StochasticDifferentialEquationModel.java

package dr.evomodel.continuous;

/**
 * @author Marc Suchard
 *         <p/>
 *         Provides a numerical approximation to a SDE solution for a diffusion model.
 *         Approximation is via the Euler-Maruyama method.
 *         <p/>
 *         Projected uses:
 *         - Diffusion on a sphere
 *         - Incorporating geographic features
 */
public class StochasticDifferentialEquationModel extends MultivariateDiffusionModel {

    private static double maxTimeIncrement = 1E-3;
    private static int defaultNumberSteps = 100;

    protected double calculateLogDensity(double[] start, double[] stop, double time) {

        int numSteps = defaultNumberSteps;
        if (time / numSteps > maxTimeIncrement) {
            numSteps = (int) (time / maxTimeIncrement);
        }

        return 0;
    }

    private double[] getDriftVector(double[] X) {
        return null;
    }

    private double[] getVarianceMatrix(double[] X) {
        return null;
    }

}