Scribd
Upload
25 views3 pages

8 Reducibility

Uploaded by

Arda Barış Örtlek
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
25 views3 pages

8 Reducibility

Uploaded by

Arda Barış Örtlek
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

CS 476 – 8.

Reducibility

1 Reducibility
Definition Reduction: transforming one problem into another, such that the solution to the second problem
yields the solution to the first one.
solve A(. . .)
...
...
solve B(. . .)
...
...
return
• If B can be solved, then A can be solved.
• Or, negatively, if A is unsolvable, then B is unsolvable.
Notation: A ≤ B:
• A can be reduced to B.
• A is no harder than B.
Example
solve Ld (. . .)
...
...
solve AT M (. . .)
...
...
return
Ld ≤ AT M

1.1 More undecidable problems / languages


• HALTT M = {< M, w > | M halts on w}

Theorem 1.1 HALTT M is undecidable.

Proof by reduction of AT M
Assume HALTT M decidable by some decider R. Then we can construct a decider S for AT M with
Algorithm 1.

Algorithm 1 S on < M, w >


Run R on < M, w >
if R rejects (i.e. M loops on w) then
reject
else
Run M on < w >
if M accepts w then
accept
else if M rejects w then
reject
end if
end if
But we know that AT M is undecidable and no such S exists. Then, no such R exists, and HALTT M
is undecidable.

• ET M = {< M > | L(M ) = ∅}

Theorem 1.2 ET M is undecidable.

Proof by reduction from AT M .


For a given AT M instance < M, w >, construct an ET M instance < M1 > such that L(M1 ) will be
∅ or not depending on whether M accepts w or not with Algorithm 2.

Algorithm 2 M1 on x
if x 6= w then
reject
else
Run M on < w >
if M accepts w then
accept
else if M rejects w then
reject
end if
end if


∅, if w 6∈ L(M )
L(M1 ) =
{w}, if w ∈ L(M )
Now, assume ET M is decidable by some TM R. We can construct a decider S for AT M using
Algorithm 3.

Algorithm 3 S on < M, w >


Construct M1 from M and w
Run R on < M1 >
if R accepts then
reject
else if R rejects then
accept
end if

We know S does not exist. Therefore ET M is undecidable.

• REGU LART M = {< M > | L(M ) is a regular language}


Similarly, we will reduce from AT M .
The idea is to construct a new machine M2 that recognizes a regular language iff M accepts w.
modified machine M2 . Initially, M2 will recognize the non-regular language {0n 1n |n ≥ 0} if M
does not accept w, and will recognize the regular language Σ∗ if M accepts w.
M2 = On input x:

1. If x has the form 0n 1n , accept.


2. If x does not have this form, run M on input w and accept if M accepts w.

2
Let R be a decider TM that decides REGU LART M , and let S be a decider for AT M .

Algorithm 4 S on < M, w >


Construct M2 from M and w.
Run R on < M2 >
if R accepts then
accept
else if R rejects then
reject
end if

We know S does not exist. Therefore REGU LART M is undecidable.

• EQT M = {< M1 , M2 > | L(M1 ) = L(M2 )}

Theorem 1.3 EQT M is undecidable.

Proof by reduction from ET M .


Assume EQT M is decidable by R. Then we can have a decider for S as outlined in Algorithm 5.

Algorithm 5 S on < M >


Construct a TM M3 that rejects everything. (i.e. L(M3 ) = ∅)
Run R on < M, M3 >
if R accepts then
accept
else if R rejects then
reject
end if

We know S does not exist. Therefore EQT M is undecidable.

1.2 Rice’s Theorem


For any non-trivial property P of the language of a TM:

PT M = {< M > | L(M ) has property P}

is undecidable.

L(M ) ⊆ Σ∗

T rivial properties
L(M ) ⊇ ∅

Example

• REGU LART M = {< M > | L(M ) is a regular language} is undecidable.

• CFT M = {< M > | L(M ) is a context f ree language} is undecidable.

You might also like