Scribd
Upload
64 views

Chomsky Hierarchy Languages 2. Turing Reducibility 3. The Class P

The document discusses Chomsky hierarchy of formal languages and describes four language classes: Type-0 (recursively enumerable languages accepted by Turing machines), Type-1 (context-sensitive languages accepted by linear-bounded automata), Type-2 (context-free languages accepted by pushdown automata), and Type-3 (regular languages accepted by finite automata). It also discusses Turing reducibility and the complexity classes P and NP, defining polynomial-time solvability and introducing NP-completeness.

Uploaded by

Divya Bhat
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
64 views

Chomsky Hierarchy Languages 2. Turing Reducibility 3. The Class P

The document discusses Chomsky hierarchy of formal languages and describes four language classes: Type-0 (recursively enumerable languages accepted by Turing machines), Type-1 (context-sensitive languages accepted by linear-bounded automata), Type-2 (context-free languages accepted by pushdown automata), and Type-3 (regular languages accepted by finite automata). It also discusses Turing reducibility and the complexity classes P and NP, defining polynomial-time solvability and introducing NP-completeness.

Uploaded by

Divya Bhat
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 8

AUTOMATA THEORY AND COMPUTABILITY 18CS54

MODULE V

1. Chomsky Hierarchy Languages


2. Turing Reducibility
3. The Class P

1. Chomsky Hierarchy of Languages

 A containment hierarchy (strictly nested sets) of classes of formal grammars

The Hierarchy

Class Grammars Languages Automaton

Type-0 Unrestricted Recursively enumerable Turing machine


(Turing-recognizable)

none Recursive Decider

(Turing-decidable)

Type-1 Context-sensitive Context-sensitive Linear-bounded

Type-2 Context-free Context-free Pushdown

Type-3 Regular Regular Finite

Type 0 Unrestricted:

Languages defined by Type-0 grammars are accepted by Turing machines .

Rules are of the form: α → β, where α and β are arbitrary strings over a vocabulary V and
α≠ε

Type 1 Context-sensitive:

Languages defined by Type-1 grammars are accepted by linear-bounded automata.

Syntax of some natural languages (Germanic)

Available At VTU HUB (Android App)


AUTOMATA THEORY AND COMPUTABILITY 18CS54
Rules are of the form:

αAβ → αBβ

S→ε

where

Available At VTU HUB (Android App)


AUTOMATA THEORY AND COMPUTABILITY 18CS54

Type 2 Context-free:

Languages defined by Type-2 grammars are accepted by push-down automata.

Natural language is almost entirely definable by type-2 tree structures

Rules are of the form:

A→α

Where

A∈N

α ∈ (N ⋃ Σ)∗

Type 3 Regular:

Languages defined by Type-3 grammars are accepted by finite state automata

Most syntax of some informal spoken dialog

Rules are of the form:

A→ε

A→α

A → αB

where

A, B ∈ N and α ∈ Σ

The Universal Turing Machine

 If Tm’s are so damned powerful, can’t we build one that simulates the behavior of any
Tm on any tape that it is given?

Available At VTU HUB (Android App)


AUTOMATA THEORY AND COMPUTABILITY 18CS54

 Yes. This machine is called the Universal Turing machine.

 How would we build a Universal Turing machine?

 We place an encoding of any Turing machine on the input tape of the Universal
Tm.

 The tape consists entirely of zeros and ones (and, of course, blanks)

 Any Tm is represented by zeros and ones, using unary notation for elements and
zeros as separators.

 Every Tm instruction consists of four parts, each a represented as a series of 1's and
separated by 0's.

 Instructions are separated by 00.

 We use unary notation to represent components of an instruction, with

 0 = 1,

 1 = 11,

 2 = 111,

 3 = 1111,

 n = 111...111 (n+1 1's).

 We encode qn as n + 1 1's

 We encode symbol an as n + 1 1's

 We encode move left as 1, and move right as 11

1111011101111101110100101101101101100

q3, a2, q4, a2, L q0, a1, q1, a1, R

 Any Turing machine can be encoded as a unique long string of zeros and ones, beginning
with a 1.

 Let Tn be the Turing machine whose encoding is the number n.

Available At VTU HUB (Android App)


AUTOMATA THEORY AND COMPUTABILITY 18CS54
2. Turing Reducibility

• A language A is Turing reducible to a language B, written A T B, if A is decidable


relative to B

• Below it is shown that ETM is Turing reducible to EQTM

• Whenever A is mapping reducible to B, then A is Turing reducible to B

– The function in the mapping reducibility could be replaced by an oracle

• An oracle Turing machine with an oracle for EQTM can decide ETM

Available At VTU HUB (Android App)


AUTOMATA THEORY AND COMPUTABILITY 18CS54
TEQ-TM = “On input <M>

1. Create TM M1 such that L(M1) = 

M1 has a transition from start state to reject state for every element of 

1. Call the EQTM oracle on input <M,M2>

2. If it accepts, accept; if it rejects, reject”

• TEQ-TM decides ETM

• ETM is decidable relative to EQTM

• Applications
• If A T B and B is decidable, then A is decidable
• If A T B and A is undecidable, then B is undecidable
• If A T B and B is Turing-recognizable, then A is Turing-recognizable
• If A T B and A is non-Turing-recognizable, then B is non-Turing-recognizable

3. The class P

A decision problem D is solvable in polynomial time or in the class P, if there exists an


algorithm A such that

• A Takes instances of D as inputs.


• A always outputs the correct answer “Yes” or “No”.
• There exists a polynomial p such that the execution of A on inputs of size n always
terminates in p(n) or fewer steps.
• EXAMPLE: The Minimum Spanning Tree Problem is in the class P.

The class P is often considered as synonymous with the class of computationally


feasible problems, although in practice this is somewhat unrealistic.

The class NP

A decision problem is nondeterministically polynomial-time solvable or in the class NP if


there exists an algorithm A such that

• A takes as inputs potential witnesses for “yes” answers to problem D.


• A correctly distinguishes true witnesses from false witnesses.

Available At VTU HUB (Android App)


AUTOMATA THEORY AND COMPUTABILITY 18CS54
• There exists a polynomial p such that for each potential witnesses of each instance of
size n of D, the execution of the algorithm A takes at most p(n) steps.
• Think of a non-deterministic computer as a computer that magically “guesses” a
solution, then has to verify that it is correct

o If a solution exists, computer always guesses it

o One way to imagine it: a parallel computer that can freely spawn an infinite
number of processes

 Have one processor work on each possible solution

 All processors attempt to verify that their solution works

 If a processor finds it has a working solution

o So: NP = problems verifiable in polynomial time

o Unknown whether P = NP (most suspect not)

NP-Complete Problems

• We will see that NP-Complete problems are the “hardest” problems in NP:
o If any one NP-Complete problem can be solved in polynomial time.
o Then every NP-Complete problem can be solved in polynomial time.
o And in fact every problem in NP can be solved in polynomial time (which would
show P = NP)
o Thus: solve hamiltonian-cycle in O(n100) time, you’ve proved that P = NP. Retire
rich & famous.
• The crux of NP-Completeness is reducibility

o Informally, a problem P can be reduced to another problem Q if any instance of P


can be “easily rephrased” as an instance of Q, the solution to which provides a
solution to the instance of P

 What do you suppose “easily” means?

 This rephrasing is called transformation

o Intuitively: If P reduces to Q, P is “no harder to solve” than Q

• An example:

o P: Given a set of Booleans, is at least one TRUE?

o Q: Given a set of integers, is their sum positive?

Available At VTU HUB (Android App)


AUTOMATA THEORY AND COMPUTABILITY 18CS54
o Transformation: (x1, x2, …, xn) = (y1, y2, …, yn) where yi = 1 if xi = TRUE, yi = 0
if xi = FALSE

• Another example:

o Solving linear equations is reducible to solving quadratic equations

 How can we easily use a quadratic-equation solver to solve linear


equations?

• Given one NP-Complete problem, we can prove many interesting problems NP-Complete

o Graph coloring (= register allocation)

o Hamiltonian cycle

o Hamiltonian path

o Knapsack problem

o Traveling salesman

o Job scheduling with penalties, etc.

NP Hard

 Definition: Optimization problems whose decision versions are NP- complete are
called NP-hard.

 Theorem: If there exists a polynomial-time algorithm for finding the optimum in


any NP-hard problem, then P = NP.
Proof: Let E be an NP-hard optimization (let us say minimization) problem, and let A
be a polynomial-time algorithm for solving it. Now an instance J of the corresponding
decision problem D is of the form (I, c), where I is an instance of E, and c is a
number. Then the answer to D for instance J can be obtained by running A on I and
checking whether the cost of the optimal solution exceeds c. Thus there exists a
polynomial-time algorithm for D, and NP-completeness of the latter implies P= NP.

Available At VTU HUB (Android App)

You might also like