Start | Index | Overview • PREV | UP | NEXT

net.simplace.client.simulation.lap.generators.LHSSimGenerator

Generate simulations with parameters created using random Latin hypercube sampling

Example

Three dimensional hypercube with 1600 samples:

|  <generator id="LHSSG" class="net.simplace.client.simulation.lap.generators.LHSSimGenerator">
|    <cal id="par_a" datatype="DOUBLE" min="0.3" max="1" target="par_a"/>
|    <cal id="par_b" datatype="DOUBLE" min="0.012" max="0.024" target="par_b"/>
|    <cal id="par_c" datatype="DOUBLE" min="0.1" max="1" target="par_c"/>
|    <input id="samples" datatype="INT">1600</input>
|  </generator>

References

LHS and permutation functions by Le Minh Nghia, NTU-Singapore (Google code link)

public class LHSSimGenerator extends
net.simplace.simulation.control.generators.DefaultSimulationGenerator {
// Public Constructors
public LHSSimGenerator(Element aGeneratorElement, FWSimSession aSession,
HashMap aFieldMap, FWSimVariable.CONTENT_TYPE aContentType);

// Public Instance Methods

public HashMap createVariables(); // Overrides
net.simplace.simulation.control.generators.DefaultSimulationGenerator

public int[][] getSimpleLH(int N, int dim);

Generate matrix (nSamples x nDim) so that each column is a permutation of 1->N

N Number of samples
dim Number of dimensions

public void generate(FWSimulationControlCache aCache, ResultSet aResultSet);
// Overrides
net.simplace.simulation.control.generators.DefaultSimulationGenerator

}

Hierarchy: java.lang.Object - net.simplace.simulation.control.FWSimulationControlContainer (net.simplace.simulation.util.FWSimFieldContainer) - net.simplace.simulation.control.FWSimulationGenerator - net.simplace.simulation.control.generators.DefaultSimulationGenerator - LHSSimGenerator