List Info

Thread: RE: Fw: lpsolve problem




RE: Fw: lpsolve problem
country flaguser name
United States
2008-07-12 11:42:35

0 put the colon after max:. ; so it is a proper lp format file

1 realize that it is poorly posed because of the 1e10 range of coefficients.

2 believe the UNBOUNDED solution given by most option configurations.

3 find a solver that is willing to do exact rational arithmetic. http://members.jcom.home.ne.jp/masashi777/exlp.html

4 solve it by hand via row elimination.

5 Re scale S1, S2 and L2 for convenience

6 try to give the model some freedom to keep away from critical points at the boundary of feasible/infeasible

  In this case the O1 = 1; .

 

//max: 1 * S1 + 1 * S2 ;

max: 0.5 * S1 + 0.5e-10 * S2 ;

2 * S1 + 2 * L1 + 2 * L2 + -2 * O1 = 0 ;

-2e10 S2 + 2e10 * L1 + 2 * L2 = 2e10 ;

//1 * O1 = 1 ;

  L2 > .001 * O1;

  O1 < 1e12;

//  O1 < 1.01;

 

William


From: lp_solveyahoogroups.com [mailto:lp_solveyahoogroups.com] On Behalf Of masoum rajabi
Sent: Saturday, July 12, 2008 2:04 AM
To: lp_solveyahoogroups.com
Subject: [lp_solve] Fw: lpsolve problem

 

 

Hi,

lpsolve fails to solve the following model. but some times it will be solved with changiing some default parameter's value and some times no.what can I do?

 
max 1 * S1 + 1 * S2 ;
1 * S1 + 2 * L1 + 1 * L2 + -2 * O1 = 0 ;
-1 * S2 + 20000000000 * L1 + 1 * L2 = 20000000000 ;
1 * O1 = 1 ;

 

I'm looking forward for your answer,

thanks

 

__._,_.___
.

__,_._,___
[1]

about | contact  Other archives ( Real Estate discussion Medical topics )