Re: [Help-glpk] multiplication of linear forms not allowed

[Jorge Tavares]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi,

On Nov 30, 2005, at 17:34 , Mihai Banciu wrote:

> I suspect that you may be wrong on one count. a[i] shouldn't be a
> variable, but a parameter. Basically, what yo need to do is solve  
> the LP
> relaxation of the original MKP, get the dual variables a[i] associated
> with the solution, and use those duals as *FIXED* parameters in the
> surrogate problem!. This way, in the surrogate, the only variables are
> x[j] and your problem is linear, since a[i]'s are fixed.

You're right, I was mistaken about a[i]! The paper I was reading that  
showed this didn't explain it very well and I tought a[i] and x[i]  
were to be determined at the same time. I got it wrong. Thanks very  
much for the clarification.

Best regards,
Jorge

- --
Jorge Tavares
University of Coimbra | http://eden.dei.uc.pt/~jast

"Sometimes the appropriate response to reality is to go insane."



-----BEGIN PGP SIGNATURE-----

iD8DBQFDje4qKHEUoQoCFcIRAq0AAKCgCS+inBZRYoT3vJSSZb3f3lS9/wCfXCv5
A3Z/iPXVdJZcmVeIlM6m1pM=
=mavn
-----END PGP SIGNATURE-----