.m
Sigma - A summation package by Carsten Schneider c© RISC-Linz
In[2]:= mySum =a∑
k=1
Hk
k(k + n)
In[3]:= rec = GenerateRecurrence[mySum,n][[1]]
Out[3]= n2SUM[n]− (n + 1)(2n + 1)SUM[n+ 1] + (n + 1)(n + 2)SUM[n+ 2] ==(−a− 1)Ha
(a+ n+ 1)(a + n+ 2)+
a
(n+ 1)(a + n+ 1)
RISC, J. Kepler University Linz
![Page 26: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/26.jpg)
A bet (at my cost)
In[1]:= << Sigma.m
Sigma - A summation package by Carsten Schneider c© RISC-Linz
In[2]:= mySum =a∑
k=1
Hk
k(k + n)
In[3]:= rec = GenerateRecurrence[mySum,n][[1]]
Out[3]= n2SUM[n]− (n + 1)(2n + 1)SUM[n+ 1] + (n + 1)(n + 2)SUM[n+ 2] ==(−a− 1)Ha
(a+ n+ 1)(a + n+ 2)+
a
(n+ 1)(a + n+ 1)
In[4]:= rec = LimitRec[rec,SUM[n], {n}, a]
Out[4]= n2SUM[n]− (n + 1)(2n + 1)SUM[n+ 1] + (n + 1)(n + 2)SUM[n+ 2] =1
n+ 1
RISC, J. Kepler University Linz
![Page 27: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/27.jpg)
A bet (at my cost)
In[1]:= << Sigma.m
Sigma - A summation package by Carsten Schneider c© RISC-Linz
In[2]:= mySum =a∑
k=1
Hk
k(k + n)
In[3]:= rec = GenerateRecurrence[mySum,n][[1]]
Out[3]= n2SUM[n]− (n + 1)(2n + 1)SUM[n+ 1] + (n + 1)(n + 2)SUM[n+ 2] ==(−a− 1)Ha
(a+ n+ 1)(a + n+ 2)+
a
(n+ 1)(a + n+ 1)
In[4]:= rec = LimitRec[rec,SUM[n], {n}, a]
Out[4]= n2SUM[n]− (n + 1)(2n + 1)SUM[n+ 1] + (n + 1)(n + 2)SUM[n+ 2] =1
n+ 1
In[5]:= recSol = SolveRecurrence[rec,SUM[n], IndefiniteSummation → True]
Out[5]= {{0, 1n}, {0,
n∑
i=1
1
i
n− 1
n2}, {1,
(
n∑
i=1
1
i
)2
2n−
n∑
i=1
1
i
n2+
n∑
i=1
1
i2
2n}}
RISC, J. Kepler University Linz
![Page 28: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/28.jpg)
A bet (at my cost)
In[1]:= << Sigma.m
Sigma - A summation package by Carsten Schneider c© RISC-Linz
In[2]:= mySum =a∑
k=1
Hk
k(k + n)
In[3]:= rec = GenerateRecurrence[mySum,n][[1]]
Out[3]= n2SUM[n]− (n + 1)(2n + 1)SUM[n+ 1] + (n + 1)(n + 2)SUM[n+ 2] ==(−a− 1)Ha
(a+ n+ 1)(a + n+ 2)+
a
(n+ 1)(a + n+ 1)
In[4]:= rec = LimitRec[rec,SUM[n], {n}, a]
Out[4]= n2SUM[n]− (n + 1)(2n + 1)SUM[n+ 1] + (n + 1)(n + 2)SUM[n+ 2] =1
n+ 1
In[5]:= recSol = SolveRecurrence[rec,SUM[n], IndefiniteSummation → True]
Out[5]= {{0, 1n}, {0,
n∑
i=1
1
i
n− 1
n2}, {1,
(
n∑
i=1
1
i
)2
2n−
n∑
i=1
1
i
n2+
n∑
i=1
1
i2
2n}}
In[6]:= FindLinearCombination[recSol,{1, {ζ2, 1/2 + ζ2/2}},n, 2]
Out[6]= −
n∑
i=1
1
i
n2+
(
n∑
i=1
1
i
)2
2n+
n∑
i=1
1
i2
2n+
ζ2
n
RISC, J. Kepler University Linz
![Page 29: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/29.jpg)
A bet (at my cost)
S =
∞∑
n=1
Hn+1 − 1
n(n+ 1)
∞∑
k=1
Hk
k(k + n)
︸ ︷︷ ︸
=ζ2n
+nH2
n − 2Hn + nH(2)n
2n2
RISC, J. Kepler University Linz
![Page 30: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/30.jpg)
A bet (at my cost)
S =
∞∑
n=1
Hn+1 − 1
n(n+ 1)
∞∑
k=1
Hk
k(k + n)
︸ ︷︷ ︸
=ζ2n
+nH2
n − 2Hn + nH(2)n
2n2
=− 4ζ2 + (ζ2 − 1)
∞∑
i=1
Hi
i2−
∞∑
i=1
H2i
i3
+1
2
∞∑
i=1
H3i
i2+
1
2
∞∑
i=1
HiH(2)i
i2
RISC, J. Kepler University Linz
![Page 31: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/31.jpg)
A bet (at my cost)
S =
∞∑
n=1
Hn+1 − 1
n(n+ 1)
∞∑
k=1
Hk
k(k + n)
︸ ︷︷ ︸
=ζ2n
+nH2
n − 2Hn + nH(2)n
2n2
=− 4ζ2 + (ζ2 − 1)
∞∑
i=1
Hi
i2−
∞∑
i=1
H2i
i3
+1
2
∞∑
i=1
H3i
i2+
1
2
∞∑
i=1
HiH(2)i
i2
=− 4ζ2 − 2ζ3 + 4ζ2ζ3 + 2ζ5 = 0.999222...
J.M. Borwein and R. Girgensohn. Evaluation of triple Euler sums. Electron. J. Combin., 3:1–27, 1996.P. Flajolet and B. Salvy. Euler sums and contour integral representations. Experim. Math., 7(1):15–35, 1998.
RISC, J. Kepler University Linz
![Page 32: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/32.jpg)
A bet (at my cost)
S =
∞∑
n=1
Hn+1 − 1
n(n+ 1)
∞∑
k=1
Hk
k(k + n)
︸ ︷︷ ︸
=ζ2n
+nH2
n − 2Hn + nH(2)n
2n2
=− 4ζ2 + (ζ2 − 1)
∞∑
i=1
Hi
i2−
∞∑
i=1
H2i
i3
+1
2
∞∑
i=1
H3i
i2+
1
2
∞∑
i=1
HiH(2)i
i2
=− 4ζ2 − 2ζ3 + 4ζ2ζ3 + 2ζ5 = 0.999222...
J.M. Borwein and R. Girgensohn. Evaluation of triple Euler sums. Electron. J. Combin., 3:1–27, 1996.P. Flajolet and B. Salvy. Euler sums and contour integral representations. Experim. Math., 7(1):15–35, 1998.
J. Blumlein and D. J. Broadhurst and J. A. M. Vermaseren, The Multiple Zeta Value Data Mine,Comput. Phys. Commun., 181:582–625, 2010.
RISC, J. Kepler University Linz
![Page 33: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/33.jpg)
Summation paradigms
Toolbox 1: Indefinite summation
Toolbox 2: Definite summation
Toolbox 3: Special function algorithms
RISC, J. Kepler University Linz
![Page 34: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/34.jpg)
Summation paradigms
Toolbox 1: Indefinite summation
RISC, J. Kepler University Linz
![Page 35: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/35.jpg)
Toolbox 1: Indefinite summation
Telescoping
GIVEN f(k) = Hk.
FIND g(k):
f(k) = g(k + 1)− g(k)
for all 1 ≤ k ≤ n and n ≥ 0.
RISC, J. Kepler University Linz
![Page 36: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/36.jpg)
Toolbox 1: Indefinite summation
Telescoping
GIVEN f(k) = Hk.
FIND g(k):
f(k) = g(k + 1)− g(k)
for all 1 ≤ k ≤ n and n ≥ 0.
We computeg(k) = (Hk − 1)k.
RISC, J. Kepler University Linz
![Page 37: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/37.jpg)
Toolbox 1: Indefinite summation
Telescoping
GIVEN f(k) = Hk.
FIND g(k):
f(k) = g(k + 1)− g(k)
for all 1 ≤ k ≤ n and n ≥ 0.
Summing this equation over k from 1 to n gives
n∑
k=1
Hk = g(n + 1)− g(1)
=(Hn+1 − 1)(n + 1).
RISC, J. Kepler University Linz
![Page 38: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/38.jpg)
Toolbox 1: Indefinite summation
Telescoping in the given difference field
FIND a closed form forn∑
k=1
Hk.
A difference field for the summand
Consider the rational function field
F
with the automorphism σ : F→ F defined by
RISC, J. Kepler University Linz
![Page 39: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/39.jpg)
Toolbox 1: Indefinite summation
Telescoping in the given difference field
FIND a closed form forn∑
k=1
Hk.
A difference field for the summand
Consider the rational function field
F := Q
with the automorphism σ : F→ F defined by
σ(c) = c ∀c ∈ Q,
RISC, J. Kepler University Linz
![Page 40: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/40.jpg)
Toolbox 1: Indefinite summation
Telescoping in the given difference field
FIND a closed form forn∑
k=1
Hk .
A difference field for the summand
Consider the rational function field
F := Q(k)
with the automorphism σ : F→ F defined by
σ(c) = c ∀c ∈ Q,
σ(k) = k + 1, S k = k + 1,
RISC, J. Kepler University Linz
![Page 41: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/41.jpg)
Toolbox 1: Indefinite summation
Telescoping in the given difference field
FIND a closed form forn∑
k=1
Hk.
A difference field for the summand
Consider the rational function field
F := Q(k)(h)
with the automorphism σ : F→ F defined by
σ(c) = c ∀c ∈ Q,
σ(k) = k + 1, S k = k + 1,
σ(h) = h+1
k + 1, SHk = Hk +
1
k + 1.
RISC, J. Kepler University Linz
![Page 42: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/42.jpg)
Toolbox 1: Indefinite summation
Telescoping in the given difference field
FIND g ∈ F:
σ(g)− g = h.
RISC, J. Kepler University Linz
![Page 43: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/43.jpg)
Toolbox 1: Indefinite summation
Telescoping in the given difference field
FIND g ∈ F:
σ(g)− g = h.
We computeg = (h− 1)k ∈ F.
RISC, J. Kepler University Linz
![Page 44: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/44.jpg)
Toolbox 1: Indefinite summation
Telescoping in the given difference field
FIND g ∈ F:
σ(g)− g = h.
We computeg = (h− 1)k ∈ F.
This gives
g(k + 1)− g(k) = Hk
withg(k) = (Hk − 1)k.
RISC, J. Kepler University Linz
![Page 45: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/45.jpg)
Toolbox 1: Indefinite summation
Telescoping in the given difference field
FIND g ∈ F:
σ(g)− g = h.
We computeg = (h− 1)k ∈ F.
This gives
g(k + 1)− g(k) = Hk
withg(k) = (Hk − 1)k.
Hence,
(Hn+1 − 1)(n+ 1) =
n∑
k=1
Hk.
RISC, J. Kepler University Linz
![Page 46: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/46.jpg)
Toolbox 1: Indefinite summation
Toolbox 1: Indefinite summation– the basic tactic
(a simplified version of Karr’s algorithm, 1981)
RISC, J. Kepler University Linz
![Page 47: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/47.jpg)
Toolbox 1: Indefinite summation
CONSTRUCT a difference field (F, σ):
◮ a rational function field (containing Q)
F := K
◮ with an automorphism
σ(c) = c ∀c ∈ K
RISC, J. Kepler University Linz
![Page 48: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/48.jpg)
Toolbox 1: Indefinite summation
CONSTRUCT a difference field (F, σ):
◮ a rational function field (containing Q)
F := K(t1)
◮ with an automorphism
σ(c) = c ∀c ∈ K
σ(t1) = a1 t1 + f1, a1 ∈ K∗, f1 ∈ K
RISC, J. Kepler University Linz
![Page 49: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/49.jpg)
Toolbox 1: Indefinite summation
CONSTRUCT a difference field (F, σ):
◮ a rational function field (containing Q)
F := K(t1)(t2)
◮ with an automorphism
σ(c) = c ∀c ∈ K
σ(t1) = a1 t1 + f1, a1 ∈ K∗, f1 ∈ K
σ(t2) = a2 t2 + f2, a2 ∈ K(t1)∗, f2 ∈ K(t1)
RISC, J. Kepler University Linz
![Page 50: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/50.jpg)
Toolbox 1: Indefinite summation
CONSTRUCT a difference field (F, σ):
◮ a rational function field (containing Q)
F := K(t1)(t2) . . . (te)
◮ with an automorphism
σ(c) = c ∀c ∈ K
σ(t1) = a1 t1 + f1, a1 ∈ K∗, f1 ∈ K
σ(t2) = a2 t2 + f2, a2 ∈ K(t1)∗, f2 ∈ K(t1)
...
σ(te) = ae te + fe, ae ∈ K(t1, . . . , te−1)∗, fe ∈ K(t1, . . . , te−1)
RISC, J. Kepler University Linz
![Page 51: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/51.jpg)
Toolbox 1: Indefinite summation
CONSTRUCT a difference field (F, σ):
◮ a rational function field (containing Q)
F := K(t1)(t2) . . . (te)
◮ with an automorphism
σ(c) = c ∀c ∈ K
σ(t1) = a1 t1 + f1, a1 ∈ K∗, f1 ∈ K
σ(t2) = a2 t2 + f2, a2 ∈ K(t1)∗, f2 ∈ K(t1)
...
σ(te) = ae te + fe, ae ∈ K(t1, . . . , te−1)∗, fe ∈ K(t1, . . . , te−1)
such that
constσF = {c ∈ K(t1)(t2) . . . (te)|σ(c) = c} = K.
RISC, J. Kepler University Linz
![Page 52: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/52.jpg)
Toolbox 1: Indefinite summation
CONSTRUCT a ΠΣ-field (F, σ):
◮ a rational function field (containing Q)
F := K(t1)(t2) . . . (te)
◮ with an automorphism
σ(c) = c ∀c ∈ K
σ(t1) = a1 t1 + f1, a1 ∈ K∗, f1 ∈ K
σ(t2) = a2 t2 + f2, a2 ∈ K(t1)∗, f2 ∈ K(t1)
...
σ(te) = ae te + fe, ae ∈ K(t1, . . . , te−1)∗, fe ∈ K(t1, . . . , te−1)
such that
constσF = {c ∈ K(t1)(t2) . . . (te)|σ(c) = c} = K.
GIVEN f ∈ F;FIND, in case of existence, a g ∈ F such that
σ(g) − g = f.
RISC, J. Kepler University Linz
![Page 53: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/53.jpg)
Toolbox 1: Indefinite summation
Telescoping in the given difference field
FIND a closed form forn∑
k=1
Hk.
A ΠΣ∗-field for the summand constσF = Q
Consider the rational function field
F := Q(k)(h)
with the automorphism σ : F→ F defined by
σ(c) = c ∀c ∈ Q,
σ(k) = k + 1, S k = k + 1,
σ(h) = h+1
k + 1, SHk = Hk +
1
k + 1.
RISC, J. Kepler University Linz
![Page 54: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/54.jpg)
Toolbox 1: Indefinite summation
FIND g ∈ Q(k)(h):σ(g)− g = h.
RISC, J. Kepler University Linz
![Page 55: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/55.jpg)
Toolbox 1: Indefinite summation
FIND g ∈ Q(k)(h):σ(g)− g = h.
Denominator bound: COMPUTE a polynomial d ∈ Q(k)[h]∗:
∀g ∈ Q(k)(h) : σ(g)− g = h ⇒ g d ∈ Q(k)[h].
FIND g′ ∈ Q(k)[h] with
σ(g′
d)−
g′
d= h.
RISC, J. Kepler University Linz
![Page 56: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/56.jpg)
Toolbox 1: Indefinite summation
FIND g ∈ Q(k)(h):σ(g)− g = h.
Denominator bound: COMPUTE a polynomial d ∈ Q(k)[h]∗:
∀g ∈ Q(k)(h) : σ(g)− g = h ⇒ g d ∈ Q(k)[h].
FIND g′ ∈ Q(k)[h] with
σ(g′
d)−
g′
d= h.
d = 1
RISC, J. Kepler University Linz
![Page 57: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/57.jpg)
Toolbox 1: Indefinite summation
FIND g ∈ Q(k)(h):σ(g)− g = h.
Denominator bound: COMPUTE a polynomial d ∈ Q(k)[h]∗:
∀g ∈ Q(k)(h) : σ(g)− g = h ⇒ g d ∈ Q(k)[h].
FIND g′ ∈ Q(k)[h] with
σ(g′
d)−
g′
d= h.
d = 1
Degree bound: COMPUTE b ≥ 0:
∀g ∈ Q(k)[h] σ(g)− g = h ⇒ deg(g) ≤ b.
RISC, J. Kepler University Linz
![Page 58: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/58.jpg)
Toolbox 1: Indefinite summation
FIND g ∈ Q(k)(h):σ(g)− g = h.
Denominator bound: COMPUTE a polynomial d ∈ Q(k)[h]∗:
∀g ∈ Q(k)(h) : σ(g)− g = h ⇒ g d ∈ Q(k)[h].
FIND g′ ∈ Q(k)[h] with
σ(g′
d)−
g′
d= h.
d = 1
Degree bound: COMPUTE b ≥ 0:
∀g ∈ Q(k)[h] σ(g)− g = h ⇒ deg(g) ≤ b.
b = 2
RISC, J. Kepler University Linz
![Page 59: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/59.jpg)
Toolbox 1: Indefinite summation
FIND g ∈ Q(k)(h):σ(g)− g = h.
Denominator bound: COMPUTE a polynomial d ∈ Q(k)[h]∗:
∀g ∈ Q(k)(h) : σ(g)− g = h ⇒ g d ∈ Q(k)[h].
FIND g′ ∈ Q(k)[h] with
σ(g′
d)−
g′
d= h.
d = 1
Degree bound: COMPUTE b ≥ 0:
∀g ∈ Q(k)[h] σ(g)− g = h ⇒ deg(g) ≤ b.
b = 2
Polynomial Solution: FIND
g = g2 h2 + g1 h+ g0 ∈ Q(k)[h].
RISC, J. Kepler University Linz
![Page 60: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/60.jpg)
Toolbox 1: Indefinite summation
ANSATZ g = g2 h2 + g1 h+ g0 ∈ Q(k)[h]
σ(g) − g = h
RISC, J. Kepler University Linz
![Page 61: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/61.jpg)
Toolbox 1: Indefinite summation
ANSATZ g = g2 h2 + g1 h+ g0 ∈ Q(k)[h]
[σ(g2)
(h+ 1
k+1
)2+ σ(g1h+ g0)
]
−[g2 h
2 + g1h+ g0]= h
RISC, J. Kepler University Linz
![Page 62: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/62.jpg)
Toolbox 1: Indefinite summation
ANSATZ g = g2 h2 + g1 h+ g0 ∈ Q(k)[h]
[σ(g2)
(h+ 1
k+1
)2+ σ(g1h+ g0)
]
−[g2 h
2 + g1h+ g0]= h coeff. comp.
%%❑❑❑
❑❑❑❑
❑
σ(g2)− g2 = 0
RISC, J. Kepler University Linz
![Page 63: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/63.jpg)
Toolbox 1: Indefinite summation
ANSATZ g = g2 h2 + g1 h+ g0 ∈ Q(k)[h]
[σ(g2)
(h+ 1
k+1
)2+ σ(g1h+ g0)
]
−[g2 h
2 + g1h+ g0]= h coeff. comp.
%%❑❑❑
❑❑❑❑
❑
σ(g2)− g2 = 0
g2 = c ∈ Q
yysssssssss
RISC, J. Kepler University Linz
![Page 64: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/64.jpg)
Toolbox 1: Indefinite summation
ANSATZ g = g2 h2 + g1 h+ g0 ∈ Q(k)[h]
[σ(g2)
(h+ 1
k+1
)2+ σ(g1h+ g0)
]
−[g2 h
2 + g1h+ g0]= h coeff. comp.
%%❑❑❑
❑❑❑❑
❑
σ(g2)− g2 = 0
g2 = c ∈ Q
yysssssssss
[σ(c)
(h+ 1
k+1
)2+ σ(g1h+ g0)
]
−[c h2 + g1h+ g0
]= h
RISC, J. Kepler University Linz
![Page 65: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/65.jpg)
Toolbox 1: Indefinite summation
ANSATZ g = g2 h2 + g1 h+ g0 ∈ Q(k)[h]
[σ(g2)
(h+ 1
k+1
)2+ σ(g1h+ g0)
]
−[g2 h
2 + g1h+ g0]= h coeff. comp.
%%❑❑❑
❑❑❑❑
❑
σ(g2)− g2 = 0
g2 = c ∈ Q
yysssssssss
[c(h+ 1
k+1
)2+ σ(g1h+ g0)
]
−[c h2 + g1h+ g0
]= h
RISC, J. Kepler University Linz
![Page 66: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/66.jpg)
Toolbox 1: Indefinite summation
ANSATZ g = g2 h2 + g1 h+ g0 ∈ Q(k)[h]
[σ(g2)
(h+ 1
k+1
)2+ σ(g1h+ g0)
]
−[g2 h
2 + g1h+ g0]= h coeff. comp.
%%❑❑❑
❑❑❑❑
❑
σ(g2)− g2 = 0
g2 = c ∈ Q
yysssssssss
σ(g1 h+ g0)− (g1 h+ g0) = h− c[2h(k+1)+1
(k+1)2
]
RISC, J. Kepler University Linz
![Page 67: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/67.jpg)
Toolbox 1: Indefinite summation
ANSATZ g = g2 h2 + g1 h+ g0 ∈ Q(k)[h]
[σ(g2)
(h+ 1
k+1
)2+ σ(g1h+ g0)
]
−[g2 h
2 + g1h+ g0]= h coeff. comp.
%%❑❑❑
❑❑❑❑
❑
σ(g2)− g2 = 0
g2 = c ∈ Q
yysssssssss
σ(g1 h+ g0)− (g1 h+ g0) = h− c[2h(k+1)+1
(k+1)2
]
coeff. comp.%%❑
❑❑❑❑
❑❑❑
σ(g1)− g1 = 1− c 2k+1
RISC, J. Kepler University Linz
![Page 68: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/68.jpg)
Toolbox 1: Indefinite summation
ANSATZ g = g2 h2 + g1 h+ g0 ∈ Q(k)[h]
[σ(g2)
(h+ 1
k+1
)2+ σ(g1h+ g0)
]
−[g2 h
2 + g1h+ g0]= h coeff. comp.
%%❑❑❑
❑❑❑❑
❑
σ(g2)− g2 = 0
g2 = c ∈ Q
yysssssssss
σ(g1 h+ g0)− (g1 h+ g0) = h− c[2h(k+1)+1
(k+1)2
]
coeff. comp.%%❑
❑❑❑❑
❑❑❑
σ(g1)− g1 = 1− c 2k+1
c = 0,g1 = k + d
d ∈ Q
yyssssssss
RISC, J. Kepler University Linz
![Page 69: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/69.jpg)
Toolbox 1: Indefinite summation
ANSATZ g = g2 h2 + g1 h+ g0 ∈ Q(k)[h]
[σ(g2)
(h+ 1
k+1
)2+ σ(g1h+ g0)
]
−[g2 h
2 + g1h+ g0]= h coeff. comp.
%%❑❑❑
❑❑❑❑
❑
σ(g2)− g2 = 0
g2 = c ∈ Q
yysssssssss
σ(g1 h+ g0)− (g1 h+ g0) = h− c[2h(k+1)+1
(k+1)2
]
coeff. comp.%%❑
❑❑❑❑
❑❑❑
σ(g1)− g1 = 1− c 2k+1
c = 0,g1 = k + d
d ∈ Q
yyssssssss
σ(g0)− g0 = −1− d1
k + 1
RISC, J. Kepler University Linz
![Page 70: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/70.jpg)
Toolbox 1: Indefinite summation
ANSATZ g = g2 h2 + g1 h+ g0 ∈ Q(k)[h]
[σ(g2)
(h+ 1
k+1
)2+ σ(g1h+ g0)
]
−[g2 h
2 + g1h+ g0]= h coeff. comp.
%%❑❑❑
❑❑❑❑
❑
σ(g2)− g2 = 0
g2 = c ∈ Q
yysssssssss
σ(g1 h+ g0)− (g1 h+ g0) = h− c[2h(k+1)+1
(k+1)2
]
coeff. comp.%%❑
❑❑❑❑
❑❑❑
σ(g1)− g1 = 1− c 2k+1
c = 0,g1 = k + d
d ∈ Q
yyssssssss
g0 = −kd = 0
← σ(g0)− g0 = −1− d1
k + 1
g=hk-k
RISC, J. Kepler University Linz
![Page 71: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/71.jpg)
Toolbox 1: Indefinite summation
Toolbox 1: Improved indefinite summation
– symbolic simplificationFor algorithmic details see:
◮ CS. Symbolic summation with single-nested sum extensions. In J. Gutierrez, editor, Proc. ISSAC’04, pages 282–289.ACM Press, 2004.
◮ CS. Product representations in ΠΣ-fields. Ann. Comb., 9(1):75–99, 2005.
◮ CS. Simplifying Sums in ΠΣ-Extensions. J. Algebra Appl., 6(3):415–441, 2007.
◮ CS. A refined difference field theory for symbolic summation. J. Symbolic Comput., 43(9):611–644, 2008.[arXiv:0808.2543v1].
◮ S.A. Abramov, M. Petkovsek. Polynomial ring automorphisms, rational (w, σ)-canonical forms, and the assignmentproblem. J. Symbolic Comput., 45(6): 684–708, 2010.
◮ CS, A Symbolic Summation Approach to Find Optimal Nested Sum Representations. In: A. Carey, D. Ellwood,S. Paycha, S. Rosenberg (eds.) Motives, Quantum Field Theory, and Pseudodifferential Operators, Clay MathematicsProceedings, vol. 12, pp. 285–308. Amer. Math. Soc (2010). ArXiv:0808.2543
◮ CS, Parameterized Telescoping Proves Algebraic Independence of Sums. Ann. Comb. 14(4), 533–552 (2010).[arXiv:0808.2596]
◮ CS. Structural Theorems for Symbolic Summation. Appl. Algebra Engrg. Comm. Comput., 21(1):1–32, 2010.
◮ CS. Fast Algorithms for Refined Parameterized Telescoping in Difference Fields. To appear in Computer Algebra and
Polynomials, Lecture Notes in Computer Science (LNCS), Springer, 2014. arXiv:1307.7887 [cs.SC].
For special cases see:◮ S.A. Abramov. On the summation of rational functions. Zh. vychisl. mat. Fiz., 11: 1071-1074, 1971.
◮ P. Paule. Greatest factorial factorization and symbolic summation, J. Symbolic Comput., 20(3): 235-268, 1995.
RISC, J. Kepler University Linz
![Page 72: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/72.jpg)
Toolbox 1: Indefinite summation
A difference field approach (M. Karr, 1981)
GIVEN a ΠΣ-field (F, σ) with f ∈ F.
FIND g ∈ F:σ(g) − g = f.
RISC, J. Kepler University Linz
![Page 73: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/73.jpg)
Toolbox 1: Indefinite summation
A symbolic summation approach
1. FIND an appropriate ΠΣ-field (F, σ) with f ∈ F.
2. FIND g ∈ F:σ(g) − g = f.
RISC, J. Kepler University Linz
![Page 74: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/74.jpg)
Toolbox 1: Indefinite summation
A symbolic summation approach
1. FIND an appropriate ΠΣ-field (F, σ) with f ∈ F.
2. FIND an appropriate extension E > F with g ∈ E:
σ(g) − g = f.
RISC, J. Kepler University Linz
![Page 75: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/75.jpg)
Toolbox 1: Indefinite summation
A symbolic summation approach
1. FIND an appropriate ΠΣ-field (F, σ) with f ∈ F.
2. FIND an appropriate extension E > F with g ∈ E:
σ(g) − g = f.
appropriate = degrees in denominators minimal
Example:
a∑
k=1
( −2 + k
10(1 + k2)+
(1− 4k − 2k2)Hk
10(1 + k2)(2 + 2k + k2)+
(1− 4k − 2k2)H(3)k
5(1 + k2)(2 + 2k + k2)
)
= ?
RISC, J. Kepler University Linz
![Page 76: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/76.jpg)
Toolbox 1: Indefinite summation
A symbolic summation approach
1. FIND an appropriate ΠΣ-field (F, σ) with f ∈ F.
2. FIND an appropriate extension E > F with g ∈ E:
σ(g) − g = f.
appropriate = degrees in denominators minimal
Example:
a∑
k=1
( −2 + k
10(1 + k2)+
(1− 4k − 2k2)Hk
10(1 + k2)(2 + 2k + k2)+
(1− 4k − 2k2)H(3)k
5(1 + k2)(2 + 2k + k2)
)
=a2 + 4a+ 5
10(a2 + 2a+ 2)Ha −
(a− 1)(a+ 1)
5(a2 + 2a+ 2)H(3)
a −2
5
a∑
k=1
1
k2
RISC, J. Kepler University Linz
![Page 77: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/77.jpg)
Toolbox 1: Indefinite summation
A symbolic summation approach
1. FIND an appropriate ΠΣ-field (F, σ) with f ∈ F.
2. FIND an appropriate extension E > F with g ∈ E:
σ(g) − g = f.
appropriate = sum representations with optimal nesting depth
Example:
n∑
k=1
k∑
j=1
j∑
i=1
1
i
j
k= ?
RISC, J. Kepler University Linz
![Page 78: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/78.jpg)
Toolbox 1: Indefinite summation
A symbolic summation approach
1. FIND an appropriate ΠΣ-field (F, σ) with f ∈ F.
2. FIND an appropriate extension E > F with g ∈ E:
σ(g) − g = f.
appropriate = sum representations with optimal nesting depth
Example:
n∑
k=1
k∑
j=1
j∑
i=1
1
i
j
k=
1
6
(n∑
i=1
1
i
)3
+1
2
(n∑
i=1
1
i2
)(n∑
i=1
1
i
)
+1
3
n∑
i=1
1
i3
depth 3 depth 1
RISC, J. Kepler University Linz
![Page 79: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/79.jpg)
Toolbox 1: Indefinite summation
A symbolic summation approach
1. FIND an appropriate ΠΣ-field (F, σ) with f ∈ F.
2. FIND an appropriate extension E > F with g ∈ E:
σ(g) − g = f.
appropriate = sum representations with minimal number of objects
Example:
a∑
k=0
(−1)kHk2
(n
k
)
=?
RISC, J. Kepler University Linz
![Page 80: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/80.jpg)
Toolbox 1: Indefinite summation
A symbolic summation approach
1. FIND an appropriate ΠΣ-field (F, σ) with f ∈ F.
2. FIND an appropriate extension E > F with g ∈ E:
σ(g) − g = f.
appropriate = sum representations with minimal number of objects
Example:
a∑
k=0
(−1)kHk2
(n
k
)
=−1
n
a∑
i1=1
(−1)i1
i1
(n
i1
)
− (a− n)(n2Ha
2 + 2nHa + 2)(−1)a
(na
)
n3−
2
n2
RISC, J. Kepler University Linz
![Page 81: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/81.jpg)
Toolbox 1: Indefinite summation
Simplification of nested product-sum expressions
A(k): nested product-sum expression (sums/products not in the denominator)
↓ SigmaReduce[A,k]
B(k): nested product-sum expression (sums/products not in the denominator)
◮ such thatA(k) = B(k)
RISC, J. Kepler University Linz
![Page 82: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/82.jpg)
Toolbox 1: Indefinite summation
Simplification of nested product-sum expressions
A(k): nested product-sum expression (sums/products not in the denominator)
↓ SigmaReduce[A,k]
B(k): nested product-sum expression (sums/products not in the denominator)
◮ such thatA(k) = B(k)
◮ such that all the sums in B(k) are simplified as above
RISC, J. Kepler University Linz
![Page 83: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/83.jpg)
Toolbox 1: Indefinite summation
Simplification of nested product-sum expressions
A(k): nested product-sum expression (sums/products not in the denominator)
↓ SigmaReduce[A,k]
B(k): nested product-sum expression (sums/products not in the denominator)
◮ such thatA(k) = B(k)
◮ such that all the sums in B(k) are simplified as above
◮ and such that the arising sums in B(k) are algebraically independent(i.e., they do not satisfy any polynomial relation)
RISC, J. Kepler University Linz
![Page 84: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/84.jpg)
Toolbox 2: Definite summation
Toolbox 2: Definite summation
RISC, J. Kepler University Linz
![Page 85: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/85.jpg)
Toolbox 2: Definite summation
Summation principles (in difference field/ring setting)
n2A(n)− (n+ 1)(2n + 1)A(n + 1) + (n+ 1)(n + 2)A(n + 2) =
1
n+ 1
Summation package Sigma(based on difference field algorithms/theory
see, e.g., Karr 1981, Bronstein 2000, Schneider 2001 –)
A(n) =
∞∑
k=1
Hk
k(k + n)=
0nHn − 1
n2+ ζ2
1
n
+nH2
n − 2Hn + nH(2)n
2n2
where
H(2)n =
n∑
i=1
1
i2ζz =
∞∑
i=1
1
iz(= ζ(z))
RISC, J. Kepler University Linz
![Page 86: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/86.jpg)
Toolbox 2: Definite summation
1. Creative telescoping (for the special case of hypergeometric terms see Zeilberger’s algorithm (1991))
GIVEN a definite sum
A(n) =n∑
k=1
f(n, k); f(n, k): indefinite nested product-sum in k;n: extra parameter
FIND a recurrence for A(n)
RISC, J. Kepler University Linz
![Page 87: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/87.jpg)
Toolbox 2: Definite summation
Back to creative telescopingGiven
f(n, k) =Hk
k(k + n);
Find g(n, k) and c0(n), c1(n), c2(n) :
g(n, k + 1)− g(n, k) = c0(n) f(n, k) + c1(n) f(n+ 1, k) + c2(n) f(n+ 2, k)
RISC, J. Kepler University Linz
![Page 88: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/88.jpg)
Toolbox 2: Definite summation
Back to creative telescopingGiven
f(n, k) =Hk
k(k + n);
Find g(n, k) and c0(n), c1(n), c2(n) :
g(n, k + 1)− g(n, k) = c0(n)Hk
k(k+n) + c1(n) f(n+ 1, k) + c2(n) f(n+ 2, k)
RISC, J. Kepler University Linz
![Page 89: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/89.jpg)
Toolbox 2: Definite summation
Back to creative telescopingGiven
f(n, k) =Hk
k(k + n);
Find g(n, k) and c0(n), c1(n), c2(n) :
g(n, k + 1)− g(n, k) = c0(n)Hk
k(k+n) + c1(n)Hk
k(k+n+1) + c2(n) f(n+ 2, k)
RISC, J. Kepler University Linz
![Page 90: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/90.jpg)
Toolbox 2: Definite summation
Back to creative telescopingGiven
f(n, k) =Hk
k(k + n);
Find g(n, k) and c0(n), c1(n), c2(n) :
g(n, k + 1)− g(n, k) = c0(n)Hk
k(k+n) + c1(n)Hk
k(k+n+1) + c2(n)Hk
k(k+n+2)
RISC, J. Kepler University Linz
![Page 91: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/91.jpg)
Toolbox 2: Definite summation
Back to creative telescopingGiven
f(n, k) =Hk
k(k + n);
Find g(n, k) and c0(n), c1(n), c2(n) :
g(n, k + 1)− g(n, k) = c0(n)Hk
k(k+n) + c1(n)Hk
k(k+n+1) + c2(n)Hk
k(k+n+2)
A difference field for the summand:Construct a rational function field
F
and a field automorphism σ : F→ F defined by
RISC, J. Kepler University Linz
![Page 92: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/92.jpg)
Toolbox 2: Definite summation
Back to creative telescopingGiven
f(n, k) =Hk
k(k +n);
Find g(n, k) and c0(n), c1(n), c2(n) :
g(n, k + 1)− g(n, k) = c0(n)Hk
k(k+n) + c1(n)Hk
k(k+n+1) + c2(n)Hk
k(k+n+2)
A difference field for the summand:Construct a rational function field
F := Q(n)
and a field automorphism σ : F→ F defined by
σ(c) = c ∀c ∈ Q(n),
RISC, J. Kepler University Linz
![Page 93: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/93.jpg)
Toolbox 2: Definite summation
Back to creative telescopingGiven
f(n, k) =Hk
k(k + n);
Find g(n, k) and c0(n), c1(n), c2(n) :
g(n, k + 1)− g(n, k) = c0(n)Hk
k(k+n) + c1(n)Hk
k(k+n+1) + c2(n)Hk
k(k+n+2)
A difference field for the summand:Construct a rational function field
F := Q(n)(k)
and a field automorphism σ : F→ F defined by
σ(c) = c ∀c ∈ Q(n),
σ(k) = k+ 1, S k = k + 1,
RISC, J. Kepler University Linz
![Page 94: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/94.jpg)
Toolbox 2: Definite summation
Back to creative telescopingGiven
f(n, k) =Hk
k(k + n);
Find g(n, k) and c0(n), c1(n), c2(n) :
g(n, k + 1)− g(n, k) = c0(n)Hk
k(k+n) + c1(n)Hk
k(k+n+1) + c2(n)Hk
k(k+n+2)
A difference field for the summand:Construct a rational function field (F, σ) is a ΠΣ-field
F := Q(n)(k)(h) Karr 1981
and a field automorphism σ : F→ F defined by
σ(c) = c ∀c ∈ Q(n),
σ(k) = k + 1, S k = k + 1,
σ(h) = h+1
k+ 1, SHk = Hk +
1
k + 1,
RISC, J. Kepler University Linz
![Page 95: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/95.jpg)
Toolbox 2: Definite summation
Back to creative telescopingGiven
f(n, k) =Hk
k(k + n);
Find g(n, k) and c0(n), c1(n), c2(n) :
g(n, k + 1)− g(n, k) = c0(n)Hk
k(k+n) + c1(n)Hk
k(k+n+1) + c2(n)Hk
k(k+n+2)
FIND g ∈ F and c0, c1, c2 ∈ Q(n):
σ(g)− g = c0h
k(k + n)+ c1
h
k(k + n+ 1)+ c2
h
k(k + n+ 2)
RISC, J. Kepler University Linz
![Page 96: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/96.jpg)
Toolbox 2: Definite summation
Back to creative telescopingGiven
f(n, k) =Hk
k(k + n);
Find g(n, k) and c0(n), c1(n), c2(n) :
g(n, k + 1)− g(n, k) = c0(n)Hk
k(k+n) + c1(n)Hk
k(k+n+1) + c2(n)Hk
k(k+n+2)
FIND g ∈ F and c0, c1, c2 ∈ Q(n):
σ(g)− g = c0h
k(k + n)+ c1
h
k(k + n+ 1)+ c2
h
k(k + n+ 2)
↓
c0 = n2, c1 = −(n+ 1)(2n + 1), c2 = (n+ 1)(n + 2)
g = −kh+ n+ k
(n+ k)(n + k + 1)
RISC, J. Kepler University Linz
![Page 97: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/97.jpg)
Toolbox 2: Definite summation
Zeilberger’s creative telescoping paradigmGIVEN
A’(n) :=
a∑
k=1
Hk
k(k + n)︸ ︷︷ ︸
=: f(n, k)
.
FIND g(n, k) and c0(n), c1(n), c2(n):
g(n, k + 1)− g(n, k) = c0(n)f(n, k) + c1(n) f(n+ 1, k) + c2(n) f(n+ 2, k)
for all n, k ≥ 1.
Sigma computes: c0(n) = n2, c1(n) = −(n+ 1)(2n+ 1), c2(n) = (n+ 1)(n+ 2)
andg(n, k) := −
kHk + n+ k
(n+ k)(n+ k + 1),
g(n, k + 1) := −(1 + n)Hk + n+ k + 2
(n+ k + 1)(n+ k + 2).
RISC, J. Kepler University Linz
![Page 98: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/98.jpg)
Toolbox 2: Definite summation
Zeilberger’s creative telescoping paradigmGIVEN
A’(n) :=
a∑
k=1
Hk
k(k + n)︸ ︷︷ ︸
=: f(n, k)
.
FIND g(n, k) and c0(n), c1(n), c2(n):
g(n, k + 1)− g(n, k) = c0(n)f(n, k) + c1(n) f(n+ 1, k) + c2(n) f(n+ 2, k)
for all n, k ≥ 1.
Summing this equation over k from 1 to a gives:
g(n, a+1)− g(n, 1) =
a∑
k=1
[
c0(n)f(n, k)+c1(n)f(n+1, k)+c2(n)f(n+2, k)]
RISC, J. Kepler University Linz
![Page 99: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/99.jpg)
Toolbox 2: Definite summation
Zeilberger’s creative telescoping paradigmGIVEN
A’(n) :=
a∑
k=1
Hk
k(k + n)︸ ︷︷ ︸
=: f(n, k)
.
FIND g(n, k) and c0(n), c1(n), c2(n):
g(n, k + 1)− g(n, k) = c0(n)f(n, k) + c1(n) f(n+ 1, k) + c2(n) f(n+ 2, k)
for all n, k ≥ 1.
Summing this equation over k from 1 to a gives:
g(n, a+1)− g(n, 1) =
a∑
k=1
c0(n)f(n, k)+
a∑
k=1
c1(n)f(n+1, k)+
a∑
k=1
c2(n)f(n+2, k)
RISC, J. Kepler University Linz
![Page 100: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/100.jpg)
Toolbox 2: Definite summation
Zeilberger’s creative telescoping paradigmGIVEN
A’(n) :=
a∑
k=1
Hk
k(k + n)︸ ︷︷ ︸
=: f(n, k)
.
FIND g(n, k) and c0(n), c1(n), c2(n):
g(n, k + 1)− g(n, k) = c0(n)f(n, k) + c1(n) f(n+ 1, k) + c2(n) f(n+ 2, k)
for all n, k ≥ 1.
Summing this equation over k from 1 to a gives:
g(n, a+1)− g(n, 1) = c0(n)
a∑
k=1
f(n, k)+c1(n)
a∑
k=1
f(n+1, k)+c2(n)
a∑
k=1
f(n+2, k)
RISC, J. Kepler University Linz
![Page 101: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/101.jpg)
Toolbox 2: Definite summation
Zeilberger’s creative telescoping paradigmGIVEN
A’(n) :=
a∑
k=1
Hk
k(k + n)︸ ︷︷ ︸
=: f(n, k)
.
FIND g(n, k) and c0(n), c1(n), c2(n):
g(n, k + 1)− g(n, k) = c0(n)f(n, k) + c1(n) f(n+ 1, k) + c2(n) f(n+ 2, k)
for all n, k ≥ 1.
Summing this equation over k from 1 to a gives:
g(n, a+1)− g(n, 1) = c0(n)A’(n) + c1(n)A’(n+ 1) + c2(n)A’(n+ 2)
RISC, J. Kepler University Linz
![Page 102: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/102.jpg)
Toolbox 2: Definite summation
Zeilberger’s creative telescoping paradigmGIVEN
A’(n) :=
a∑
k=1
Hk
k(k + n)︸ ︷︷ ︸
=: f(n, k)
.
FIND g(n, k) and c0(n), c1(n), c2(n):
g(n, k + 1)− g(n, k) = c0(n)f(n, k) + c1(n) f(n+ 1, k) + c2(n) f(n+ 2, k)
for all n, k ≥ 1.
Summing this equation over k from 1 to a gives:
g(n, a+1)− g(n, 1) = c0(n)A’(n) + c1(n)A’(n+ 1) + c2(n)A’(n+ 2)
|| ||a
(n+1)(a+n+1) n2A’(n)− (n+1)(2n+1)A’(n+ 1) + (n+1)(n+2)A’(n+2)
− (a+1)Ha
(a+n+1)(a+n+2)
RISC, J. Kepler University Linz
![Page 103: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/103.jpg)
Toolbox 2: Definite summation
1. Creative telescoping (for the special case of hypergeometric terms see Zeilberger’s algorithm (1991))
GIVEN a definite sum
A(n) =n∑
k=1
f(n, k); f(n, k): indefinite nested product-sum in k;n: extra parameter
FIND a recurrence for A(n)
RISC, J. Kepler University Linz
![Page 104: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/104.jpg)
Toolbox 2: Definite summation
1. Creative telescoping (for the special case of hypergeometric terms see Zeilberger’s algorithm (1991))
GIVEN a definite sum
A(n) =n∑
k=1
f(n, k); f(n, k): indefinite nested product-sum in k;n: extra parameter
FIND a recurrence for A(n)
2. Recurrence solving
GIVEN a recurrence a0(n), . . . , ad(n), h(n):indefinite nested product-sum expressions in n.
a0(n)A(n) + · · ·+ ad(n)A(n + d) = h(n);
FIND all solutions expressible by indefinite nested products/sums in n.(d’Alembertian solutions)(Abramov/Bronstein/Petkovsek/CS, in preparation)
RISC, J. Kepler University Linz
![Page 105: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/105.jpg)
Toolbox 2: Definite summation
1. Creative telescoping (for the special case of hypergeometric terms see Zeilberger’s algorithm (1991))
GIVEN a definite sum
A(n) =n∑
k=1
f(n, k); f(n, k): indefinite nested product-sum in k;n: extra parameter
FIND a recurrence for A(n)
2. Recurrence solving
GIVEN a recurrence a0(n), . . . , ad(n), h(n):indefinite nested product-sum expressions in n.
a0(n)A(n) + · · ·+ ad(n)A(n + d) = h(n);
FIND all solutions expressible by indefinite nested products/sums in n.(d’Alembertian solutions)(Abramov/Bronstein/Petkovsek/CS, in preparation)
Note: the sum solutions are highly nested(possibly with denominators of high degrees)
RISC, J. Kepler University Linz
![Page 106: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/106.jpg)
Toolbox 2: Definite summation
1. Creative telescoping (for the special case of hypergeometric terms see Zeilberger’s algorithm (1991))
GIVEN a definite sum
A(n) =n∑
k=1
f(n, k); f(n, k): indefinite nested product-sum in k;n: extra parameter
FIND a recurrence for A(n)
2. Recurrence solving
GIVEN a recurrence a0(n), . . . , ad(n), h(n):indefinite nested product-sum expressions in n.
a0(n)A(n) + · · ·+ ad(n)A(n + d) = h(n);
FIND all solutions expressible by indefinite nested products/sums in n.(d’Alembertian solutions)(Abramov/Bronstein/Petkovsek/CS, in preparation)
3. Simplify the solutions (using difference field theory) s.t.◮ the sums are algebraically independent;◮ the sums are flattened;◮ the sums can be given in terms of special functions.
RISC, J. Kepler University Linz
![Page 107: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/107.jpg)
Toolbox 2: Definite summation
1. Creative telescoping (for the special case of hypergeometric terms see Zeilberger’s algorithm (1991))
GIVEN a definite sum
A(n) =n∑
k=1
f(n, k); f(n, k): indefinite nested product-sum in k;n: extra parameter
FIND a recurrence for A(n)
2. Recurrence solving
GIVEN a recurrence a0(n), . . . , ad(n), h(n):indefinite nested product-sum expressions in n.
a0(n)A(n) + · · ·+ ad(n)A(n + d) = h(n);
FIND all solutions expressible by indefinite nested products/sums in n.(d’Alembertian solutions)(Abramov/Bronstein/Petkovsek/CS, in preparation)
4. Find a “closed form”
A(n)=combined solutions in terms of indefinite nested sums in n.
RISC, J. Kepler University Linz
![Page 108: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/108.jpg)
Toolbox 2: Definite summation
In[1]:= << Sigma.m
Sigma - A summation package by Carsten Schneider c© RISC-Linz
In[2]:= mySum =a∑
k=1
Hk
k(k + n)
RISC, J. Kepler University Linz
![Page 109: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/109.jpg)
Toolbox 2: Definite summation
In[1]:= << Sigma.m
Sigma - A summation package by Carsten Schneider c© RISC-Linz
In[2]:= mySum =a∑
k=1
Hk
k(k + n)
In[3]:= rec = GenerateRecurrence[mySum,n][[1]]
Out[3]= n2SUM[n]− (n + 1)(2n + 1)SUM[n+ 1] + (n + 1)(n + 2)SUM[n+ 2] ==(−a− 1)Ha
(a+ n+ 1)(a + n+ 2)+
a
(n+ 1)(a + n+ 1)
In[4]:= rec = LimitRec[rec,SUM[n], {n}, a]
Out[4]= n2SUM[n]− (n + 1)(2n + 1)SUM[n+ 1] + (n + 1)(n + 2)SUM[n+ 2] =1
n+ 1
RISC, J. Kepler University Linz
![Page 110: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/110.jpg)
Toolbox 2: Definite summation
In[1]:= << Sigma.m
Sigma - A summation package by Carsten Schneider c© RISC-Linz
In[2]:= mySum =a∑
k=1
Hk
k(k + n)
In[3]:= rec = GenerateRecurrence[mySum,n][[1]]
Out[3]= n2SUM[n]− (n + 1)(2n + 1)SUM[n+ 1] + (n + 1)(n + 2)SUM[n+ 2] ==(−a− 1)Ha
(a+ n+ 1)(a + n+ 2)+
a
(n+ 1)(a + n+ 1)
In[4]:= rec = LimitRec[rec,SUM[n], {n}, a]
Out[4]= n2SUM[n]− (n + 1)(2n + 1)SUM[n+ 1] + (n + 1)(n + 2)SUM[n+ 2] =1
n+ 1
In[5]:= recSol = SolveRecurrence[rec,SUM[n], IndefiniteSummation → False]
RISC, J. Kepler University Linz
![Page 111: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/111.jpg)
Toolbox 2: Definite summation
In[1]:= << Sigma.m
Sigma - A summation package by Carsten Schneider c© RISC-Linz
In[2]:= mySum =a∑
k=1
Hk
k(k + n)
In[3]:= rec = GenerateRecurrence[mySum,n][[1]]
Out[3]= n2SUM[n]− (n + 1)(2n + 1)SUM[n+ 1] + (n + 1)(n + 2)SUM[n+ 2] ==(−a− 1)Ha
(a+ n+ 1)(a + n+ 2)+
a
(n+ 1)(a + n+ 1)
In[4]:= rec = LimitRec[rec,SUM[n], {n}, a]
Out[4]= n2SUM[n]− (n + 1)(2n + 1)SUM[n+ 1] + (n + 1)(n + 2)SUM[n+ 2] =1
n+ 1
In[5]:= recSol = SolveRecurrence[rec,SUM[n], IndefiniteSummation → False]
Out[5]= {{0, 1n}, {0,− 1
n2+
n∑
i=1
1
i
n}, {1,−
n∑
i=1
1
i
n2+
n∑
k=1
k∑
i=1
1
i
k
n}}
RISC, J. Kepler University Linz
![Page 112: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/112.jpg)
Toolbox 2: Definite summation
In[1]:= << Sigma.m
Sigma - A summation package by Carsten Schneider c© RISC-Linz
In[2]:= mySum =a∑
k=1
Hk
k(k + n)
In[3]:= rec = GenerateRecurrence[mySum,n][[1]]
Out[3]= n2SUM[n]− (n + 1)(2n + 1)SUM[n+ 1] + (n + 1)(n + 2)SUM[n+ 2] ==(−a− 1)Ha
(a+ n+ 1)(a + n+ 2)+
a
(n+ 1)(a + n+ 1)
In[4]:= rec = LimitRec[rec,SUM[n], {n}, a]
Out[4]= n2SUM[n]− (n + 1)(2n + 1)SUM[n+ 1] + (n + 1)(n + 2)SUM[n+ 2] =1
n+ 1
In[5]:= recSol = SolveRecurrence[rec,SUM[n], IndefiniteSummation → True]
Out[5]= {{0, 1n}, {0,
n∑
i=1
1
i
n− 1
n2}, {1,
(
n∑
i=1
1
i
)2
2n−
n∑
i=1
1
i
n2+
n∑
i=1
1
i2
2n}}
RISC, J. Kepler University Linz
![Page 113: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/113.jpg)
Toolbox 2: Definite summation
In[1]:= << Sigma.m
Sigma - A summation package by Carsten Schneider c© RISC-Linz
In[2]:= mySum =a∑
k=1
Hk
k(k + n)
In[3]:= rec = GenerateRecurrence[mySum,n][[1]]
Out[3]= n2SUM[n]− (n + 1)(2n + 1)SUM[n+ 1] + (n + 1)(n + 2)SUM[n+ 2] ==(−a− 1)Ha
(a+ n+ 1)(a + n+ 2)+
a
(n+ 1)(a + n+ 1)
In[4]:= rec = LimitRec[rec,SUM[n], {n}, a]
Out[4]= n2SUM[n]− (n + 1)(2n + 1)SUM[n+ 1] + (n + 1)(n + 2)SUM[n+ 2] =1
n+ 1
In[5]:= recSol = SolveRecurrence[rec,SUM[n], IndefiniteSummation → True]
Out[5]= {{0, 1n}, {0,
n∑
i=1
1
i
n− 1
n2}, {1,
(
n∑
i=1
1
i
)2
2n−
n∑
i=1
1
i
n2+
n∑
i=1
1
i2
2n}}
In[6]:= FindLinearCombination[recSol,{1, {ζ2, 1/2 + ζ2/2}},n, 2]
Out[6]= −
n∑
i=1
1
i
n2+
(
n∑
i=1
1
i
)2
2n+
n∑
i=1
1
i2
2n+
ζ2
n
RISC, J. Kepler University Linz
![Page 114: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/114.jpg)
Toolbox 2: Definite summation
Sigma’s summation spiral
RISC, J. Kepler University Linz
![Page 115: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/115.jpg)
Toolbox 3: Special function algorithms
Toolbox 3: Special function algorithms
RISC, J. Kepler University Linz
![Page 116: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/116.jpg)
Toolbox 3: Special function algorithms
Computer algebra and special functions:Harmonic sums (Borwein, Hoffman, Broadhurst, Vermaseren, Remmiddi, Blumlein,. . . )
n∑
i=1
1
i2
i∑
j=1
1
j
(J. Ablinger, J. Blumlein, CS; J. Math. Phys. 2011 [arXiv:1105.6063 [math-ph]])RISC, J. Kepler University Linz
![Page 117: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/117.jpg)
Toolbox 3: Special function algorithms
Computer algebra and special functions:Harmonic sums (Borwein, Hoffman, Broadhurst, Vermaseren, Remmiddi, Blumlein,. . . )
n∑
i=1
1
i2
i∑
j=1
1
j
Integral representation:
=∫ 1
0
xn − 1
1− x
(
∫ x
0
∫ y
01
1−zdz
ydy − ζ2
)
dx, ζz :=
∞∑
i=1
1/iz
(J. Ablinger, J. Blumlein, CS; J. Math. Phys. 2011 [arXiv:1105.6063 [math-ph]])RISC, J. Kepler University Linz
![Page 118: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/118.jpg)
Toolbox 3: Special function algorithms
Computer algebra and special functions:Harmonic sums (Borwein, Hoffman, Broadhurst, Vermaseren, Remmiddi, Blumlein,. . . )
n∑
i=1
1
i2
i∑
j=1
1
j
Integral representation:
=∫ 1
0
xn − 1
1− x
(
∫ x
0
∫ y
01
1−zdz
ydy − ζ2
)
dx, ζz :=
∞∑
i=1
1/iz
Asymptotic expansion:
=(
1
30n5−
1
6n3+
1
2n2−
1
n
)
ln(n)
− 1100n5 − 1
6n4 + 1336n3 − 1
4n2 − 1n+ 2ζ3 +O( ln(n)
n6 ).
limit computations numerical evaluation
(J. Ablinger, J. Blumlein, CS; J. Math. Phys. 2011 [arXiv:1105.6063 [math-ph]])RISC, J. Kepler University Linz
![Page 119: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/119.jpg)
Toolbox 3: Special function algorithms
◮ Generalized algorithms for generalized harmonic sums
N∑
k=1
2kk∑
i=1
2−i
i∑
j=1
Hj
j
i
k= −
21ζ2220
1
N+
1
8N2+
295
216N3−
1115
96N4+O(N−5)
+( 1
2N−
3
4N2+
19
12N3−
5
N4+O(N−5)
)ζ2
+ 2N( 3
2N+
3
2N2+
9
2N3+
39
2N4+O(N−5)
)ζ3
+( 1
N+
3
4N2−
157
36N3+
19
N4+O(N−5)
)(log(N) + γ)
+( 1
2N−
3
4N2+
19
12N3−
5
N4+O(N−5)
)(log(N) + γ)2)
[Ablinger, Blumlein, CS, J. Math. Phys. 54, 2013, arXiv:1302.0378 [math-ph]]
RISC, J. Kepler University Linz
![Page 120: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/120.jpg)
Toolbox 3: Special function algorithms
◮ Generalized algorithms for cyclotomic harmonic sums
N∑
k=1
k∑
j=1
j∑
i=1
1
1 + 2i
j2
(1 + 2k)2=(
− 3 +35ζ316
)
ζ2 −31ζ58
+1
N−
33
32N2+
17
16N3−
4795
4608N4+O(N−5)
+ log(2)(
6ζ2 −1
N+
9
8N2−
7
6N3+
209
192N4+O(N−5)
)
+(
−7
4−
7
16N+
7
16N2−
77
192N3+
21
64N4+O(N−5)
)
ζ3
+( 1
16N2−
1
8N3+
65
384N4+O(N−5)
)
(log(N) + γ)
[Ablinger, Blumlein, CS, J. Math. Phys. 52, 2011, arXiv:1302.0378 [math-ph]]
RISC, J. Kepler University Linz
![Page 121: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/121.jpg)
Toolbox 3: Special function algorithms
◮ Generalized algorithms for nested binomial sums
N∑
j=1
4jHj−1(2jj
)j2
= 7ζ3 +√π√N
{[
− 2
N+
5
12N2− 21
320N3− 223
10752N4+
671
49152N5
+11635
1441792N6− 1196757
136314880N7− 376193
50331648N8+
201980317
18253611008N9
+O(N−10)
]
ln(N)− 4
N+
5
18N2− 263
2400N3+
579
12544N4+ 10123
1105920N5
− 1705445
71368704N6− 27135463
11164188672N7+
197432563
7927234560N8+ 405757489
775778467840N9
+O(N−10)
}
Ablinger, Blumlein, CS, ACAT 2013, arXiv:1310.5645 [math-ph]
Ablinger, Blumlein, Raab, CS, J. Math. Phys. 55, 2014. arXiv:1407.1822 [hep-th]
RISC, J. Kepler University Linz
![Page 122: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/122.jpg)
Toolbox 3: Special function algorithms
Discovery of algebraic relations
mutliple Zeta-values∞∑
i=1
1
i3
i∑
j=1
(−1)j
j2
j∑
k=1
1
k
(Comprehensive literature: M.E. Hoffman, D. Zagier,P. Cartier, M. Petitot/H.N. Minh/C. Costermans,D.J. Broadhurst, D. Kreimer, M. Waldschmidt,D.M. Bradley, J. Vermaseren, J. Bumlein, etc.)
combining known relations of thesum and integral representations
RISC, J. Kepler University Linz
![Page 123: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/123.jpg)
Toolbox 3: Special function algorithms
Discovery of algebraic relations(J. Ablinger, J. Blumlein, CS)
mutliple Zeta-values∞∑
i=1
1
i3
i∑
j=1
(−1)j
j2
j∑
k=1
1
k
(Comprehensive literature: M.E. Hoffman, D. Zagier,P. Cartier, M. Petitot/H.N. Minh/C. Costermans,D.J. Broadhurst, D. Kreimer, M. Waldschmidt,D.M. Bradley, J. Vermaseren, J. Bumlein, etc.)
combining known relations of thesum and integral representations
cyclotomic Zeta-values
∞∑
i=1
1
i3
i∑
j=1
(−1)j
(2j + 1)2
j∑
k=1
1
k
RISC, J. Kepler University Linz
![Page 124: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/124.jpg)
Toolbox 3: Special function algorithms
Discovery of algebraic relations(J. Ablinger, J. Blumlein, CS)
mutliple Zeta-values∞∑
i=1
1
i3
i∑
j=1
(−1)j
j2
j∑
k=1
1
k
(Comprehensive literature: M.E. Hoffman, D. Zagier,P. Cartier, M. Petitot/H.N. Minh/C. Costermans,D.J. Broadhurst, D. Kreimer, M. Waldschmidt,D.M. Bradley, J. Vermaseren, J. Bumlein, etc.)
combining known relations of thesum and integral representations
cyclotomic Zeta-values
∞∑
i=1
1
i3
i∑
j=1
(−1)j
(2j + 1)2
j∑
k=1
1
k
generalized multipe Zeta-values
∞∑
i=1
1
i3
i∑
j=1
1
2jj2
j∑
k=1
1
k
RISC, J. Kepler University Linz
![Page 125: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/125.jpg)
Toolbox 3: Special function algorithms
The full machinery:
Toolbox 1 + Toolbox 2 + Toolbox 3
RISC, J. Kepler University Linz
![Page 126: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/126.jpg)
Toolbox 3: Special function algorithms
The problem
From: Robin Pemantle [University of Pennsylvania]
To: herb wilf; doron zeilberger
Herb, Doron,
I have a sum that, when I evaluate numerically, looks suspiciously
like it comes out to exactly 1.
Is there a way I can automatically decide this?
The sum may be written in many ways, but one is:
∞∑
n,k=1
Hk(Hn+1 − 1)
kn(n+ 1)(k + n); Hk :=
k∑
i=1
1
i
[Arose in the analysis of the simplex algorithm on the Klee-Minty cube(J. Balogh, R. Pemantle)]
RISC, J. Kepler University Linz
![Page 127: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/127.jpg)
Toolbox 3: Special function algorithms
The full machinery:In[1]:= << Sigma.m
Sigma by Carsten Schneider c© RISC-Linz
In[2]:= << EvaluateMultiSums.m
EvaluateMultiSums by Carsten Schneider c© RISC-Linz
In[4]:= EvaluateMultiSum[∞∑
n=1
∞∑
k=1
Hk(Hn+1 − 1)
kn(n + 1)(k + n)]
RISC, J. Kepler University Linz
![Page 128: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/128.jpg)
Toolbox 3: Special function algorithms
The full machinery:In[1]:= << Sigma.m
Sigma by Carsten Schneider c© RISC-Linz
In[2]:= << EvaluateMultiSums.m
EvaluateMultiSums by Carsten Schneider c© RISC-Linz
In[4]:= EvaluateMultiSum[∞∑
n=1
∞∑
k=1
Hk(Hn+1 − 1)
kn(n + 1)(k + n)]
Out[4]= 3
∞∑
i=1
i∑
j=1
j∑
k=1
k∑
l=1
1
l
k
j
i2− 2
∞∑
j=1
j∑
k=1
k∑
l=1
1
l
k
j3+
1
3
(
3
∞∑
j=1
j∑
k=1
k∑
l=1
1
l
k2
j2− 3
∞∑
k=1
k∑
l=1
1
l
k4
)
−
2
∞∑
k=1
k∑
l=1
1
l3
k2+(
∞∑
l=1
1
l2
)(
−∞∑
k=1
k∑
l=1
1
l
k2+
∞∑
l=1
1
l3− 1)
+ z2(
∞∑
k=1
k∑
l=1
1
l
k2− 1)
+∞∑
l=1
1
l5
RISC, J. Kepler University Linz
![Page 129: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/129.jpg)
Toolbox 3: Special function algorithms
The full machinery:In[1]:= << Sigma.m
Sigma by Carsten Schneider c© RISC-Linz
In[2]:= << EvaluateMultiSums.m
EvaluateMultiSums by Carsten Schneider c© RISC-Linz
In[3]:= << HarmonicSums.m
HarmonicSums by Jakob Ablinger c© RISC-Linz
In[4]:= EvaluateMultiSum[∞∑
n=1
∞∑
k=1
Hk(Hn+1 − 1)
kn(n + 1)(k + n)]
RISC, J. Kepler University Linz
![Page 130: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/130.jpg)
Toolbox 3: Special function algorithms
The full machinery:In[1]:= << Sigma.m
Sigma by Carsten Schneider c© RISC-Linz
In[2]:= << EvaluateMultiSums.m
EvaluateMultiSums by Carsten Schneider c© RISC-Linz
In[3]:= << HarmonicSums.m
HarmonicSums by Jakob Ablinger c© RISC-Linz
In[4]:= EvaluateMultiSum[∞∑
n=1
∞∑
k=1
Hk(Hn+1 − 1)
kn(n + 1)(k + n)]
Out[4]= −4ζ2 − 2ζ3 + 4ζ2ζ3 + 2ζ5
RISC, J. Kepler University Linz
![Page 131: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/131.jpg)
Toolbox 3: Special function algorithms
The full machinery:In[1]:= << Sigma.m
Sigma by Carsten Schneider c© RISC-Linz
In[2]:= << EvaluateMultiSums.m
EvaluateMultiSums by Carsten Schneider c© RISC-Linz
In[3]:= << HarmonicSums.m
HarmonicSums by Jakob Ablinger c© RISC-Linz
In[4]:= EvaluateMultiSum[∞∑
n=1
∞∑
k=1
Hk(Hn+1 − 1)
kn(n + 1)(k + n)]
Out[4]= −4ζ2 − 2ζ3 + 4ζ2ζ3 + 2ζ5
In[5]:= EvaluateMultiSum[∞∑
n=1
∞∑
k=1
H2k(Hn+1 − 1)2k (k + n) n
]
RISC, J. Kepler University Linz
![Page 132: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/132.jpg)
Toolbox 3: Special function algorithms
The full machinery:In[1]:= << Sigma.m
Sigma by Carsten Schneider c© RISC-Linz
In[2]:= << EvaluateMultiSums.m
EvaluateMultiSums by Carsten Schneider c© RISC-Linz
In[3]:= << HarmonicSums.m
HarmonicSums by Jakob Ablinger c© RISC-Linz
In[4]:= EvaluateMultiSum[∞∑
n=1
∞∑
k=1
Hk(Hn+1 − 1)
kn(n + 1)(k + n)]
Out[4]= −4ζ2 − 2ζ3 + 4ζ2ζ3 + 2ζ5
In[5]:= EvaluateMultiSum[∞∑
n=1
∞∑
k=1
H2k(Hn+1 − 1)2k (k + n) n
]
Out[5]= −10ζ3 + ζ22(58ζ3
5− 29
5
)
− 10ζ5 + ζ2(−ζ3 + 13ζ5 − 4) +457ζ7
8
RISC, J. Kepler University Linz
![Page 133: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/133.jpg)
Toolbox 3: Special function algorithms
The full machinery:In[1]:= << Sigma.m
Sigma by Carsten Schneider c© RISC-Linz
In[2]:= << EvaluateMultiSums.m
EvaluateMultiSums by Carsten Schneider c© RISC-Linz
In[3]:= << HarmonicSums.m
HarmonicSums by Jakob Ablinger c© RISC-Linz
In[4]:= EvaluateMultiSum[∞∑
n=1
∞∑
k=1
Hk(Hn+1 − 1)
kn(n + 1)(k + n)]
Out[4]= −4ζ2 − 2ζ3 + 4ζ2ζ3 + 2ζ5
In[5]:= EvaluateMultiSum[∞∑
k=1
∞∑
j=1
Hk(Hn+1 − 1)
k (k + n)2n2]
RISC, J. Kepler University Linz
![Page 134: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/134.jpg)
Toolbox 3: Special function algorithms
The full machinery:In[1]:= << Sigma.m
Sigma by Carsten Schneider c© RISC-Linz
In[2]:= << EvaluateMultiSums.m
EvaluateMultiSums by Carsten Schneider c© RISC-Linz
In[3]:= << HarmonicSums.m
HarmonicSums by Jakob Ablinger c© RISC-Linz
In[4]:= EvaluateMultiSum[∞∑
n=1
∞∑
k=1
Hk(Hn+1 − 1)
kn(n + 1)(k + n)]
Out[4]= −4ζ2 − 2ζ3 + 4ζ2ζ3 + 2ζ5
In[5]:= EvaluateMultiSum[∞∑
k=1
∞∑
j=1
Hk(Hn+1 − 1)
k (k + n)2n2]
Out[5]= 2ζ3 + ζ22(17ζ3
10+
17
10
)
+ ζ2(2ζ3 − 3ζ5 − 4)− 9ζ5
2+
3ζ7
16
RISC, J. Kepler University Linz
![Page 135: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/135.jpg)
Toolbox 3: Special function algorithms
The full machinery:In[1]:= << Sigma.m
Sigma by Carsten Schneider c© RISC-Linz
In[2]:= << EvaluateMultiSums.m
EvaluateMultiSums by Carsten Schneider c© RISC-Linz
In[3]:= << HarmonicSums.m
HarmonicSums by Jakob Ablinger c© RISC-Linz
In[4]:= EvaluateMultiSum[∞∑
n=1
∞∑
k=1
Hk(Hn+1 − 1)
kn(n + 1)(k + n)]
Out[4]= −4ζ2 − 2ζ3 + 4ζ2ζ3 + 2ζ5
In[5]:= EvaluateMultiSum[∞∑
n=1
∞∑
k=1
∞∑
l=1
HkHnHn+l+k
k (k + n) (k + n + l + 1)2]
RISC, J. Kepler University Linz
![Page 136: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/136.jpg)
Toolbox 3: Special function algorithms
The full machinery:In[1]:= << Sigma.m
Sigma by Carsten Schneider c© RISC-Linz
In[2]:= << EvaluateMultiSums.m
EvaluateMultiSums by Carsten Schneider c© RISC-Linz
In[3]:= << HarmonicSums.m
HarmonicSums by Jakob Ablinger c© RISC-Linz
In[4]:= EvaluateMultiSum[∞∑
n=1
∞∑
k=1
Hk(Hn+1 − 1)
kn(n + 1)(k + n)]
Out[4]= −4ζ2 − 2ζ3 + 4ζ2ζ3 + 2ζ5
In[5]:= EvaluateMultiSum[∞∑
n=1
∞∑
k=1
∞∑
l=1
HkHnHn+l+k
k (k + n) (k + n + l + 1)2]
Out[5]= 3ζ23 − 15ζ5
2+ ζ2(9ζ5 − 6ζ3) +
149ζ7
16+
114
35ζ32
RISC, J. Kepler University Linz
![Page 137: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/137.jpg)
Example 1: Unfair permutations
Example 1: Unfair permutations
joint work with H. Prodinger, S. Wagner
RISC, J. Kepler University Linz
![Page 138: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/138.jpg)
Example 1: Unfair permutations
◮ We are given n players.
RISC, J. Kepler University Linz
![Page 139: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/139.jpg)
Example 1: Unfair permutations
◮ We are given n players.
◮ Player i: chooses randomly a number (all different)
RISC, J. Kepler University Linz
![Page 140: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/140.jpg)
Example 1: Unfair permutations
◮ We are given n players.
◮ Player i: chooses randomly a number (all different)
◮ The player with the highest number gets n dices
RISC, J. Kepler University Linz
![Page 141: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/141.jpg)
Example 1: Unfair permutations
◮ We are given n players.
◮ Player i: chooses randomly a number (all different)
◮ The player with the highest number gets n dicesThe player with the second highest number gets n− 1 dices.
RISC, J. Kepler University Linz
![Page 142: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/142.jpg)
Example 1: Unfair permutations
◮ We are given n players.
◮ Player i: chooses randomly a number (all different)
◮ The player with the highest number gets n dicesThe player with the second highest number gets n− 1 dices.
...
The player with the lowest number (looser) gets 1 dice.
RISC, J. Kepler University Linz
![Page 143: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/143.jpg)
Example 1: Unfair permutations
◮ We are given n players.
◮ Player i: chooses randomly a number (all different)
◮ The player with the highest number gets n dicesThe player with the second highest number gets n− 1 dices.
...
The player with the lowest number (looser) gets 1 dice.
◮ We get a random permutation
playerdices
(1 2 3 . . . na1 a2 a3 . . . an
)
∈ Sn
RISC, J. Kepler University Linz
![Page 144: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/144.jpg)
Example 1: Unfair permutations
◮ We are given n players.
◮ Player i: chooses randomly i numbers and takes the largest (best) one
◮ The player with the highest number gets n dicesThe player with the second highest number gets n− 1 dices.
...
The player with the lowest number (looser) gets 1 dice.
◮ We get an unfair permutation
playerdices
(1 2 3 . . . na1 a2 a3 . . . an
)
∈ Sn
RISC, J. Kepler University Linz
![Page 145: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/145.jpg)
Example 1: Unfair permutations
◮ We are given n players.
◮ Player i: chooses randomly i numbers and takes the largest (best) one
◮ The player with the highest number gets n dicesThe player with the second highest number gets n− 1 dices.
...
The player with the lowest number (looser) gets 1 dice.
◮ We get an unfair permutation
playerdices
(1 2 3 . . . na1 a2 a3 . . . an
)
∈ Sn
anti-inversion:
i < j and ai < ajm
i < j and j beats iRISC, J. Kepler University Linz
![Page 146: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/146.jpg)
Example 1: Unfair permutations
◮ We are given n players.
◮ Player i: chooses randomly i numbers and takes the largest (best) one
◮ The player with the highest number gets n dicesThe player with the second highest number gets n− 1 dices.
...
The player with the lowest number (looser) gets 1 dice.
◮ We get an unfair permutation
playerdices
(1 2 3 . . . na1 a2 a3 . . . an
)
∈ Sn
anti-inversion:
i < j and ai < ajm
i < j and j beats i
probability:
j
i+ j
RISC, J. Kepler University Linz
![Page 147: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/147.jpg)
Example 1: Unfair permutations
◮ We are given n players.
◮ Player i: chooses randomly i numbers and takes the largest (best) one
◮ The player with the highest number gets n dicesThe player with the second highest number gets n− 1 dices.
...
The player with the lowest number (looser) gets 1 dice.
◮ We get an unfair permutation
playerdices
(1 2 3 . . . na1 a2 a3 . . . an
)
∈ Sn
anti-inversion:
i < j and ai < ajm
i < j and j beats i
probability:
j
i+ j
expected numberof anti-inversions:
∑
1≤i<j≤n
j
i+ j
RISC, J. Kepler University Linz
![Page 148: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/148.jpg)
Example 1: Unfair permutations
Theorem (Prodinger, Wagner).
An = no. of anti-inversions of a random unfair permutation of length n.
Then the mean of An is∑
1≤i<j≤n
j
i+ j
RISC, J. Kepler University Linz
![Page 149: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/149.jpg)
Example 1: Unfair permutations
Theorem (Prodinger, Wagner).
An = no. of anti-inversions of a random unfair permutation of length n.
Then the mean of An is∑
1≤i<j≤n
j
i+ j=
1
16
(− 8n2 − 8n− 1
)Hn +
1
8(2n + 1)2H2n −
5n
8
RISC, J. Kepler University Linz
![Page 150: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/150.jpg)
Example 1: Unfair permutations
Theorem (Prodinger, Wagner).
An = no. of anti-inversions of a random unfair permutation of length n.
Then the mean of An is∑
1≤i<j≤n
j
i+ j=
1
16
(− 8n2 − 8n− 1
)Hn +
1
8(2n + 1)2H2n −
5n
8
= 0.3465735903n2 − 0.4034264097n +O(log n)
fair case = 0.25n2 − 0.25n
RISC, J. Kepler University Linz
![Page 151: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/151.jpg)
Example 1: Unfair permutations
Theorem (Prodinger, Wagner).
An = no. of anti-inversions of a random unfair permutation of length n.
Then the mean of An is∑
1≤i<j≤n
j
i+ j=
1
16
(− 8n2 − 8n− 1
)Hn +
1
8(2n + 1)2H2n −
5n
8
The variance of An is
2∑
1≤i<j<k≤n
kj
(i+ j)(i + j + k)+ 2
∑
1≤i<j<k≤n
k
i+ j + k
+ 2∑
1≤i<j<k≤n
kj
(i+ j)(i + j + k)+ 2
∑
1≤i<j<k≤n
kj
(i+ k)(i + j + k)
− 2∑
1≤i<j<k≤n
j
i+ j·
k
j + k− 2
∑
1≤i<j<k≤n
k
i+ k·
k
j + k
− 2∑
1≤i<j<k≤n
j
i+ j·
k
i+ k−
∑
1≤i<j≤n
j2
(i+ j)2+
∑
1≤i<j≤n
j
i+ j
RISC, J. Kepler University Linz
![Page 152: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/152.jpg)
Example 1: Unfair permutations
Theorem (Prodinger, Wagner).
An = no. of anti-inversions of a random unfair permutation of length n.
Then the mean of An is∑
1≤i<j≤n
j
i+ j=
1
16
(− 8n2 − 8n− 1
)Hn +
1
8(2n + 1)2H2n −
5n
8
The variance of An is
2∑
1≤i<j<k≤n
kj
(i+ j)(i + j + k)+ 2
∑
1≤i<j<k≤n
k
i+ j + k
+ 2∑
1≤i<j<k≤n
kj
(i+ j)(i + j + k)+ 2
∑
1≤i<j<k≤n
kj
(i+ k)(i + j + k)
− 2∑
1≤i<j<k≤n
j
i+ j·
k
j + k− 2
∑
1≤i<j<k≤n
k
i+ k·
k
j + k
− 2∑
1≤i<j<k≤n
j
i+ j·
k
i+ k−
∑
1≤i<j≤n
j2
(i+ j)2+
∑
1≤i<j≤n
j
i+ j
RISC, J. Kepler University Linz
![Page 153: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/153.jpg)
Example 1: Unfair permutations
n∑
k=3
k−1∑
j=2
j−1∑
i=1
jk
(i+ j)(j + k)
RISC, J. Kepler University Linz
![Page 154: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/154.jpg)
Example 1: Unfair permutations
n∑
k=3
k−1∑
j=2
j−1∑
i=1
jk
(i+ j)(j + k)
RISC, J. Kepler University Linz
![Page 155: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/155.jpg)
Example 1: Unfair permutations
n∑
k=3
k−1∑
j=2
j−1∑
i=1
jk
(i+ j)(j + k)
|| summation spiral
1
j + kjk
j∑
r=1
1
−1 + 2r−
jkHj
2(j + k)−
k
2(j + k)
RISC, J. Kepler University Linz
![Page 156: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/156.jpg)
Example 1: Unfair permutations
n∑
k=3
k−1∑
j=2
[ 1
j + kjk
j∑
r=1
1
−1 + 2r−
jkHj
2(j + k)−
k
2(j + k)
]
RISC, J. Kepler University Linz
![Page 157: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/157.jpg)
Example 1: Unfair permutations
n∑
k=3
k−1∑
j=2
[ 1
j + kjk
j∑
r=1
1
−1 + 2r−
jkHj
2(j + k)−
k
2(j + k)
]
RISC, J. Kepler University Linz
![Page 158: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/158.jpg)
Example 1: Unfair permutations
n∑
k=3
k−1∑
j=2
[ 1
j + kjk
j∑
r=1
1
−1 + 2r−
jkHj
2(j + k)−
k
2(j + k)
]
|| summation spiral
− k2k∑
s=1
s∑
r=1
1
−1 + 2r
s+((k − 1)k + k2Hk
)k∑
r=1
1
−1 + 2r
−1
4k2H2
k −1
4k2H
(2)k −
1
4k(2k − 3)Hk +
1
4
RISC, J. Kepler University Linz
![Page 159: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/159.jpg)
Example 1: Unfair permutations
n∑
k=3
[
− k2k∑
s=1
s∑
r=1
1
−1 + 2r
s+((k − 1)k + k2Hk
)k∑
r=1
1
−1 + 2r
−1
4k2H2
k −1
4k2H
(2)k −
1
4k(2k − 3)Hk +
1
4
]
RISC, J. Kepler University Linz
![Page 160: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/160.jpg)
Example 1: Unfair permutations
n∑
k=3
[
− k2k∑
s=1
s∑
r=1
1
−1 + 2r
s+((k − 1)k + k2Hk
)k∑
r=1
1
−1 + 2r
−1
4k2H2
k −1
4k2H
(2)k −
1
4k(2k − 3)Hk +
1
4
]
|| summation spiral
n(n+ 1)(2n + 1)[
−1
6
n∑
s=1
s∑
r=1
1
−1 + 2r
s−
1
12
)Hn −
1
24H2
n
+(1
6
n∑
r=1
1
−1 + 2r−
1
24H(2)
n +1
6
n∑
r=1
1
−1 + 2r
]
−1
8(2n + 1)2
n∑
r=1
1
−1 + 2r+
1
12(n+ 1)(4n + 1)Hn +
7n
24
RISC, J. Kepler University Linz
![Page 161: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/161.jpg)
Example 1: Unfair permutations
n∑
k=3
k−1∑
j=2
j−1∑
i=1
jk
(i+ j)(j + k)
||
n(n+ 1)(2n + 1)[
−1
6
n∑
s=1
s∑
r=1
1
−1 + 2r
s−
1
12
)Hn −
1
24H2
n
+(1
6
n∑
r=1
1
−1 + 2r−
1
24H(2)
n +1
6
n∑
r=1
1
−1 + 2r
]
−1
8(2n + 1)2
n∑
r=1
1
−1 + 2r+
1
12(n+ 1)(4n + 1)Hn +
7n
24
RISC, J. Kepler University Linz
![Page 162: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/162.jpg)
Example 1: Unfair permutations
Theorem (Prodinger, Wagner).
An = no. of anti-inversions of a random unfair permutation of length n.
Then the mean of An is∑
1≤i<j≤n
j
i+ j=
1
16
(− 8n2 − 8n− 1
)Hn +
1
8(2n + 1)2H2n −
5n
8
The variance of An is
2∑
1≤i<j<k≤n
kj
(i+ j)(i + j + k)+ 2
∑
1≤i<j<k≤n
k
i+ j + k
+ 2∑
1≤i<j<k≤n
kj
(i+ j)(i + j + k)+ 2
∑
1≤i<j<k≤n
kj
(i+ k)(i + j + k)
− 2∑
1≤i<j<k≤n
j
i+ j·
k
j + k− 2
∑
1≤i<j<k≤n
k
i+ k·
k
j + k
− 2∑
1≤i<j<k≤n
j
i+ j·
k
i+ k−
∑
1≤i<j≤n
j2
(i+ j)2+
∑
1≤i<j≤n
j
i+ j
RISC, J. Kepler University Linz
![Page 163: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/163.jpg)
Example 1: Unfair permutations
Theorem (Prodinger, Wagner, CS).
An = no. of anti-inversions of a random unfair permutation of length n.
Then the mean of An is∑
1≤i<j≤n
j
i+ j=
1
16
(− 8n2 − 8n− 1
)Hn +
1
8(2n + 1)2H2n −
5n
8
The variance of An is
n(29 + 126n + 72n2)
216+
35 + 108n + 81n2 − 27n3
162Hn
+−3− 16n − 10n2 + 8n3
12H2n +
−16 + 27n − 54n3
108H3n
+n(1 + 3n + 2n2)
6
(
3H(2)2n − 2H(2)
n + 4∑
1≤i≤2n
(−1)iHi
i
)
+8
27
n∑
i=1
1
3i− 2+
(−1)nn
4
( n∑
i=1
(−1)i
i−
3n∑
i=1
(−1)i
i
)
,
RISC, J. Kepler University Linz
![Page 164: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/164.jpg)
Example 2: Super-congruences
Example 2: Super-congruences
(S. Ahlgren, E. Mortenson, R. Osburn, Sigma)
RISC, J. Kepler University Linz
![Page 165: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/165.jpg)
Example 2: Super-congruences
Sigma’s contribution to harmonic number congruences◮ S. Ahlgren (2001):
p−12∑
j=0
(p−12
j
)2(j + p−1
2
j
)
(Hj+ p−1
2−H p−1
2) ≡ 0 mod p
RISC, J. Kepler University Linz
![Page 166: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/166.jpg)
Example 2: Super-congruences
Sigma’s contribution to harmonic number congruences◮ S. Ahlgren (2001):
p−12∑
j=0
(p−12
j
)2(j + p−1
2
j
)
(Hj+ p−1
2−H p−1
2) ≡ 0 mod p
◮ E. Mortenson (2003):p−12∑
j=0
(p−12
j
)2(j + p−1
2
j
)
(1 + 3jHj+ p−1
2− 3jHj) ≡ 0 mod p
p−12∑
j=0
(p−12
j
)(j + p−1
2
j
)(1 + 2jH
j+ p−12− 2jHj) ≡ 0 mod p
RISC, J. Kepler University Linz
![Page 167: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/167.jpg)
Example 2: Super-congruences
Sigma’s contribution to harmonic number congruences◮ S. Ahlgren (2001):
p−12∑
j=0
(p−12
j
)2(j + p−1
2
j
)
(Hj+ p−1
2−H p−1
2) ≡ 0 mod p
◮ E. Mortenson (2003):p−12∑
j=0
(p−12
j
)2(j + p−1
2
j
)
(1 + 3jHj+ p−1
2− 3jHj) ≡ 0 mod p
p−12∑
j=0
(p−12
j
)(j + p−1
2
j
)(1 + 2jH
j+ p−12− 2jHj) ≡ 0 mod p
◮ R. Osburn:
p2 E2(p) + pE1(p) + p0E0(p) ≡ (−1)p−12 mod p3
RISC, J. Kepler University Linz
![Page 168: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/168.jpg)
Example 2: Super-congruences
For a prime p > 2,
p2E2(p)
+pE1(p)
+p0E0(p) ≡ (−1)p−12 mod p3
RISC, J. Kepler University Linz
![Page 169: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/169.jpg)
Example 2: Super-congruences
For a prime p > 2,
p2E2(p)
+pE1(p)
+p0
p−12∑
j=0
(2j
j
)2
16−j ≡ (−1)p−12 mod p3
RISC, J. Kepler University Linz
![Page 170: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/170.jpg)
Example 2: Super-congruences
For a prime p > 2,
p2E2(p)
+p[
p−12∑
j=0
(−1)j(p−1
2
j
)(j + p−1
2
j
)(1 + j
(+H
j+ p−12
+H−j+ p−12− 2Hj
))]
+p0
p−12∑
j=0
(2j
j
)2
16−j ≡ (−1)p−12 mod p3
RISC, J. Kepler University Linz
![Page 171: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/171.jpg)
Example 2: Super-congruences
For a prime p > 2,
p2[
p−32∑
j=1
( (−1)j( p−1
2j
)(j+ p−12
j
)
+
p−12∑
j=0
(−1)j(p−1
2
j
)(j + p−1
2
j
)(1 + 4j
(H
j+ p−12−Hj
)
+ j2(2(H
j+ p−12−Hj
)2+H
(2)j −H
(2)
j+ p−12
))]
+p[
p−12∑
j=0
(−1)j(p−1
2
j
)(j + p−1
2
j
)(1 + j
(+H
j+ p−12
+H−j+ p−12− 2Hj
))]
+p0
p−12∑
j=0
(2j
j
)2
16−j ≡ (−1)p−12 mod p3
RISC, J. Kepler University Linz
![Page 172: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/172.jpg)
Example 2: Super-congruences
For a prime p > 2,
p2[
p−32∑
j=1
( (−1)j
( p−12j
)(j+ p−12
j
)
+
p−12∑
j=0
(−1)j(p−1
2
j
)(j + p−1
2
j
)(1 + 4j
(H
j+ p−12−Hj
)
+j2(2(H
j+ p−12−Hj
)2+H
(2)j −H
(2)
j+ p−12
)) ]
+p[
p−12∑
j=0
(−1)j(p−1
2
j
)(j + p−1
2
j
)(1 + j
(+H
j+ p−12
+H−j+ p−12− 2Hj
))]
+p0
p−12∑
j=0
(2j
j
)2
16−j ≡ (−1)p−12 mod p3
RISC, J. Kepler University Linz
![Page 173: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/173.jpg)
Example 2: Super-congruences
p−32∑
j=1
( (−1)j
( p−12j
)(j+ p−12
j
)
+
p−12∑
j=0
(−1)j(p−1
2
j
)(j + p−1
2
j
)(1 + 4j
(H
j+ p−12−Hj
)
+ j2(2(H
j+ p−12−Hj
)2+H
(2)j −H
(2)
j+ p−12
))
RISC, J. Kepler University Linz
![Page 174: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/174.jpg)
Example 2: Super-congruences
n− 1∑
j=1
( (−1)j(nj
)(j+nj
)
+
n∑
j=0
(−1)j(n
j
)(j + n
j
)(1 + 4j
(Hj+n −Hj
)
+ j2(2(Hj+n −Hj
)2+H
(2)j −H
(2)j+n
))
RISC, J. Kepler University Linz
![Page 175: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/175.jpg)
Example 2: Super-congruences
n− 1∑
j=1
( (−1)j(nj
)(j+nj
)
+
n∑
j=0
(−1)j(n
j
)(j + n
j
)(1 + 4j
(Hj+n −Hj
)
+ j2(2(Hj+n −Hj
)2+H
(2)j −H
(2)j+n
))
|| summation spiral
(−1)n((n + 1)(2n + 1)−
(2n
n
)
)
RISC, J. Kepler University Linz
![Page 176: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/176.jpg)
Example 2: Super-congruences
p−32∑
j=1
( (−1)j
( p−12j
)(j+ p−12
j
)
+
p−12∑
j=0
(−1)j(p−1
2
j
)(j + p−1
2
j
)(1 + 4j
(H
j+ p−12−Hj
)
+ j2(2(H
j+ p−12−Hj
)2+H
(2)j −H
(2)
j+ p−12
))
||
(−1)p−12 ((p−1
2 + 1)p−
(p− 1p−12
)
)
RISC, J. Kepler University Linz
![Page 177: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/177.jpg)
Example 2: Super-congruences
For a prime p > 2,
p2[
p−32∑
j=1
( (−1)j
( p−12j
)(j+ p−12
j
)
+
p−12∑
j=0
(−1)j(p−1
2
j
)(j + p−1
2
j
)(1 + 4j
(H
j+ p−12−Hj
)
+j2(2(H
j+ p−12−Hj
)2+H
(2)j −H
(2)
j+ p−12
)) ]
+p[
p−12∑
j=0
(−1)j(p−1
2
j
)(j + p−1
2
j
)(1 + j
(+H
j+ p−12
+H−j+ p−12− 2Hj
))]
+p0
p−12∑
j=0
(2j
j
)2
16−j ≡ (−1)p−12 mod p3
RISC, J. Kepler University Linz
![Page 178: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/178.jpg)
Example 2: Super-congruences
For a prime p > 2,
p2[
(−1)p−12
(
(p−12 + 1)p −
(p− 1p−12
))
]
+p[
p−12∑
j=0
(−1)j(p−1
2
j
)(j + p−1
2
j
)(1 + j
(+H
j+ p−12
+H−j+ p−12− 2Hj
))]
+p0
p−12∑
j=0
(2j
j
)2
16−j ≡ (−1)p−12 mod p3
RISC, J. Kepler University Linz
![Page 179: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/179.jpg)
Example 2: Super-congruences
For a prime p > 2,
p2[
− (−1)p−12
(p− 1p−12
)
]
+p[
p−12∑
j=0
(−1)j(p−1
2
j
)(j + p−1
2
j
)(1 + j
(+H
j+ p−12
+H−j+ p−12− 2Hj
))]
+p0
p−12∑
j=0
(2j
j
)2
16−j ≡ (−1)p−12 mod p3
RISC, J. Kepler University Linz
![Page 180: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/180.jpg)
Example 2: Super-congruences
For a prime p > 2,
p2[
− (−1)p−12
(p− 1p−12
)
]
+p[
p−12∑
j=0
(−1)j(p−1
2
j
)(j + p−1
2
j
)(1 + j
(+H
j+ p−12
+H−j+ p−12− 2Hj
)) ]
+p0
p−12∑
j=0
(2j
j
)2
16−j ≡ (−1)p−12 mod p3
RISC, J. Kepler University Linz
![Page 181: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/181.jpg)
Example 2: Super-congruences
p−12∑
j=0
(−1)j(p−1
2
j
)(j + p−1
2
j
)(1 + j
(− 2Hj +H
j+ p−12
+H−j+ p−12
))
RISC, J. Kepler University Linz
![Page 182: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/182.jpg)
Example 2: Super-congruences
n∑
j=0
(−1)j(n
j
)(j + n
j
)(1 + j
(− 2Hj +Hj+n +H−j+n
))
RISC, J. Kepler University Linz
![Page 183: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/183.jpg)
Example 2: Super-congruences
n∑
j=0
(−1)j(n
j
)(j + n
j
)(1 + j
(− 2Hj +Hj+n +H−j+n
))
|| summation spiral
−3
2(−1)nn(n+ 1)
n∑
j=1
(2jj
)
j+ (−1)n(2n + 1)
(2n
n
)
RISC, J. Kepler University Linz
![Page 184: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/184.jpg)
Example 2: Super-congruences
p−12∑
j=0
(−1)j(p−1
2
j
)(j + p−1
2
j
)(1 + j
(− 2Hj +H
j+ p−12
+H−j+ p−12
))
||
−3
2(−1)
p−12
(p2
4−
1
4
)p−12∑
j=1
(2jj
)
j+ (−1)
p−12 p
(p− 1p−12
)
RISC, J. Kepler University Linz
![Page 185: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/185.jpg)
Example 2: Super-congruences
For a prime p > 2,
p2[
− (−1)p−12
(p− 1p−12
)
]
+p[
p−12∑
j=0
(−1)j(p−1
2
j
)(j + p−1
2
j
)(1 + j
(+H
j+ p−12
+H−j+ p−12− 2Hj
)) ]
+p0
p−12∑
j=0
(2j
j
)2
16−j ≡ (−1)p−12 mod p3
RISC, J. Kepler University Linz
![Page 186: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/186.jpg)
Example 2: Super-congruences
For a prime p > 2,
p2[
− (−1)p−12
(p− 1p−12
)
]
+p[
−3
2(−1)
p−12
(p2
4−
1
4
)p−12∑
j=1
(2jj
)
j+ (−1)
p−12 p
(p− 1p−12
)]
+p0
p−12∑
j=0
(2j
j
)2
16−j ≡ (−1)p−12 mod p3
RISC, J. Kepler University Linz
![Page 187: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/187.jpg)
Example 2: Super-congruences
For a prime p > 2,
p2[
0]
+p[
−3
2(−1)
p−12
(p2
4−
1
4
)p−12∑
j=1
(2jj
)
j
]
+p0
p−12∑
j=0
(2j
j
)2
16−j ≡ (−1)p−12 mod p3
RISC, J. Kepler University Linz
![Page 188: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/188.jpg)
Example 2: Super-congruences
For a prime p > 2,
p2[
0]
+p[3
8(−1)
p−12
p−12∑
j=1
(2jj
)
j
]
+p0
p−12∑
j=0
(2j
j
)2
16−j ≡ (−1)p−12 mod p3
RISC, J. Kepler University Linz
![Page 189: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/189.jpg)
Example 2: Super-congruences
For a prime p > 2,
p2[
p−32∑
j=1
( (−1)j( p−1
2j
)(j+ p−12
j
)
+
p−12∑
j=0
(−1)j(p−1
2
j
)(j + p−1
2
j
)(1 + 4j
(H
j+ p−12−Hj
)
+ j2(2(H
j+ p−12−Hj
)2+H
(2)j −H
(2)
j+ p−12
))]
+p[
p−12∑
j=0
(−1)j(p−1
2
j
)(j + p−1
2
j
)(1 + j
(+H
j+ p−12
+H−j+ p−12− 2Hj
))]
+p0
p−12∑
j=0
(2j
j
)2
16−j ≡ (−1)p−12 mod p3
RISC, J. Kepler University Linz
![Page 190: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/190.jpg)
Example 2: Super-congruences
Sigma’s contribution to harmonic number congruences◮ S. Ahlgren (2001):
p−12∑
j=0
(p−12
j
)2(j + p−1
2
j
)
(Hj+ p−1
2−H p−1
2) ≡ 0 mod p
◮ E. Mortenson (2003):p−12∑
j=0
(p−12
j
)2(j + p−1
2
j
)
(1 + 3jHj+ p−1
2− 3jHj) ≡ 0 mod p
p−12∑
j=0
(p−12
j
)(j + p−1
2
j
)(1 + 2jH
j+ p−12− 2jHj) ≡ 0 mod p
◮ R. Osburn/CS (2008):
p3
8(−1)
p−12
p−12∑
j=1
(2jj
)
j+
p−12∑
j=0
(2j
j
)2
16−j ≡ (−1)p−12 mod p3
RISC, J. Kepler University Linz
![Page 191: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/191.jpg)
Example 2: Super-congruences
Example 3: Feynman integrals
joint work with J. Ablinger, A. Behring, J. Blumlein, A. Hasselhuhn,A. de Freitas, C. Raab, M. Round, F. Wissbrock (RISC–DESY)
RISC, J. Kepler University Linz
![Page 192: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/192.jpg)
Example 3: Feynman diagrams
Evaluation of Feynman diagrams(long term project with J. Blumlein, Deutsches Elektronen–Synchrotron)
Behavior of particles
RISC, J. Kepler University Linz
![Page 193: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/193.jpg)
Example 3: Feynman diagrams
Evaluation of Feynman diagrams(long term project with J. Blumlein, Deutsches Elektronen–Synchrotron)
Behavior of particles
//
∫
Φ(n, ǫ, x)dx
Feynman integrals
RISC, J. Kepler University Linz
![Page 194: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/194.jpg)
Example 3: Feynman diagrams
Evaluation of Feynman diagrams(long term project with J. Blumlein, Deutsches Elektronen–Synchrotron)
Behavior of particles
//
∫
Φ(n, ǫ, x)dx
Feynman integrals
DESY
��
∑
f(n, ǫ, k)
multi sumsRISC, J. Kepler University Linz
![Page 195: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/195.jpg)
Example 3: Feynman diagrams
Evaluation of Feynman diagrams(long term project with J. Blumlein, Deutsches Elektronen–Synchrotron)
Behavior of particles
//
∫
Φ(n, ǫ, x)dx
Feynman integrals
DESY
��
simple sum expressions∑
f(n, ǫ, k)
multi sums
symbolic summationoo
RISC, J. Kepler University Linz
![Page 196: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/196.jpg)
Example 3: Feynman diagrams
Evaluation of Feynman diagrams(long term project with J. Blumlein, Deutsches Elektronen–Synchrotron)
Behavior of particles
//
∫
Φ(n, ǫ, x)dx
Feynman integrals
DESY
��
Evaluations required for theLHC experiment at CERN
simple sum expressions
processable by physicists
99
∑
f(n, ǫ, k)
multi sums
symbolic summationoo
RISC, J. Kepler University Linz
![Page 197: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/197.jpg)
Example 3: Feynman diagrams
= F−3(n)ε−3 + F−2(n)ε
−2 + F−1(n)ε−1 + F0(n) + . . .
RISC, J. Kepler University Linz
![Page 198: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/198.jpg)
Example 3: Feynman diagrams
= F−3(n)ε−3 + F−2(n)ε
−2 + F−1(n)ε−1 + F0(n) + . . .
Simplify ||
n−3∑
j=0
j∑
k=0
k∑
l=0
−j+n−3∑
q=0
−l+n−q−3∑
s=1
−l+n−q−s−3∑
r=0
(−1)−j+k−l+n−q−3×
×(j+1k+1)(
kl)(
n−1j+2)(
−j+n−3q )(−l+n−q−3
s ) (−l+n−q−s−3r )r!(−l+n−q−r−s−3)!(s−1)!
(−l+n−q−2)!(−j+n−1)(n−q−r−s−2)(q+s+1)[
4H−j+n−1 − 4H−j+n−2 − 2Hk
− (H−l+n−q−2 +H−l+n−q−r−s−3 − 2Hr+s)
+ 2Hs−1 − 2Hr+s
]
+ 3 further 6–fold sums
RISC, J. Kepler University Linz
![Page 199: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/199.jpg)
Example 3: Feynman diagrams
F0(N) =
7
12HN
4 +(17N + 5)HN
3
3N(N + 1)+( 35N2 − 2N − 5
2N2(N + 1)2+
13H(2)N
2+
5(−1)N
2N2
)
HN2
+(
− 4(13N + 5)
N2(N + 1)2+(4(−1)N (2N + 1)
N(N + 1)− 13
N
)
H(2)N
+(29
3− (−1)N
)
H(3)N
+(
2 + 2(−1)N)
S2,1(N) − 28S−2,1(N) +20(−1)N
N2(N + 1)
)
HN +( 3
4+ (−1)N
)
H(2)N
2
− 2(−1)NS−2(N)2 + S−3(N)( 2(3N − 5)
N(N + 1)+(
26 + 4(−1)N)
HN +4(−1)N
N + 1
)
+( (−1)N (5 − 3N)
2N2(N + 1)− 5
2N2
)
H(2)N
+ S−2(N)(
10HN2 +
( 8(−1)N (2N + 1)
N(N + 1)
+4(3N − 1)
N(N + 1)
)
HN +8(−1)N (3N + 1)
N(N + 1)2+(
− 22 + 6(−1)N)
H(2)N
− 16
N(N + 1)
)
+( (−1)N (9N + 5)
N(N + 1)− 29
3N
)
H(3)N
+(19
2− 2(−1)N
)
H(4)N
+(
− 6 + 5(−1)N)
S−4(N)
+(
− 2(−1)N (9N + 5)
N(N + 1)− 2
N
)
S2,1(N) +(
20 + 2(−1)N)
S2,−2(N) +(
− 17 + 13(−1)N)
S3,1(N)
− 8(−1)N (2N + 1) + 4(9N + 1)
N(N + 1)S−2,1(N) −
(
24 + 4(−1)N)
S−3,1(N) +(
3− 5(−1)N)
S2,1,1(N)
+ 32S−2,1,1(N) +
(
3
2HN
2 − 3HN
N+
3
2(−1)NS−2(N)
)
ζ(2)
RISC, J. Kepler University Linz
![Page 200: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/200.jpg)
Example 3: Feynman diagrams
F0(N) =
7
12HN
4 +(17N + 5)HN
3
3N(N + 1)+( 35N2 − 2N − 5
2N2(N + 1)2+
13H(2)N
2+
5(−1)N
2N2
)
HN2
+(
− 4(13N + 5)
N2(N + 1)2+(4(−1)N (2N + 1)
N(N + 1)− 13
N
)
H(2)N
+(29
3− (−1)N
)
H(3)N
+(
2 + 2(−1)N)
S2,1(N) − 28S−2,1(N) +20(−1)N
N2(N + 1)
)
HN +( 3
4+ (−1)N
)
H(2)N
2
− 2(−1)NS−2(N)2 + S−3(N)( 2(3N − 5)
N(N + 1)+(
26 + 4(−1)N)
HN +4(−1)N
N + 1
)
+( (−1)N (5 − 3N)
2N2(N + 1)− 5
2N2
)
H(2)N
+ S−2(N)(
10HN2 +
( 8(−1)N (2N + 1)
N(N + 1)
+4(3N − 1)
N(N + 1)
)
HN +8(−1)N (3N + 1)
N(N + 1)2+(
− 22 + 6(−1)N)
H(2)N
− 16
N(N + 1)
)
+( (−1)N (9N + 5)
N(N + 1)− 29
3N
)
H(3)N
+(19
2− 2(−1)N
)
H(4)N
+(
− 6 + 5(−1)N)
S−4(N)
+(
− 2(−1)N (9N + 5)
N(N + 1)− 2
N
)
S2,1(N) +(
20 + 2(−1)N)
S2,−2(N) +(
− 17 + 13(−1)N)
S3,1(N)
− 8(−1)N (2N + 1) + 4(9N + 1)
N(N + 1)S−2,1(N) −
(
24 + 4(−1)N)
S−3,1(N) +(
3− 5(−1)N)
S2,1,1(N)
+ 32S−2,1,1(N) +
(
3
2HN
2 − 3HN
N+
3
2(−1)NS−2(N)
)
ζ(2)
HN =N∑
i=1
1
i
RISC, J. Kepler University Linz
![Page 201: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/201.jpg)
Example 3: Feynman diagrams
F0(N) =
7
12HN
4 +(17N + 5)HN
3
3N(N + 1)+( 35N2 − 2N − 5
2N2(N + 1)2+
13H(2)N
2+
5(−1)N
2N2
)
HN2
+(
− 4(13N + 5)
N2(N + 1)2+(4(−1)N (2N + 1)
N(N + 1)− 13
N
)
H(2)N
+(29
3− (−1)N
)
H(3)N
+(
2 + 2(−1)N)
S2,1(N) − 28S−2,1(N) +20(−1)N
N2(N + 1)
)
HN +( 3
4+ (−1)N
)
H(2)N
2
− 2(−1)NS−2(N)2 + S−3(N)( 2(3N − 5)
N(N + 1)+(
26 + 4(−1)N)
HN +4(−1)N
N + 1
)
+( (−1)N (5 − 3N)
2N2(N + 1)− 5
2N2
)
H(2)N
+ S−2(N)(
10HN2 +
( 8(−1)N (2N + 1)
N(N + 1)
+4(3N − 1)
N(N + 1)
)
HN +8(−1)N (3N + 1)
N(N + 1)2+(
− 22 + 6(−1)N)
H(2)N
− 16
N(N + 1)
)
+( (−1)N (9N + 5)
N(N + 1)− 29
3N
)
H(3)N
+(19
2− 2(−1)N
)
H(4)N
+(
− 6 + 5(−1)N)
S−4(N)
+(
− 2(−1)N (9N + 5)
N(N + 1)− 2
N
)
S2,1(N) +(
20 + 2(−1)N)
S2,−2(N) +(
− 17 + 13(−1)N)
S3,1(N)
− 8(−1)N (2N + 1) + 4(9N + 1)
N(N + 1)S−2,1(N) −
(
24 + 4(−1)N)
S−3,1(N) +(
3− 5(−1)N)
S2,1,1(N)
+ 32S−2,1,1(N) +
(
3
2HN
2 − 3HN
N+
3
2(−1)NS−2(N)
)
ζ(2)
HN =N∑
i=1
1
i
H(2)N =
N∑
i=1
1
i2
RISC, J. Kepler University Linz
![Page 202: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/202.jpg)
Example 3: Feynman diagrams
F0(N) =
7
12HN
4 +(17N + 5)HN
3
3N(N + 1)+( 35N2 − 2N − 5
2N2(N + 1)2+
13H(2)N
2+
5(−1)N
2N2
)
HN2
+(
− 4(13N + 5)
N2(N + 1)2+(4(−1)N (2N + 1)
N(N + 1)− 13
N
)
H(2)N
+(29
3− (−1)N
)
H(3)N
+(
2 + 2(−1)N)
S2,1(N) − 28S−2,1(N) +20(−1)N
N2(N + 1)
)
HN +( 3
4+ (−1)N
)
H(2)N
2
− 2(−1)NS−2(N)2 + S−3(N)( 2(3N − 5)
N(N + 1)+(
26 + 4(−1)N)
HN +4(−1)N
N + 1
)
+( (−1)N (5 − 3N)
2N2(N + 1)− 5
2N2
)
H(2)N
+ S−2(N)(
10HN2 +
( 8(−1)N (2N + 1)
N(N + 1)
+4(3N − 1)
N(N + 1)
)
HN +8(−1)N (3N + 1)
N(N + 1)2+(
− 22 + 6(−1)N)
H(2)N
− 16
N(N + 1)
)
+( (−1)N (9N + 5)
N(N + 1)− 29
3N
)
H(3)N
+(19
2− 2(−1)N
)
H(4)N
+(
− 6 + 5(−1)N)
S−4(N)
+(
− 2(−1)N (9N + 5)
N(N + 1)− 2
N
)
S2,1(N) +(
20 + 2(−1)N)
S2,−2(N) +(
− 17 + 13(−1)N)
S3,1(N)
− 8(−1)N (2N + 1) + 4(9N + 1)
N(N + 1)S−2,1(N) −
(
24 + 4(−1)N)
S−3,1(N) +(
3− 5(−1)N)
S2,1,1(N)
+ 32S−2,1,1(N) +
(
3
2HN
2 − 3HN
N+
3
2(−1)NS−2(N)
)
ζ(2)
HN =N∑
i=1
1
i
H(2)N =
N∑
i=1
1
i2
S−2,1,1(N) =N∑
i=1
(−1)ii∑
j=1
j∑
k=1
1
k
j
i2
RISC, J. Kepler University Linz
![Page 203: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/203.jpg)
Summary
Summarizing:
If you have
RISC, J. Kepler University Linz
![Page 204: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/204.jpg)
Summary
Summarizing:
If you have
unfair permutations/monster sums/. . .
RISC, J. Kepler University Linz
![Page 205: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/205.jpg)
Summary
Summarizing:
If you have
unfair permutations/monster sums/. . .
super congruences/identities/. . . ,
RISC, J. Kepler University Linz
![Page 206: SymbolicSummation for Combinatorialand …dmg.tuwien.ac.at/aofa15/slides/schneider.pdfq-hypergeometric telescoping. In: M. Ismail, M. Rahman (eds.) Special Functions, q-Series and](https://reader031.fdocuments.in/reader031/viewer/2022030505/5ab1e5587f8b9ad9788cd817/html5/thumbnails/206.jpg)
Summary
Summarizing:
If you have
unfair permutations/monster sums/. . .
super congruences/identities/. . . ,
give the presented machinery a try!
RISC, J. Kepler University Linz