edu.stanford.nlp.optimization
Class MinimizationExample
java.lang.Object
|
+--edu.stanford.nlp.optimization.MinimizationExample
- public class MinimizationExample
- extends Object
Example of using the minimization classes. The class contents is included
here in the documentation to make the example easy to see.
public class MinimizationExample {
// Sum of squares objective
private static class SumSquaresFunction implements DiffFunction {
public int domainDimension() { return 4; }
public double valueAt(double[] x) {
double sum = 0.0;
for (int i=0; i= 0.3
ineqConstraints[1] = new ComponentMinimumConstraint(2,0.5); // x2 >= 0.5
DiffFunction[] noConstraints = new DiffFunction[0];
double[] initial = {0.25, 0.25, 0.25, 0.25}; // note that it's not actual feasible
// unconstrained minimization
double[] x = unconstrainedMinimizer.minimize(objective, 1e-4, initial);
System.out.println("Unconstrained minimum is : "+objective.valueAt(x)+" at "+arrayToString(x));
// with equality constraints
double[] xEQ = constrainedMinimizer.minimize(objective, 1e-4, eqConstraints, 1e-8, noConstraints, 0.0, initial);
System.out.println("Equality-conconstrained minimum is : "+objective.valueAt(xEQ)+" at "+arrayToString(xEQ));
// with inequality constraints
double[] xINEQ = constrainedMinimizer.minimize(objective, 1e-4, noConstraints, 0, ineqConstraints, 1e-8, initial);
System.out.println("Inequality-conconstrained minimum is :"+objective.valueAt(xINEQ)+" at "+arrayToString(xINEQ));
// with all constraints
double[] xBOTH = constrainedMinimizer.minimize(objective, 1e-4, eqConstraints, 1e-8, ineqConstraints, 1e-8, initial);
System.out.println("Mixed-conconstrained minimum is : "+objective.valueAt(xBOTH)+" at "+arrayToString(xBOTH));
}
}
- Since:
- 1.0
Methods inherited from class java.lang.Object |
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait |
MinimizationExample
public MinimizationExample()
main
public static void main(String[] args)
Stanford NLP Group