Skip to content

Turing Machines & Undecidability

Short Notes

Turing Machines (TM) are the most powerful computational models. Undecidability deals with problems that cannot be solved by any algorithm (TM).

Chomsky Hierarchy & Power

Regular \(\subset\) CFL \(\subset\) CSL \(\subset\) Recursive (Rec) \(\subset\) Recursively Enumerable (RE).

Language Class Recognizer
Regular Finite Automata
CFL Pushdown Automata
CSL Linear Bounded Automata
Recursive TM which always halts
RE TM which halts for YES, may loop for NO

Key Theories & Formulas

1. Halting Problem

The problem of determining whether a TM \(M\) will halt on input \(w\) is undecidable and RE.

2. Rice's Theorem

Every non-trivial property of the languages recognized by TMs is undecidable.

  • "Is \(L(M)\) empty?" - Undecidable.
  • "Is \(M\) starting with 5 states?" - Decidable (it's a property of the machine, not its language).

Example Problems

Problem: Is it decidable if a Regular Language is finite?

  • Result: YES. All properties of Regular languages are decidable.

Hardest GATE Questions

Topic: Decidability Table (The "Master Chart") Tricky Question (GATE 2013/2016/2020): Is \(\{ \langle M \rangle \mid M \text{ is a TM and } |L(M)| \ge 2 \}\) recursively enumerable?

  • Analysis: YES. We can non-deterministically guess two different strings and run \(M\) on them. If both halt/accept, we accept.
  • The "Trap": Checking if it is Recursive. It's not. If \(M\) recognizes a language with only 1 string, we might loop forever.
  • Hard Aspect: Identifying "Trivial" vs "Non-trivial" in Rice's Theorem.
  • Trivial: Property is true for all RE languages or none.
  • Complexity: Post Correspondence Problem (PCP).
  • PCP is undecidable for alphabet size \(\ge 2\).
  • Modified PCP is also undecidable

References