[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Help-glpk] FOR HELP! - Specifying initial solution

From: glpk xypron
Subject: Re: [Help-glpk] FOR HELP! - Specifying initial solution
Date: Tue, 02 Dec 2008 19:04:40 +0100

Hello Andrew,

> >> Yes. Ampl allows specifying an initial estimation of variables in the
> >> same way as computable values for parameters, e.g.
> >> For mip models there might be additional attributes of variables and
> >>constraints, for example, branching priority, etc.
> I have looked at other languages:
> Each variable has 4 attributes:
> .lo = lower bound
> .l = level or primal value
> .up = upper bound
> .m = marginal or dual value
> which can be set like this:
> x.lo(i,j) = 10.0 ;
> .l is used as a guess for the optimum of the constrained problem, the
> initial value is 0.0 unless zero is not within the bounds, else it takes
> the
> value of the bound closesest to zero.

I have had a close look at GAMS:

GAMS 22.8 allows the following values to be specified and output for variables:
.l      level
.m      marginal (output only)
.lo     lower bound
.up     upper bound
.scale  scale value
.prior  priority for discrete variables
.fx     shorthand for setting .l,.lo and .up to the same value

For constraints the following can be output:
.l      level
.m      marginal 
.lo     lower bound
.up     upper bound

.m is the amount by which the objective function would change if the equation 
level were moved one unit
.l is the left hand side of the equation
.up is the upper bound for the left hand side of the equation (= right hand 
side or +infinity)
.lo is the lower bound for the left hand side of the equation (= right hand 
side or -infinity)

(.l - .up) or (.l - .lo) is the solution error of the equation

If GMPL could output marginal costs in post processing these values could be 
made available for post processing using the table OUT statement.

Best regards


Psssst! Schon vom neuen GMX MultiMessenger gehört? Der kann`s mit allen:

reply via email to

[Prev in Thread] Current Thread [Next in Thread]