A java-based implementation of Stochastic Ranking Evolutionary Strategy (SRES).
In a nutshell, this is a heuristic algorithm that finds a set of parameters Xi that minimize/maximize the value of a given function aka objective F(Xi) and does not violate the constraints specified in form of Gk(Xi)<=0 inequalities.
For more details, please see the Runarsson TP and Yao X paper which has served as a basis for implementation of this algorithm. In my experience, this is one of the best algorithms
for solving non-linear real multivariate function constrained optimization problems. Note that this implementation is optimized by calculating the objective and constraints in parallel over the population of evolving solutions.
Java 1.8 is required for this library. In order to use, compile using
Maven and add the following dependency to your pom.xml file:
<dependency>
<groupId>com.antigenomics</groupId>
<artifactId>jsres</artifactId>
<version>1.0.0</version>
</dependency>The following code demonstrates how to specify the objective function and constraints:
import static java.lang.Math.*;
...
Objective myObjective = new Objective(
new double[]{0, 0}, // lower bounds of the search space
new double[]{10, 10}, // upper bounds of the search space
true // this is a maximization problem
) {
@Override
public Result evaluate(double[] features) {
// objective, first and second constraint values
double f = pow(sin(2 * PI * features[0]), 3) * sin(2 * PI * features[1]) /
(features[0] * features[0] * features[0] * (features[0] + features[1])),
g1 = features[0] * features[0] - features[1] + 1,
g2 = 1 - features[0] + (features[1] - 4) * (features[1] - 4);
return new Result(f, new double[]{g1, g2});
}
}The following code can be used to execute the SRES algorithm:
int numberOfGenerations = 1000; // run for 1000 generations
SRES sres = new SRES(myObjective); // set objective
SRES.SolutionSet solutionSet = sres.run(numberOfGenerations); // holds solutions after last SRES generationAccessing the results is performed as follows:
SRES.Solution bestSolution = solutionSet.getBestSolution();
System.out.println(solution.toString()); // print the best solution
solution.getFeatures(); // optimal objective parameters
solution.getObjectiveResult().getValue(); // objective function value at optimum
solution.getObjectiveResult().getConstraintValues(); // values of constraints at optimum