From: Piotr Wyderski on
Hello,

I would like to learn a bit about two-way Deterministic Finite
Automata. I am particularly interested in their practical applications,
namely the procedure of conversion from an NFA to a 2-DFA
and its minimization. Could you please provide me some reference
on the subject?

Best regards
Piotr Wyderski
From: Torben �gidius Mogensen on
"Piotr Wyderski" <piotr.wyderski(a)mothers.against.spam.gmail.com> writes:

> Hello,
>
> I would like to learn a bit about two-way Deterministic Finite
> Automata. I am particularly interested in their practical applications,
> namely the procedure of conversion from an NFA to a 2-DFA
> and its minimization. Could you please provide me some reference
> on the subject?

Since any DFA is also a 2-DFA, you can use the normal NFA to DFA
contruction. A 2-DFA can not recognize any language that is not also
recognized by a DFA, but it may be more succint. I don't know about
minimization, but I have a feeling it is not easy (or at least not both
easy and efficient).

Torben
From: Piotr Wyderski on
Torben Agidius Mogensen wrote:

> Since any DFA is also a 2-DFA, you can use the normal NFA to DFA
> contruction.

Well, of course it would be valid by definition, but the point is to skip
the potentially exponential state space explosion during the process
of conversion from NFA to DFA. If there is no algorithmic advantage
in constructing 2-DFAs, then they are only of academic interest. But
I believe there is huge potential to be explored...

Best regards
Piotr Wyderski

From: Daniel Pitts on
On 6/22/2010 9:24 AM, Piotr Wyderski wrote:
> Torben Agidius Mogensen wrote:
>
>> Since any DFA is also a 2-DFA, you can use the normal NFA to DFA
>> contruction.
>
> Well, of course it would be valid by definition, but the point is to skip
> the potentially exponential state space explosion during the process
> of conversion from NFA to DFA. If there is no algorithmic advantage
> in constructing 2-DFAs, then they are only of academic interest. But
> I believe there is huge potential to be explored...
Without having read anything on 2-DFAs, it sounds like a disguise for a
NDA. If you might *try* going back, then you have reintroduced
backtracking, which would cause potential run-time exponential growth
for savings in the state graph.

--
Daniel Pitts' Tech Blog: <http://virtualinfinity.net/wordpress/>
From: Torben �gidius Mogensen on
"Piotr Wyderski" <piotr.wyderski(a)mothers.against.spam.gmail.com> writes:

> Torben Agidius Mogensen wrote:
>
>> Since any DFA is also a 2-DFA, you can use the normal NFA to DFA
>> contruction.
>
> Well, of course it would be valid by definition, but the point is to skip
> the potentially exponential state space explosion during the process
> of conversion from NFA to DFA. If there is no algorithmic advantage
> in constructing 2-DFAs, then they are only of academic interest. But
> I believe there is huge potential to be explored...

You can escape exponential explosion in some cases, but there are still
many NFAs that have exponentially larger 2DFAs. So you won't escape
exponential explosion by using 2DFAs.

Torben