Adding the ERP starter which allows running the external runtime problem directly from the console. And some minor cosmetics.

src/eva2/server/go/problems/AbstractProblemDoubleOffset.java

@@ -11,7 +11,7 @@ public abstract class AbstractProblemDoubleOffset extends AbstractProblemDouble
 
 //	protected AbstractEAIndividual      m_OverallBest       = null;
     protected int                       m_ProblemDimension  = 10;
-    protected double                    m_XOffSet           = 0.0;
+    protected double                    m_XOffSet           = 0.0; // TODO make them private, implement eval() and create abstract evalWithoutOffsets
     protected double                    m_YOffSet           = 0.0;
 //    protected boolean                   m_UseTestConstraint = false;
 

src/eva2/server/go/problems/ERPStarter.java

@@ -0,0 +1,171 @@
+package eva2.server.go.problems;
+
+import eva2.OptimizerFactory;
+import eva2.OptimizerRunnable;
+import eva2.client.EvAClient;
+import eva2.gui.BeanInspector;
+import eva2.server.go.individuals.AbstractEAIndividual;
+import eva2.server.go.operators.crossover.CrossoverESDefault;
+import eva2.server.go.operators.moso.MOSONoConvert;
+import eva2.server.go.operators.moso.MOSOWeightedFitness;
+import eva2.server.go.operators.mutation.MutateESRankMuCMA;
+import eva2.server.go.operators.selection.SelectBestIndividuals;
+import eva2.server.go.operators.terminators.EvaluationTerminator;
+import eva2.server.go.populations.InterfaceSolutionSet;
+import eva2.server.go.populations.Population;
+import eva2.server.go.strategies.GeneticAlgorithm;
+import eva2.server.go.strategies.InterfaceOptimizer;
+import eva2.server.modules.GOParameters;
+import eva2.server.stat.StatisticsStandalone;
+import eva2.tools.BasicResourceLoader;
+import eva2.tools.StringTools;
+import eva2.tools.ToolBox;
+import eva2.tools.math.Mathematics;
+
+public class ERPStarter {
+	
+	/**
+	 * Start an external runtime problem with some basic configs. The
+	 * default variant uses equally weighs all objectives through an equally weighted sum.  
+	 * Arguments: --csv <path-to-csv-config-file> --cmd <command-with-full-path> --maxEvals <maxEvals> [--multiObjective] [--gui]
+	 * @param args
+	 */
+	public static void main(String[] args) {
+		ExternalRuntimeProblem erp;
+		boolean useMultiObjective=false, startGUI=false;
+		// define the number of evaluations
+		int maxEvals = 5000;
+		// prefix for data output file - set null to deactivate
+		String outputFilePrefix = "erpTest";
+		
+		/** Argument handling ****************/
+		String[] keys= new String[]{"--gui", "--multiObjective", "--csv", "--cmd", "--maxEvals"};
+		int[] arities = new int[]{0, 0, 1, 1, 1};
+		Object[] values = new Object[6];
+		
+		Integer[] unknownArgs = StringTools.parseArguments(args, keys, arities, values, true);
+		
+		if (unknownArgs.length>0 || (values[2]==null) || (values[3]==null) || (values[4]==null)) {
+			System.err.println("Missing or unrecognized command line options: ");
+			for (int i=0; i<unknownArgs.length; i++) System.err.println("   " + args[unknownArgs[i]]);
+			System.err.println("Use with: --csv <path-to-csv-config-file> --cmd <command-with-full-path> --maxEvals <maxEvals> [--multiObjective] [--gui]");
+			return;
+		}
+		System.out.println("Parsing file: " + values[2]);
+		erp = parseCSV((String)values[2]);
+		String cmd = (String)values[3];
+		
+		maxEvals = Integer.parseInt((String)values[4]);
+		System.out.println("Setting maxEvals: " + maxEvals);
+		useMultiObjective=(values[1]!=null);
+		startGUI=(values[0]!=null);
+		// set target function and working directory
+		System.out.println("Setting target function command: " + cmd);
+		erp.setCommand(cmd);
+		int indexOfLastSlash = cmd.lastIndexOf(System.getProperty("file.separator"));
+		erp.setWorkingDirectory(cmd.substring(0, indexOfLastSlash));
+		
+		/******************/
+		// create optimization instance
+		InterfaceOptimizer opt = null;
+		if (useMultiObjective) {
+			// this uses a multi-objective strategy
+			erp.setMosoConverter(new MOSONoConvert());
+			opt = OptimizerFactory.createMultiObjectiveEA(new GeneticAlgorithm(), 50, erp, null);
+		} else {
+			// this uses a single objective (5,20)-CMA-ES with equally weighted criteria
+//			erp.setMosoConverter(new MOSOWeightedFitness());
+			opt = OptimizerFactory.createEvolutionStrategy(5, 20, false, new MutateESRankMuCMA(), 1., new CrossoverESDefault(), 0., new SelectBestIndividuals(), erp, null);
+		}
+
+		System.out.println("Optimizer instance: " + BeanInspector.toString(opt));
+		System.out.println("Problem instance: " + BeanInspector.toString(erp));
+		//	Instantiate optimization
+		GOParameters params = new GOParameters(opt, erp, new EvaluationTerminator(maxEvals));
+		if (startGUI) {
+			EvAClient.initClientGUI(params, null, null, null);
+		} else {
+			OptimizerRunnable rnbl = new OptimizerRunnable(params, new StatisticsStandalone(outputFilePrefix, 1, 3, true), false);
+			// actually start the optimization
+			OptimizerFactory.optimize(rnbl);
+
+			// some data output
+			InterfaceSolutionSet solSet = opt.getAllSolutions();
+			Population sols = solSet.getSolutions();
+			System.out.println("*** Solutions found: " + sols.size());
+			System.out.println("Best allover solution: " + AbstractEAIndividual.getDefaultStringRepresentation((AbstractEAIndividual)rnbl.getResult()));
+			System.out.println("Best last solution: " + AbstractEAIndividual.getDefaultStringRepresentation(solSet.getCurrentPopulation().getBestEAIndividual()));
+			System.out.println("Last solution set:");
+			for (int i=0; i<sols.size(); i++) {
+				System.out.println(AbstractEAIndividual.getDefaultStringRepresentation(sols.getEAIndividual(i)));
+			}
+		}
+	}
+	
+	public static ExternalRuntimeProblem parseCSV(String fname) {
+//		return parseCSV(fname, 5, 7, 2, 4, 3, 9); //
+		//String fname, int isVariableColIndex, int isObjectiveColIndex, 
+		//int lowerBoundIndex, int upperBoundIndex, int initialPosColIndex, int fitWeightsColIndex) {
+		return parseCSV(fname, 5, 2, 4, 3, -1, 7, 9);
+		//public static ExternalRuntimeProblem parseCSV(String fname, int isVariableColIndex, 
+		//int lowerBoundIndex, int upperBoundIndex, int initialPosColIndex, int initialPosBoxLenColIndex, 
+		//int isObjectiveColIndex, int fitWeightsColIndex) {
+	}
+	/**
+	 * Parse a csv file as in the example. Reads columns 2,3,4,5 where counting starts at zero.
+	 * Checks for the column entries at index 5 to contain a value of 1. These columns are interpreted as
+	 * target function variables with boundaries and seed. Note that all boundary values must contain
+	 * valid numbers. If a single seed value is empty, no initialization range is used.
+	 * 
+	 * @param fname name of the csv data file
+	 * @return an instance of ExternalRuntimeProblem
+	 */
+	public static ExternalRuntimeProblem parseCSV(String fname, int isVariableColIndex, int lowerBoundIndex, 
+			int upperBoundIndex, int initialPosColIndex, int initialPosBoxLenColIndex, 
+			int isObjectiveColIndex, int fitWeightsColIndex) {
+		double defaultInitialBoxLenRatio = 0.05; // by default, the init range is 5% of the domain range in a component 
+		// this assumes that data colums are as in the example csv
+		// parse only columns with index 2-9. Note that these will be reindexed to 0,1,2,3
+		int[] filterCols = new int[7];
+		filterCols[0]=isVariableColIndex;
+		filterCols[1]=lowerBoundIndex;
+		filterCols[2]=upperBoundIndex;
+		filterCols[3]=initialPosColIndex;
+		filterCols[4]=initialPosBoxLenColIndex;
+		filterCols[5]=isObjectiveColIndex;
+		filterCols[6]=fitWeightsColIndex;
+//		double[][] dat = BasicResourceLoader.loadDoubleData(fname, null, ";", 1, -1, new int[]{2,3,4,5,6,7,8,9});
+		double[][] dat = BasicResourceLoader.loadDoubleData(fname, null, ";", 1, -1, filterCols);
+//		System.out.println(BeanInspector.toString(dat));
+		double[][] filteredVars = ToolBox.filterBy(dat, 0, 1., 1.); // filter only those with a value of one in column isVariable
+		double[][] range = ToolBox.getCols(filteredVars, 1, 2); // get columns with lower and upper bound
+		double[][] initPos = ToolBox.getCols(filteredVars, 3); // get column with initial position center
+		double[][] initBoxLen = ToolBox.getCols(filteredVars, 4); // get column with initialization range box length
+		double[][] filteredObjectives = ToolBox.filterBy(dat, 5, 1., 1.); // filter those with a value of one in isObjective
+
+		ExternalRuntimeProblem erp = new ExternalRuntimeProblem();
+		if (filteredObjectives!=null && (filteredObjectives.length>1)) {
+			double[][] weights = ToolBox.getCols(filteredObjectives, 6);
+			if (Mathematics.areFinite(weights)<0) erp.setMosoConverter(new MOSOWeightedFitness(weights));
+		}
+		erp.setProblemDimension(filteredVars.length);
+		erp.setRange(range);
+//		erp.setInitialRange(range);
+		// produce an initial range around the seed position if available, otherwise use domain range
+		double[][] initialRange = new double[filteredVars.length][2];
+		for (int i=0; i<filteredVars.length; i++) {
+			if (Mathematics.isFinite(initPos[i][0])) { // if the init pos is valid (non NaN)...
+				double dv = defaultInitialBoxLenRatio*(range[i][1]-range[i][0]); // default box length for initial interval
+				if (Mathematics.isFinite(initBoxLen[i][0])) dv = initBoxLen[i][0]; // or box length defined in the file
+				if (!Mathematics.isInRange(initPos[i][0], range[i][0], range[i][1])) System.err.println("Warning: initial seed is not in range in dim. " + i + " when parsing from " + fname);
+				initialRange[i][0]=initPos[i][0]-dv;
+				initialRange[i][1]=initPos[i][0]+dv;
+			} else {
+				initialRange[i][0]=range[i][0];
+				initialRange[i][1]=range[i][1];
+			}
+		}
+		erp.setInitialRange(initialRange);
+		return erp;
+	}
+}

src/eva2/server/go/strategies/ParticleSwarmOptimization.java

@@ -997,6 +997,7 @@ public class ParticleSwarmOptimization implements InterfaceOptimizer, java.io.Se
 			for (int x = -this.m_TopologyRange; x <= this.m_TopologyRange; x++) {
 				for (int y = -this.m_TopologyRange; y <= this.m_TopologyRange; y++) {
 					tmpIndex = index + x + (y*corner);
+					if (wrapTopology) tmpIndex=(tmpIndex + pop.size())% pop.size(); // wrap the grid toroidal
 					if ((x != index) && (tmpIndex >= 0) && (tmpIndex < pop.size())) {
 						this.compareAndSetAttractor(localBestFitness, localBestPosition, (AbstractEAIndividual)pop.get(tmpIndex), useHistoric);
 					}
@@ -1820,7 +1821,7 @@ public class ParticleSwarmOptimization implements InterfaceOptimizer, java.io.Se
 		// tree, hpso
 //		GenericObjectEditor.setShowProperty(cls, "treeBranchDegree", (topology==PSOTopologyEnum.tree) || (topology==PSOTopologyEnum.hpso));
 		// linear
-		GenericObjectEditor.setShowProperty(cls, "wrapTopology", (topology==PSOTopologyEnum.linear));
+		GenericObjectEditor.setShowProperty(cls, "wrapTopology", (topology==PSOTopologyEnum.linear) || (topology==PSOTopologyEnum.grid));
 		// dms
 		GenericObjectEditor.setShowProperty(cls, "dmsRegroupGens", (topology==PSOTopologyEnum.dms));
 	}
@@ -1888,7 +1889,10 @@ public class ParticleSwarmOptimization implements InterfaceOptimizer, java.io.Se
 		m_Show = show;
 		if (!show) m_Plot = null;
 	}
-
+	public String showTipText() {
+		return "Activate for debugging in 2D";
+	}
+	
 	/**
 	 * @return the checkSpeedLimit
 	 **/
@@ -1929,7 +1933,10 @@ public class ParticleSwarmOptimization implements InterfaceOptimizer, java.io.Se
 	public void setSleepTime(int sleepTime) {
 		this.sleepTime = sleepTime;
 	}
-
+	public String sleepTimeTipText() {
+		return "Sleep for a time between iterations - to be used with debugging and the show option.";
+	}
+	
 	public double getSubSwarmRadius() {
 		return m_swarmRadius;
 	}