Re: [eclipse-clp-users] ECLiPSe - Real variables domain propagation problem

Ah I see, I should have read the documentation more...
And yes sorry, I meant branch_and_bound.

Thanks everyone for your help and apologies for the questions!

Thanks again,
Wayne

On Mon, Jul 27, 2009 at 4:22 PM, Kish Shen <kisshen_at_...5...> wrote:

> Wayne Mac Adams wrote:
>
>> It might be a seperate issue but I am getting an unexpected result from
>> the
>> following query also (again assuming ic loaded)
>>
>> ?- [A, B] :: 0.0 .. 0.5, C $= A + B, Cost $= -(C), minimize((locate([A, B,
>> C], 1.0), get_min(Cost, Min)), Min).
>> A = A{0.0 .. 0.5}
>> B = B{0.0 .. 0.5}
>> C = C{0.0 .. 0.5}
>> Cost = Cost{-0.5 .. -0.0}
>> Min = -0.5
>> There are 2 delayed goals.
>> Yes (0.00s cpu)
>>
>> Surely the maximum for C is 1.0, and there should be a solution with Cost
>> -1.0?
>>
>> Am I simply using minimize wrong, or is there another problem?
>>
>>  Hi Wayne,
>
> There is no minimize/2 in lib(ic), and if you were using
> lib(branch_and_bound)'s minimize/2, you have not shown what it prints out
> with the query you gave:
> Found a solution with cost -0.5
> Found no solution with cost -1.0Inf .. -1.5
>
> notice that the interval between -0.5 and -1.5 was not covered -- this
> should give you the clue to the problem.
>
> If you read the documentation, you will see that miimize/2 is a short-hand
> for bb_min/3, with no options (i.e. run everything with the default). One of
> the option is delta, i.e. the minimum absolute improvement required for each
> step, and the default value is 1.0. Thus,
> after finding a solution of -0.5, it will only try to find a solution that
> is at least -1.5, so there is no better solution is found.
>
> Change your query to use bb_min, with a delta option, e.g.
>
> [A,B]:: 0.0..0.5, C $= A + B, Cost $= -C, bb_min((locate([A,B,C], 1.0),
> get_min(Cost, Min)), Min, bb_options{delta:0.5}).
>
> and this will give you Min = -1.0 `solution' [as Joachim pointed out, this
> is strictly not a solution, because you still have delayed goals and
> uninstantiated variables]
>
>
> Cheers,
>
> Kish
>
> --
> This e-mail may contain confidential and privileged material for the
> sole use of the intended recipient. Any review, use, distribution or
> disclosure by others is strictly prohibited. If you are not the intended
> recipient (or authorized to receive for the recipient), please contact
> the sender by reply e-mail and delete all copies of this message.
> Cisco Systems Limited (Company Number: 02558939), is registered in
> England and Wales with its registered office at 1 Callaghan Square,
> Cardiff, South Glamorgan CF10 5BT.
>