Re: [ojAlgo-user] Problem with mixed integer linear programming

[Anders P. <an...@op...>]

Haven't looked at this in great detail, but if you change:

	objective[1].setInteger(true);
to
	objective[1].setInteger(false);

and

	model.minimise();
to
        model.setMinimisation(true);
        final OptimisationSolver tmpSolver = model.getDefaultSolver();
        final Result tmpResult = tmpSolver.solve();

Then you'll see that ojAlgo reports the relaxed problem to be unbounded. If that's true it's a (big) problem for the integer solver. The relaxed problems need to be able to be solved by the linear or quadratic solvers.

I'm not confident that ojAlgo reports INFEASIBLE and/or UNBOUNDED problems right all the time, but you should definitely not get a NullPointerException!

X=0, Y=0 & Z=0 is a valid solution to your problem, right? In that case there is a bug with the LinearSolver.

/Anders


On 10 okt 2011, at 12:18, Julien.Cojan wrote:

> Hello,
> 
> I am using OjAlgo to solve mixed linear programming problems.
> However I often get a NullPointerException when running on my examples.
> I tried version 30.0 and 30.6, I get the same error.
> 
> This is the smallest example I could find where I get this exception :
> 
> 
> import static org.ojalgo.constant.BigMath.ONE;
> import static org.ojalgo.constant.BigMath.ZERO;
> 
> import java.math.BigDecimal;
> 
> import org.ojalgo.optimisation.Expression;
> import org.ojalgo.optimisation.Variable;
> import org.ojalgo.optimisation.linear.LinearExpressionsModel;
> 
> 
> public class TestOjAlgo {
> 
> 	public static void testBug1(){
> 		Variable[] objective = new Variable[] {
> 				new Variable("X").weight(ONE),
> 				new Variable("Y").weight(ZERO),
> 				new Variable("Z").weight(ZERO)};
> 
> 		objective[1].setInteger(true);
> 
> 		LinearExpressionsModel model = new LinearExpressionsModel(objective);
> 
> 		// c1: X =0
> 		Expression c1 = model.addEmptyLinearExpression("c1");
> 		c1.level(ZERO);
> 		c1.setLinearFactor(0, ONE);
> 
> 		// c2: -X +5Y =0
> 		Expression c2 = model.addEmptyLinearExpression("c2");
> 		c2.level(ZERO);
> 
> 		c2.setLinearFactor(0, new BigDecimal(-1));
> 		c2.setLinearFactor(1, ONE);
> 
> 		// c3: X -Z =0
> 		Expression c3 = model.addEmptyLinearExpression("c3");
> 		c3.level(ZERO);
> 		// bugs with this constraint
> 		c3.setLinearFactor(0, ONE);
> 		c3.setLinearFactor(2, new BigDecimal(-1));
> 		// but not with this one ???
> //		c3.setLinearFactor(0, new BigDecimal(-1));
> //		c3.setLinearFactor(2, ONE);
> 
> 		model.minimise();
> 	}
> 
> 	public static void main(String[] args){
> 		testBug1();
> 	}
> 
> }
> 
> 
> 
> 
> 
> I get the following message
> Exception in thread "main" java.lang.NullPointerException
> 	at org.ojalgo.optimisation.OptimisationSolver$Result.<init>(OptimisationSolver.java:99)
> 	at org.ojalgo.optimisation.integer.IntegerSolver.solve(IntegerSolver.java:318)
> 	at org.ojalgo.optimisation.ExpressionsBasedModel.solve(ExpressionsBasedModel.java:741)
> 	at org.ojalgo.optimisation.ExpressionsBasedModel.minimise(ExpressionsBasedModel.java:458)
> 	at TestOjAlgo.testBug1(TestOjAlgo.java:57)
> 	at TestOjAlgo.main(TestOjAlgo.java:61)
> 
> 
> Is it a bug or am I not using ojAlgo properly ?
> 
> 
> -- 
> The Open University is incorporated by Royal Charter (RC 000391), an exempt charity in England & Wales and a charity registered in Scotland (SC 038302).
> 
> 
> ------------------------------------------------------------------------------
> All the data continuously generated in your IT infrastructure contains a
> definitive record of customers, application performance, security
> threats, fraudulent activity and more. Splunk takes this data and makes
> sense of it. Business sense. IT sense. Common sense.
> http://p.sf.net/sfu/splunk-d2dcopy1
> _______________________________________________
> ojAlgo-user mailing list
> ojA...@li...
> https://lists.sourceforge.net/lists/listinfo/ojalgo-user
> 
>