Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: Branch & bound for not everywhere defiend constraints



> Date: Thu, 3 Sep 2009 12:53:16 -0400
> From: Nate Hayes <nh@xxxxxxxxxxxxxxxxx>
> Subject: Re: Branch & bound for not everywhere defiend constraints
> To: STDS-1788@xxxxxxxxxxxxxxxxx
> 
> Dan Zuras Intervals wrote:
> > The algorithm could be made more efficient by using
> > many NaIs.  Or by using none as I suggested.
> 
> I don't entirely follow what you've suggested. For example, any assertion 
> about the empty set is always vacuously true. So if f(X) is empty, then every 
> if-then clause in
> 
> If f(X) > 0 color me gray
> else if 0 \in f(X) color me blue
> else if width(f(X)) == +oo color me red
> else if f(X) \subset [-oo,0] color me green
> else if empty(f(X)) color me yellow
> 
> is true. For example, f(X) > 0 is true if f(X) is empty. So I don't see how any 
> yellow boxes will show up in your paving.
> 
> Nate Hayes

	Ah, you spotted that. :-) :-) :-)

	Yes, I noticed it just after I posted it.

	I believe (one of) the correct sequence(s) is:

		if empty(f(X)) color me yellow
		else if f(X) > 0 color me gray
		else if 0 \in f(X) color me blue
		else if width(f(X)) == +oo color me red
		else if f(X) \subset [-oo,0] color me green

	And still recurse on red & blue.

	I think the logical constraints are yellow before
	everyone (as you point out), both red & blue (in
	either order) before green, & which leaves gray to
	go anywhere after yellow.

	If the domain of your search is expanded to include
	a neighborhood of {x0 = 0, x1 = 0} then there is a
	slight behavioral difference near the beginning of
	the search if gray follows red but I don't think it
	invalidates the results.

	As I said, no computers were harmed in this process
	so even this order may have problems.

	What do you think now?  Am I there yet?

	That's what I get for wetware compiling. :-)

	Enjoy,

				Dan