CS 3240 – Chapters. PDAs with 2 stacks or 1 queue TMs with 1-way infinite tape TMs with n tapes...

Post on 17-Jan-2016

218 views 0 download

Transcript of CS 3240 – Chapters. PDAs with 2 stacks or 1 queue TMs with 1-way infinite tape TMs with n tapes...

CS 3240 – Chapters

PDAs with 2 stacks or 1 queue TMs with 1-way infinite tape TMs with n tapes or n heads or n-dim tapes TMs with various “move” options

Move by a number of cells (including 0), random cell access by position

Adding non-determinism to a Standard TM Church's Lambda Calculus (LISP, Haskell) Unrestricted Grammars Matrix Grammars Post Systems Markov Algorithms Structured Programming with unlimited

memory Any abstract machine yet imagined by man!

Queues pop at front, push at back 2 I/O points gives more flexibility Equivalent to a TM!

Begin by pushing all initial data Can simulate “moving around” by

cycling through the data circular shift via pop and push of same

character use a delimiter character ($) to keep track

of the start of data

$aabbccaabbcc$abbcc$Xbbcc$Xabcc$XaYcc$XaYbc$XaYbZ$XaBbZcXaYbZc$aYbZc$XYbZc$XX

bZc$XXYZc$XXYYc$XXYYZ$XXYYZZXXYYZZ$YYZZ$XXZZ$XXYY$XXYYZZXXYYZZ$<accept>

Just store the tuple (a,b,c) as the “symbol” in a single cell.Change part or all of the data as needed.

Just “wrap around the fold” to simulate a 2-way tape

Figure 10.04:

Figure 10.05:

Just place in different sections of a single tape

Figure 10.11:

Code the machine so it replicates and calls itself for each choice. (A form of backtracking). Do a breadth-first search for a halting configuration.

The TM we’ve seen so far have been “special-purpose computers” they implement only one algorithm, or they accept only one language

But TMs can take another TM as input, and simulate (run) it “stored program computer” general purpose computer as we know it

A TM that simulates other TMs. Mu takes M as input, along with data input.See page 267 for a sample encoding for TMs.

The input TM must be encoded as a string States, transitions, etc.

One way: Order the tape alphabet, Γ = {a1, a2, a3…}

▪ a1 = ☐, a2 = L, a3 = R, a4 = 0, a5 = 1, …

Encode each state by its index+1▪ q0= 1, q1 = 11 …

Encode each transition in δ Use 0 as the separator everywhere

δ(q0,1) = (q0,1,R)Consider it as the quintuple

(q0,1,q0,1,R)Which encodes as:

(1,5,1,5,3)= 1 11111 1 11111 111= 1011111010111110111

Notation: TU = the UTM; TM = the input machine TU has a memory location for the current state of

TM

TU reserves a portion of the tape for the encoding of the TM

TU reserves an infinitely large portion of the tape for the working memory of the TM (in one direction), initialized with the input to TM

And keeps track of the read/write position TU has its own infinitely large memory area (in the

other direction)

We won’t investigate the inner workings of UTMs further

We’ll just assume they exist (If they didn’t, neither would computers :-)

What we will do is examine the consequences of the existence of UTMs especially the consequences of encoding

TMs as strings

http://ironphoenix.org/tril/tm/Here are the instructions for the

swap machine: 1,a,1,b,> 1,b,1,a,> 1,_,2,_,< 2,a,2,a,< 2,b,2,b,< 2,_,H,_,>

Every TM can be encoded as a bitstring Not all strings represent TMs. TM form:

(11*0)50((11*0)50)*

Some TMs can have more than one string representation e.g., the name(number) of the state is

immaterial

How many strings are there? How many TMs are there?

Some infinite sets are countable like the set of positive even numbers {0, 2, 4, …} = {2n | n ∈ N}; function: p(i) =

2i There is a one-to-one mapping between

the set’s elements and the Natural numbers

Others are not countable: e.g., the set of real numbers Consequence: |(0,1)| = |(-∞,∞)| !!!

-2, -4, -6, … Maps to 1, 2, 3, … By the function: n(i) = -2i

e(0) = 0 e(1) = 2 e(2) = -2 e(3) = 4 e(4) = -4 e(5) = 6 e(6) = -6

The “formula”:

If i is evene(i) = -i

elsee(i) = i + 1

Take a cue from number systems’ positional notation

123 = 1*102 + 2*10 + 3aba = v(a)*n2 + v(b)*n + v(a), where

n = |Σ|, v(a) = 1, v(b) = 2, f(λ) = 0So f(aba) = 1*22 + 2*2 + 1 = 9aba is the 9th string (0-based) in Σ*:

{λ,a,b,aa,ab,ba,bb,aaa,aab,aba,…}

A Language is a subset of Σ*

We can enumerate the elements of Σ* (or any subset thereof) in proper order lexicographically in groups by increasing

length Therefore, there is a first one, a second one,

etc. So, the strings of a language are

enumerable ⇒ Every language is a countable set of

strings

aka enumerable They can be arranged in a sequence

▪ TM0, TM1, TM2, etc.Why?

Because they can be encoded as strings Strings over any alphabet can be

enumerated in proper order▪ by length groups, lexicographically

So there you have it!

Some TMs only need a fixed size of working storage never grows beyond (a factor of the) input size

Example: anbncn

Doesn’t expand working storage (remember?) These are called Linear Bounded Automata

LBA; they use end-markers: [<data goes here>] Why do we care?

We don’t, much; will use briefly in Chapter 11