reductions
- From: "surf" <surfunbear@xxxxxxxxx>
- Date: 25 Nov 2006 08:16:40 -0800
I've been having some problems with this.
first of all.
~hp = {<m,x> | m is a turing machine that does not halt on x}
~hp is not r.e. If I am given any language A that is not r.e. Should
it be easy to reduce ~HP to A ?
Some books imply that if you are reducing X to Y, Y should be simpler,
but on the other hand sometimes you want to show Y is not r.e. or
decidable, in that case you pick X that is not r.e or decidable, for
that situation does X need to be simpler ? Or what does it mean to be
simpler ?
The hello world tested in the Hopcroft, Ullman automata theory book on
page 324 is an example of reducing H to H2 ?
I was given the following homework problem:
let A = {<M,x> | M is a turing machine and on input x outputs the
reverse of x}
show that A is r.e. but not decidable.
It seems to me that it is only r.e. up to a given size of the input.
Suppose you can test all possible strings for x up to length 1 million.
Isn't it possible at 1 trillion someplace that M will not reverse it's
input ? Since the size of the input is potentially infinite, how can
you ever be sure a given program allways will reverse it's input even
if it has for many such input ? I used the method in the Ullman book
page 324 to show it was not decidable, though perhaps there was a
simpler way ?
.
- Follow-Ups:
- Re: reductions
- From: Patricia Shanahan
- Re: reductions
- Prev by Date: Re: Discussion regarding Mr. Diabys algorithm
- Next by Date: Re: Discussion regarding Mr. Diabys algorithm
- Previous by thread: maximal 3-colorable subset
- Next by thread: Re: reductions
- Index(es):
Relevant Pages
|