The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe...
Transcript of The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe...
![Page 1: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/1.jpg)
The Two-phase Simplex Method
Two-phase simplex method1 Given an LP in standard from, first run phase I.
2 If phase I yields a basic feasible solution for the original LP, enter “phase II” (seeabove).
Possible outcomes of the two-phase simplex methodi Problem is infeasible (detected in phase I).
ii Problem is feasible but rows of A are linearly dependent (detected and corrected atthe end of phase I by eliminating redundant constraints.)
iii Optimal cost is �1 (detected in phase II).
iv Problem has optimal basic feasible solution (found in phase II).
Remark: (ii) is not an outcome but only an intermediate result leading to outcome (iii)or (iv).
112
![Page 2: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/2.jpg)
Big-M MethodAlternative idea: Combine the two phases into one by introducing sufficiently largepenalty costs for artificial variables.
This way, the LP
minPn
i=1 ci xis.t. A · x = b
x � 0
becomes:
minPn
i=1 ci xi + M ·Pm
j=1 yjs.t. A · x + Im · y = b
x , y � 0
Remark: If M is sufficiently large and the original program has a feasible solution, allartificial variables will be driven to zero by the simplex method.
113
![Page 3: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/3.jpg)
Big-M MethodAlternative idea: Combine the two phases into one by introducing sufficiently largepenalty costs for artificial variables.
This way, the LP
minPn
i=1 ci xis.t. A · x = b
x � 0
becomes:
minPn
i=1 ci xi + M ·Pm
j=1 yjs.t. A · x + Im · y = b
x , y � 0
Remark: If M is sufficiently large and the original program has a feasible solution, allartificial variables will be driven to zero by the simplex method.
113
![Page 4: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/4.jpg)
Big-M MethodAlternative idea: Combine the two phases into one by introducing sufficiently largepenalty costs for artificial variables.
This way, the LP
minPn
i=1 ci xis.t. A · x = b
x � 0
becomes:
minPn
i=1 ci xi + M ·Pm
j=1 yjs.t. A · x + Im · y = b
x , y � 0
Remark: If M is sufficiently large and the original program has a feasible solution, allartificial variables will be driven to zero by the simplex method.
113
![Page 5: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/5.jpg)
Big-M MethodAlternative idea: Combine the two phases into one by introducing sufficiently largepenalty costs for artificial variables.
This way, the LP
minPn
i=1 ci xis.t. A · x = b
x � 0
becomes:
minPn
i=1 ci xi + M ·Pm
j=1 yjs.t. A · x + Im · y = b
x , y � 0
Remark: If M is sufficiently large and the original program has a feasible solution, allartificial variables will be driven to zero by the simplex method.
113
![Page 6: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/6.jpg)
How to Choose M?
ObservationInitially, M only occurs in the zeroth row. As the zeroth row never becomes pivot row,this property is maintained while the simplex method is running.
All we need to have is an order on all values that can appear as reduced cost coefficients.
Order on cost coefficientsaM + b < c M + d :() (a < c) _ (a = c ^ b < d)
In particular, �aM + b < 0 < aM + b for any positive a and arbitrary b, and we candecide whether a cost coefficient is negative or not.
! There is no need to give M a fixed numerical value.
114
![Page 7: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/7.jpg)
How to Choose M?
ObservationInitially, M only occurs in the zeroth row. As the zeroth row never becomes pivot row,this property is maintained while the simplex method is running.
All we need to have is an order on all values that can appear as reduced cost coefficients.
Order on cost coefficientsaM + b < c M + d :() (a < c) _ (a = c ^ b < d)
In particular, �aM + b < 0 < aM + b for any positive a and arbitrary b, and we candecide whether a cost coefficient is negative or not.
! There is no need to give M a fixed numerical value.
114
![Page 8: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/8.jpg)
How to Choose M?
ObservationInitially, M only occurs in the zeroth row. As the zeroth row never becomes pivot row,this property is maintained while the simplex method is running.
All we need to have is an order on all values that can appear as reduced cost coefficients.
Order on cost coefficientsaM + b < c M + d :() (a < c) _ (a = c ^ b < d)
In particular, �aM + b < 0 < aM + b for any positive a and arbitrary b, and we candecide whether a cost coefficient is negative or not.
! There is no need to give M a fixed numerical value.
114
![Page 9: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/9.jpg)
How to Choose M?
ObservationInitially, M only occurs in the zeroth row. As the zeroth row never becomes pivot row,this property is maintained while the simplex method is running.
All we need to have is an order on all values that can appear as reduced cost coefficients.
Order on cost coefficientsaM + b < c M + d :() (a < c) _ (a = c ^ b < d)
In particular, �aM + b < 0 < aM + b for any positive a and arbitrary b, and we candecide whether a cost coefficient is negative or not.
! There is no need to give M a fixed numerical value.
114
![Page 10: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/10.jpg)
How to Choose M?
ObservationInitially, M only occurs in the zeroth row. As the zeroth row never becomes pivot row,this property is maintained while the simplex method is running.
All we need to have is an order on all values that can appear as reduced cost coefficients.
Order on cost coefficientsaM + b < c M + d :() (a < c) _ (a = c ^ b < d)
In particular, �aM + b < 0 < aM + b for any positive a and arbitrary b, and we candecide whether a cost coefficient is negative or not.
! There is no need to give M a fixed numerical value.
114
![Page 11: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/11.jpg)
ExampleExample:
min x1 + x2 + x3s.t. x1 + 2 x2 + 3 x3 = 3
�x1 + 2 x2 + 6 x3 = 24 x2 + 9 x3 = 5
3 x3 + x4 = 1x1, . . . , x4 � 0
115
![Page 12: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/12.jpg)
Introducing Artificial Variables and M
Auxiliary problem:
min x1 +x2 +x3 +M x5 +M x6 +M x7s.t. x1 +2 x2 +3 x3 x5 = 3
�x1 +2 x2 +6 x3 +x6 = 24 x2 +9 x3 +x7 = 5
3 x3 +x4 = 1x1, . . . , x4, x5, x6, x7 � 0
Note that this time the unnecessary artificial variable x8 has been omitted.
We start off with (x5, x6, x7, x4) = (3, 2, 5, 1).
116
€14' then
Ca
ao
![Page 13: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/13.jpg)
Forming the Initial Tableau
x1 x2 x3 x4 x5 x6 x70 1 1 1 0 M M M
3 1 2 3 0 1 0 02 �1 2 6 0 0 1 05 0 4 9 0 0 0 11 0 0 3 1 0 0 0
Compute reduced costs by eliminating the nonzero entries for the basic variables.
117
⇒I
![Page 14: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/14.jpg)
Forming the Initial Tableau
x1 x2 x3 x4 x5 x6 x70 1 1 1 0 M M M
3 1 2 3 0 1 0 02 �1 2 6 0 0 1 05 0 4 9 0 0 0 11 0 0 3 1 0 0 0
Compute reduced costs by eliminating the nonzero entries for the basic variables.
117
![Page 15: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/15.jpg)
Forming the Initial Tableau
x1 x2 x3 x4 x5 x6 x7�3M �M + 1 �2M + 1 �3M + 1 0 0 M M
3 1 2 3 0 1 0 02 �1 2 6 0 0 1 05 0 4 9 0 0 0 11 0 0 3 1 0 0 0
Compute reduced costs by eliminating the nonzero entries for the basic variables.
117
![Page 16: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/16.jpg)
Forming the Initial Tableau
x1 x2 x3 x4 x5 x6 x7�5M 1 �4M + 1 �9M + 1 0 0 0 M
3 1 2 3 0 1 0 02 �1 2 6 0 0 1 05 0 4 9 0 0 0 11 0 0 3 1 0 0 0
Compute reduced costs by eliminating the nonzero entries for the basic variables.
117
![Page 17: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/17.jpg)
Forming the Initial Tableau
x1 x2 x3 x4 x5 x6 x7�10M 1 �8M + 1 �18M + 1 0 0 0 0
3 1 2 3 0 1 0 02 �1 2 6 0 0 1 05 0 4 9 0 0 0 11 0 0 3 1 0 0 0
Compute reduced costs by eliminating the nonzero entries for the basic variables.
117
![Page 18: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/18.jpg)
Forming the Initial Tableau
x1 x2 x3 x4 x5 x6 x7�10M 1 �8M + 1 �18M + 1 0 0 0 0
3 1 2 3 0 1 0 02 �1 2 6 0 0 1 05 0 4 9 0 0 0 11 0 0 3 1 0 0 0
Compute reduced costs by eliminating the nonzero entries for the basic variables.
117
I "o
Oa
oO
![Page 19: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/19.jpg)
First Iteration
x1 x2 x3 x4 x5 x6 x7�10M 1 �8M + 1 �18M + 1 0 0 0 0
3 1 2 3 0 1 0 02 �1 2 6 0 0 1 05 0 4 9 0 0 0 11 0 0 3 1 0 0 0
Reduced costs for x2 and x3 are negative.
Basis change: x3 enters the basis, x4 leaves.
118
![Page 20: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/20.jpg)
First Iteration
x1 x2 x3 x4 x5 x6 x7�10M 1 �8M + 1 �18M + 1 0 0 0 0
3 1 2 3 0 1 0 02 �1 2 6 0 0 1 05 0 4 9 0 0 0 11 0 0 3 1 0 0 0
Reduced costs for x2 and x3 are negative.
Basis change: x3 enters the basis, x4 leaves.
118
![Page 21: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/21.jpg)
First Iteration
x1 x2 x3 x4 x5 x6 x7�10M 1 �8M + 1 �18M + 1 0 0 0 0
3 1 2 3 0 1 0 02 �1 2 6 0 0 1 05 0 4 9 0 0 0 11 0 0 3 1 0 0 0
Reduced costs for x2 and x3 are negative.
Basis change: x3 enters the basis, x4 leaves.
118
⇐ I
add ( Ets
GM -
'
zE I
times Xs E I
![Page 22: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/22.jpg)
Second Iteration
x1 x2 x3 x4 x5 x6 x7�4M � 1/3 1 �8M + 1 0 6M � 1/3 0 0 0
2 1 2 0 �1 1 0 00 �1 2 0 �2 0 1 02 0 4 0 �3 0 0 1
1/3 0 0 1 1/3 0 0 0
Basis change: x2 enters the basis, x6 leaves.
119
0E I
so
c- 'z
-
![Page 23: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/23.jpg)
Second Iteration
x1 x2 x3 x4 x5 x6 x7�4M � 1/3 1 �8M + 1 0 6M � 1/3 0 0 0
2 1 2 0 �1 1 0 00 �1 2 0 �2 0 1 02 0 4 0 �3 0 0 1
1/3 0 0 1 1/3 0 0 0
Basis change: x2 enters the basis, x6 leaves.
119
![Page 24: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/24.jpg)
Third Iteration
x1 x2 x3 x4 x5 x6 x7�4M � 1/3 �4M + 3/2 0 0 �2M + 2/3 0 4M � 1/2 0
2 2 0 0 1 1 �1 00 �1/2 1 0 �1 0 1/2 02 2 0 0 1 0 �2 1
1/3 0 0 1 1/3 0 0 0
Basis change: x1 enters the basis, x5 leaves.
120
@o
C 0o O0 O
Current solution :
Xs -
- 2 Obj =4Mt.}
Xz 2 V
xy-
-
2×3=13
![Page 25: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/25.jpg)
Third Iteration
x1 x2 x3 x4 x5 x6 x7�4M � 1/3 �4M + 3/2 0 0 �2M + 2/3 0 4M � 1/2 0
2 2 0 0 1 1 �1 00 �1/2 1 0 �1 0 1/2 02 2 0 0 1 0 �2 1
1/3 0 0 1 1/3 0 0 0
Basis change: x1 enters the basis, x5 leaves.
120
![Page 26: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/26.jpg)
Fourth Iteration
x1 x2 x3 x4 x5 x6 x7�11/6 0 0 0 �1/12 2M � 3/4 2M + 1/4 0
21 1 0 0 1/2 1/2 �1/2 01/2 0 1 0 �3/4 1/4 1/4 00 0 0 0 0 �1 �1 1
1/3 0 0 1 1/3 0 0 0
Note that all artificial variables have already been driven to 0.
Basis change: x4 enters the basis, x3 leaves.
121
![Page 27: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/27.jpg)
Fourth Iteration
x1 x2 x3 x4 x5 x6 x7�11/6 0 0 0 �1/12 2M � 3/4 2M + 1/4 0
21 1 0 0 1/2 1/2 �1/2 01/2 0 1 0 �3/4 1/4 1/4 00 0 0 0 0 �1 �1 1
1/3 0 0 1 1/3 0 0 0
Note that all artificial variables have already been driven to 0.
Basis change: x4 enters the basis, x3 leaves.
121
![Page 28: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/28.jpg)
Fourth Iteration
x1 x2 x3 x4 x5 x6 x7�11/6 0 0 0 �1/12 2M � 3/4 2M + 1/4 0
21 1 0 0 1/2 1/2 �1/2 01/2 0 1 0 �3/4 1/4 1/4 00 0 0 0 0 �1 �1 1
1/3 0 0 1 1/3 0 0 0
Note that all artificial variables have already been driven to 0.
Basis change: x4 enters the basis, x3 leaves.
121
± 42-
÷
![Page 29: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/29.jpg)
Fifth Iteration
x1 x2 x3 x4 x5 x6 x7�7/4 0 0 1/4 0 2M � 3/4 2M + 1/4 01/2 1 0 �3/2 0 1/2 �1/2 05/4 0 1 9/4 0 1/4 1/4 00 0 0 0 0 �1 �1 11 0 0 3 1 0 0 0
We now have an optimal solution of the auxiliary problem, as all costs are nonnegative(M presumed large enough).
By elimiating the third row as in the previous example, we get a basic feasible and alsooptimal solution to the original problem.
122
![Page 30: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/30.jpg)
Fifth Iteration
x1 x2 x3 x4 x5 x6 x7�7/4 0 0 1/4 0 2M � 3/4 2M + 1/4 01/2 1 0 �3/2 0 1/2 �1/2 05/4 0 1 9/4 0 1/4 1/4 00 0 0 0 0 �1 �1 11 0 0 3 1 0 0 0
We now have an optimal solution of the auxiliary problem, as all costs are nonnegative(M presumed large enough).
By elimiating the third row as in the previous example, we get a basic feasible and alsooptimal solution to the original problem.
122
![Page 31: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/31.jpg)
Fifth Iteration
x1 x2 x3 x4 x5 x6 x7�7/4 0 0 1/4 0 2M � 3/4 2M + 1/4 01/2 1 0 �3/2 0 1/2 �1/2 05/4 0 1 9/4 0 1/4 1/4 00 0 0 0 0 �1 �1 11 0 0 3 1 0 0 0
We now have an optimal solution of the auxiliary problem, as all costs are nonnegative(M presumed large enough).
By elimiating the third row as in the previous example, we get a basic feasible and alsooptimal solution to the original problem.
122
X. = I × i = 8 otherwise
Xe = I Tpt = Iq4Xy
= O
Xc,
= I
![Page 32: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/32.jpg)
Computational Efficiency of the Simplex Method
ObservationThe computational efficiency of the simplex method is determined by
i the computational effort of each iteration;ii the number of iterations.
Question: How many iterations are needed in the worst case?
Idea for negative answer (lower bound)DescribeI a polyhedron with an exponential number of vertices;I a path that visits all vertices and always moves from a vertex to an adjacent one
that has lower costs.
123
![Page 33: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/33.jpg)
Computational Efficiency of the Simplex Method
ObservationThe computational efficiency of the simplex method is determined by
i the computational effort of each iteration;ii the number of iterations.
Question: How many iterations are needed in the worst case?
Idea for negative answer (lower bound)DescribeI a polyhedron with an exponential number of vertices;I a path that visits all vertices and always moves from a vertex to an adjacent one
that has lower costs.
123
![Page 34: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/34.jpg)
Computational Efficiency of the Simplex Method
ObservationThe computational efficiency of the simplex method is determined by
i the computational effort of each iteration;ii the number of iterations.
Question: How many iterations are needed in the worst case?
Idea for negative answer (lower bound)DescribeI a polyhedron with an exponential number of vertices;I a path that visits all vertices and always moves from a vertex to an adjacent one
that has lower costs.
123
![Page 35: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/35.jpg)
Computational Efficiency of the Simplex Method
Unit cubeConsider the unit cube in Rn, defined by the constraints
0 xi 1, i = 1, . . . , n
The unit cube hasI 2n vertices;I a spanning path, i. e., a path traveling the edges of the cube visiting each vertex
exactly once.
x1
x2
x1
x2
x3
124
![Page 36: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/36.jpg)
Computational Efficiency of the Simplex Method (cont.)
Klee-Minty cubeConsider a perturbation of the unit cube in Rn, defined by the constraints
0 x1 1,✏xi�1 xi 1 � ✏xi�1, i = 2, . . . , n
for some ✏ 2 (0, 1/2).
x1
x2
x1
x2
x3
125
![Page 37: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/37.jpg)
Computational Efficiency of the Simplex Method (cont.)
Klee-Minty cube
0 x1 1,✏xi�1 xi 1 � ✏xi�1, i = 2, . . . , n, ✏ 2 (0, 1/2)
Theorem 3.1.Consider the linear programming problem of minimizing �xn subject to the constraintsabove. Then,
a the feasible set has 2n vertices;b the vertices can be ordered so that each one is adjacent to and has lower cost than
the previous one;c there exists a pivoting rule under which the simplex method requires 2n � 1
changes of basis before it terminates.
126
![Page 38: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/38.jpg)
Computational Efficiency of the Simplex Method (cont.)
Klee-Minty cube
0 x1 1,✏xi�1 xi 1 � ✏xi�1, i = 2, . . . , n, ✏ 2 (0, 1/2)
Theorem 3.1.Consider the linear programming problem of minimizing �xn subject to the constraintsabove. Then,
a the feasible set has 2n vertices;b the vertices can be ordered so that each one is adjacent to and has lower cost than
the previous one;c there exists a pivoting rule under which the simplex method requires 2n � 1
changes of basis before it terminates.
126
![Page 39: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/39.jpg)
Diameter of Polyhedra
Definition 3.2.I The distance d(x , y) between two vertices x , y is the minimum number of edges
required to reach y starting from x .
I The diameter D(P) of polyhedron P is the maximum d(x , y) over all pairs ofvertices (x , y).
I �(n,m) is the maximum D(P) over all polytopes in Rn that are represented interms of m inequality constraints.
I �u(n,m) is the maximum D(P) over all polyhedra in Rn that are represented interms of m inequality constraints.
127
![Page 40: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/40.jpg)
Diameter of Polyhedra
Definition 3.2.I The distance d(x , y) between two vertices x , y is the minimum number of edges
required to reach y starting from x .
I The diameter D(P) of polyhedron P is the maximum d(x , y) over all pairs ofvertices (x , y).
I �(n,m) is the maximum D(P) over all polytopes in Rn that are represented interms of m inequality constraints.
I �u(n,m) is the maximum D(P) over all polyhedra in Rn that are represented interms of m inequality constraints.
�(2, 8) =⌅82⇧= 4
127
![Page 41: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/41.jpg)
Diameter of Polyhedra
Definition 3.2.I The distance d(x , y) between two vertices x , y is the minimum number of edges
required to reach y starting from x .
I The diameter D(P) of polyhedron P is the maximum d(x , y) over all pairs ofvertices (x , y).
I �(n,m) is the maximum D(P) over all polytopes in Rn that are represented interms of m inequality constraints.
I �u(n,m) is the maximum D(P) over all polyhedra in Rn that are represented interms of m inequality constraints.
�(2, 8) =⌅82⇧= 4
�(2,m) =⌅m2⇧
127
![Page 42: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/42.jpg)
Diameter of Polyhedra
Definition 3.2.I The distance d(x , y) between two vertices x , y is the minimum number of edges
required to reach y starting from x .
I The diameter D(P) of polyhedron P is the maximum d(x , y) over all pairs ofvertices (x , y).
I �(n,m) is the maximum D(P) over all polytopes in Rn that are represented interms of m inequality constraints.
I �u(n,m) is the maximum D(P) over all polyhedra in Rn that are represented interms of m inequality constraints.
�(2, 8) =⌅82⇧= 4
�(2,m) =⌅m2⇧
�u(2, 8) = 8� 2 = 6
127
• 7-7
![Page 43: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/43.jpg)
Diameter of Polyhedra
Definition 3.2.I The distance d(x , y) between two vertices x , y is the minimum number of edges
required to reach y starting from x .
I The diameter D(P) of polyhedron P is the maximum d(x , y) over all pairs ofvertices (x , y).
I �(n,m) is the maximum D(P) over all polytopes in Rn that are represented interms of m inequality constraints.
I �u(n,m) is the maximum D(P) over all polyhedra in Rn that are represented interms of m inequality constraints.
�(2, 8) =⌅82⇧= 4
�(2,m) =⌅m2⇧
�u(2, 8) = 8� 2 = 6�u(2,m) = m � 2
127
![Page 44: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/44.jpg)
Hirsch ConjectureObservation: The diameter of the feasible set in a linear programming problem is alower bound on the number of steps required by the simplex method, no matter whichpivoting rule is being used.
Polynomial Hirsch Conjecture
�(n,m) poly(m, n)
RemarksI Known lower bounds: �u(n,m) � m � n +
⌅n5⇧
I Known upper bounds:
�(n,m) �u(n,m) < m1+log2 n = (2n)log2 m
I The Strong Hirsch Conjecture
�(n,m) m � n
was disproven in 2010 by Paco Santos for n = 43, m = 86.
128
![Page 45: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/45.jpg)
Hirsch ConjectureObservation: The diameter of the feasible set in a linear programming problem is alower bound on the number of steps required by the simplex method, no matter whichpivoting rule is being used.
Polynomial Hirsch Conjecture
�(n,m) poly(m, n)
RemarksI Known lower bounds: �u(n,m) � m � n +
⌅n5⇧
I Known upper bounds:
�(n,m) �u(n,m) < m1+log2 n = (2n)log2 m
I The Strong Hirsch Conjecture
�(n,m) m � n
was disproven in 2010 by Paco Santos for n = 43, m = 86.
128
![Page 46: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/46.jpg)
Hirsch ConjectureObservation: The diameter of the feasible set in a linear programming problem is alower bound on the number of steps required by the simplex method, no matter whichpivoting rule is being used.
Polynomial Hirsch Conjecture
�(n,m) poly(m, n)
RemarksI Known lower bounds: �u(n,m) � m � n +
⌅n5⇧
I Known upper bounds:
�(n,m) �u(n,m) < m1+log2 n = (2n)log2 m
I The Strong Hirsch Conjecture
�(n,m) m � n
was disproven in 2010 by Paco Santos for n = 43, m = 86.
128
![Page 47: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/47.jpg)
Hirsch ConjectureObservation: The diameter of the feasible set in a linear programming problem is alower bound on the number of steps required by the simplex method, no matter whichpivoting rule is being used.
Polynomial Hirsch Conjecture
�(n,m) poly(m, n)
RemarksI Known lower bounds: �u(n,m) � m � n +
⌅n5⇧
I Known upper bounds:
�(n,m) �u(n,m) < m1+log2 n = (2n)log2 m
I The Strong Hirsch Conjecture
�(n,m) m � n
was disproven in 2010 by Paco Santos for n = 43, m = 86.
128
![Page 48: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/48.jpg)
Hirsch ConjectureObservation: The diameter of the feasible set in a linear programming problem is alower bound on the number of steps required by the simplex method, no matter whichpivoting rule is being used.
Polynomial Hirsch Conjecture
�(n,m) poly(m, n)
RemarksI Known lower bounds: �u(n,m) � m � n +
⌅n5⇧
I Known upper bounds:
�(n,m) �u(n,m) < m1+log2 n = (2n)log2 m
I The Strong Hirsch Conjecture
�(n,m) m � n
was disproven in 2010 by Paco Santos for n = 43, m = 86.
128
![Page 49: The Two-phase Simplex Method - University of Liverpoolgairing/COMP557/board/20181025.pdfThe Two-phase Simplex Method Two-phase simplex method 1 Given an LP in standard from, first](https://reader033.fdocuments.in/reader033/viewer/2022052303/5f0cd6ed7e708231d4376358/html5/thumbnails/49.jpg)
Average Case Behavior of the Simplex Method
I Despite the exponential lower bounds on the worst case behavior of the simplexmethod (Klee-Minty cubes etc.), the simplex method usually behaves well inpractice.
I The number of iterations is “typically” O(m).
I There have been several attempts to explain this phenomenon from a moretheoretical point of view.
I These results say that “on average” the number of iterations is O(·) (usuallypolynomial).
I One main difficulty is to come up with a meaningful and, at the same time,manageable definition of the term “on average”.
129