1. CSCI 2670 Introduction to Theory of Computing September 28, 2005

2. Agenda This week –Turing machines –Read section 3.1

3. September 28, 2005 Announcements Tests will be returned tomorrow Tutorial sessions are suspended until further notice –Extended office hours while tutorials are suspended Monday 11:00 – 12:00 Tuesday 3:00 – 4:00 Wednesday 3:00 – 5:00

4. September 28, 2005 Recap to date Finite automata (both deterministic and nondeterministic) machines accept regular languages –Weakness: no memory Pushdown automata accept context- free grammars –Add memory in the form of a stack –Weakness: stack is restrictive

5. September 28, 2005 Turing machines Similar to a finite automaton –Unrestricted memory in the form of a tape Can do anything a real computer can do! –Still cannot solve some problems Church-Turing thesis: any effective computation can be carried out by some Turing machine

6. September 28, 2005 Touring machine schematic Control a b a ~ Initially tape contains the input string –Blanks everywhere else (denoted ~ in class … different symbol in book) Machine may write information on the tape

7. September 28, 2005 Touring machine schematic Control a b a ~ Can move tape head to read information written to tape Continues computing until output produced –Output values accept or reject

8. September 28, 2005 Touring machine schematic Control a b a ~ Turing machine results –Accept –Reject –Never halts We may not be able to tell result by observation

9. September 28, 2005 Differences between TM and FA 1.TM has tape you can read from and write to 2.Read-write head can be moved in either direction 3.Tape is infinite 4.Accept and reject states take immediate effect

10. September 28, 2005 Example How can we design a Turing machine to find the middle of a string? –If string length is odd, return middle symbol –If string length is even, reject string Make multiple passes over string Xing out symbols at end until only middle remains

11. September 28, 2005 Processing input 1.Check if string is empty  If so, return reject 2.Write X over first and last non-X symbols  After this, the head will be at the second X 3.Move left one symbol  If symbol is an X, return reject (string is even in length) 4.Move left one symbol  If symbol is an X, return accept (string is even in length) 5.Go to step 2

12. September 28, 2005 Example 00110~ –First check if string is empty –X first and last non-X symbols X011X~ –Move left one symbol X011X~ –Is symbol an X? No –Move left one symbol X011X~ –Is symbol an X? No –Write X over first and last non-X symbols

13. September 28, 2005 Example XX1XX~ –Move left one symbol XX1XX~ –Is symbol an X? No –Move left one symbol XX1XX~ –Is symbol an X? Yes –Return accept

14. September 28, 2005 Formal definition of a TM Definition: A Turing machine is a 7- tuple (Q, , , ,q 0,q accept,q reject ), where Q, , and  are finite sets and 1.Q is the set of states, 2.  is the input alphabet not containing the special blank symbol ~ 3.  is the tape alphabet, where ~  and , 4.  : Q  Q  {L,R} is the transition function,

15. September 28, 2005 What is the Transition Function?? Q = set of states,  = tape alphabet  : Q  Q  {L,R} Given: The current internal state  Q The symbol on the current tape cell Then  tells us what the TM does: Changes to new internal state  Q Either writes new symbol   Or moves one cell left or right

16. September 28, 2005 Formal definition of a TM Definition: A Turing machine is a 7- tuple (Q, , , ,q 0,q accept,q reject ), where Q, , and  are finite sets and 5.q 0  Q is the start state, 6.q accept  Q is the accept state, and 7.q reject  Q is the reject state, where q reject  q accept

17. September 28, 2005 Computing with a TM M receives input w = w 1 w 2 …w n  * on leftmost n squares of tape –Rest of tape is blank (all ~ symbols) Head position begins at leftmost square of tape Computation follows rules of  Head never moves left of leftmost square of the tape –If  says to move L, head stays put!

18. September 28, 2005 Completing computation Continue following  transition rules until M reaches q accept or q reject –Halt at these states May never halt if the machine never transitions to one of these states!

19. September 28, 2005 Another TM example We want to create a TM to add two numbers Use a simple tape alphabet {0,1} plus the blank symbol Represent a number n by a string of n+1 1’s terminated by a zero Input to compute 3+4 looks like this:

20. September 28, 2005 Result of the TM addition example Note that the TM is initially positioned on the leftmost cell of the input. When the TM halts in the accept state, it must also be on the leftmost cell of the output:

21. September 28, 2005 Breaking down the addition problem Good computer scientists like to simplify A successor TM appends a 1 to the right end of a string of 1’s

22. September 28, 2005 The successor subroutine The TM starts in the initial state s 0, positioned on the leftmost of a string of 1’s If it sees a 1, it writes a 1, moves right, and stays in state s 0 If it sees a 0, it writes a 1 and moves to

23. September 28, 2005 Successor subroutine state transitions (original state, input, new state, action)

24. September 28, 2005 TM state interpretation S 0 – the TM has seen only 1’s so far and is scanning right S 1 – the TM has seen its first zero and is scanning left S 2 – the TM has returned to the leftmost 1 and halts. View movie of this TM

25. September 28, 2005 From successor TM to addition TM The successor TM will join the two blocks of n+1 1’s and m+1 1’s into a single block of n+m+3 1’s To complete the computation, knock off two 1’s from left end (states s 2 and s 3 )

26. September 28, 2005 TM configurations The configuration of a Turing machine is the current setting –Current state –Current tape contents –Current tape location Notation uqv –Current state = q –Current tape contents = uv Only ~ symbols after last symbol of v –Current tape location = first symbol of v

27. September 28, 2005 Acknowledgements TM addition example from Stanford, http://plato.stanford.edu/entries/turing-machine/