halfdan/eva2
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Merge branch '30-dynamic-cli'

Browse Source
This commit is contained in:
Fabian Becker
2014-11-02 14:03:59 +01:00
parent ee73f5dbe5 edaa58bd5a
commit 97d79612c2
12 changed files with 236 additions and 805 deletions
Download Patch File Download Diff File Expand all files Collapse all files

21
pom.xml
View File

@@ -27,27 +27,6 @@
<artifactId>javahelp</artifactId> <artifactId>javahelp</artifactId>
<version>2.0.05</version> <version>2.0.05</version>
</dependency> </dependency>
<dependency>
<groupId>org.apache.commons</groupId>
<artifactId>commons-math3</artifactId>
<version>3.1.1</version>
</dependency>
<dependency>
<groupId>commons-cli</groupId>
<artifactId>commons-cli</artifactId>
<version>1.2</version>
</dependency>
<dependency>
<groupId>com.google.code.gson</groupId>
<artifactId>gson</artifactId>
<version>2.3</version>
<scope>compile</scope>
</dependency>
<dependency>
<groupId>org.reflections</groupId>
<artifactId>reflections</artifactId>
<version>0.9.9</version>
</dependency>
</dependencies> </dependencies>

712
src/eva2/cli/Main.java
View File

@@ -1,719 +1,15 @@
package eva2.cli; package eva2.cli;
import com.google.gson.*; import eva2.optimization.go.InterfaceOptimizationParameters;
import eva2.OptimizerFactory;
import eva2.optimization.OptimizationStateListener;
import eva2.optimization.enums.DEType;
import eva2.optimization.enums.PSOTopology;
import eva2.optimization.population.InterfacePopulationChangedEventListener;
import eva2.optimization.modules.OptimizationParameters;
import eva2.optimization.operator.crossover.CrossoverESDefault;
import eva2.optimization.operator.crossover.InterfaceCrossover;
import eva2.optimization.operator.mutation.InterfaceMutation;
import eva2.optimization.operator.mutation.MutateDefault;
import eva2.optimization.operator.selection.InterfaceSelection;
import eva2.optimization.operator.selection.SelectXProbRouletteWheel;
import eva2.optimization.operator.terminators.CombinedTerminator;
import eva2.optimization.operator.terminators.FitnessValueTerminator;
import eva2.optimization.population.Population;
import eva2.problems.AbstractProblemDouble;
import eva2.problems.AbstractProblemDoubleOffset;
import eva2.optimization.strategies.DifferentialEvolution;
import eva2.optimization.strategies.InterfaceOptimizer;
import org.apache.commons.cli.*;
import org.reflections.Reflections;
import java.lang.reflect.Modifier;
import java.util.Arrays;
import java.util.Map;
import java.util.Set;
import java.util.TreeMap;
import java.util.logging.Level;
import java.util.logging.Logger;
/** /**
* Main Class for the EvA2 Command Line Interface * Created by becker on 01.11.2014.
*
* The command line interface features a limited subset of the EvA2
* optimization suite since it's difficult to parameterize all available
* classes in EvA2 from the command line.
* Supported Features:
* - Select all optimization problems that implement InterfaceOptimizationProblem
* - Select all optimizers that implement InterfaceOptimizer
* * Not all optimizers are configurable on the command line and will run
* with default parameters.
* * Optimizers can use the @Description / @Parameter annotations to specify
* parameters for CLI
* - Configure default parameters
* * Population size
* * Number of optimization runs (multi-runs)
* - Termination:
* * Not configurable!
* * Default: EvaluationTerminator(20000)
*/ */
public class Main implements OptimizationStateListener, InterfacePopulationChangedEventListener { public class Main {
private static Logger LOGGER = Logger.getLogger(Main.class.getName());
private int populationSize = 20;
private int numberOfRuns = 1;
private int dimension = 30;
private long seed = System.currentTimeMillis();
private AbstractProblemDoubleOffset problem;
private InterfaceOptimizer optimizer;
private InterfaceMutation mutator;
private InterfaceCrossover crossover;
private InterfaceSelection selection;
private JsonObject jsonObject;
private JsonArray optimizationRuns;
private JsonArray generationsArray;
private double[] fBias = { -4.5000000e+002, -4.5000000e+002, -4.5000000e+002, -4.5000000e+002, -3.1000000e+002,
3.9000000e+002, -1.8000000e+002, -1.4000000e+002, -3.3000000e+002, -3.3000000e+002, 9.0000000e+001,
-4.6000000e+002, -1.3000000e+002, -3.0000000e+002, 1.2000000e+002, 1.2000000e+002, 1.2000000e+002,
1.0000000e+001, 1.0000000e+001, 1.0000000e+001, 3.6000000e+002, 3.6000000e+002, 3.6000000e+002,
2.6000000e+002, 2.6000000e+002};
/**
* Creates a set of default options used in all optimizations.
*
* @return Options Default options used for optimizations
*/
private Options createDefaultCommandLineOptions() {
Options opt = new Options();
opt.addOption(OptionBuilder
.withLongOpt("optimizer")
.withDescription("Optimizer")
.hasArg()
.create("op")
);
opt.addOption("ps", "popsize", true, "Population size");
opt.addOption("n", "runs", true, "Number of runs to perform");
opt.addOption("s", "seed", true, "Random seed");
// Those two only make sense when used in an algorithm with mutation/crossover
opt.addOption("pc", true, "Crossover Probability");
opt.addOption("pm", true, "Mutation Probability");
opt.addOption("mutator", true, "Mutator Operator");
opt.addOption("crossover", true, "Crossover Operator");
opt.addOption("selection", true, "Selection Operator");
opt.addOption(OptionBuilder
.withLongOpt("help")
.withDescription("Shows this help message or specific help for [optimizer]")
.hasOptionalArgs(1)
.create('h')
);
opt.addOption(OptionBuilder
.withLongOpt("problem")
.withDescription("Select Optimization Problem to optimize.")
.hasArg()
.create('p')
);
opt.addOption("dim", true, "Problem Dimension");
return opt;
}
@Override
public void performedStop() {
LOGGER.info("Optimization stopped.");
}
@Override
public void performedStart(String infoString) {
LOGGER.info("Optimization started.");
}
@Override
public void performedRestart(String infoString) {
LOGGER.info("Optimization restarted.");
}
@Override
public void updateProgress(int percent, String msg) {
printProgressBar(percent);
}
public static void printProgressBar(int percent) {
StringBuilder bar = new StringBuilder("[");
for (int i = 0; i < 50; i++) {
if (i < (percent / 2)) {
bar.append("=");
} else if (i == (percent / 2)) {
bar.append(">");
} else {
bar.append(" ");
}
}
bar.append("] " + percent + "% ");
System.out.print("\r" + bar.toString());
}
public static Map<String, Class<? extends InterfaceOptimizer>> createOptimizerList() {
Map<String, Class<? extends InterfaceOptimizer>> optimizerList = new TreeMap<>();
Reflections reflections = new Reflections("eva2.optimization.strategies");
Set<Class<? extends InterfaceOptimizer>> optimizers = reflections.getSubTypesOf(InterfaceOptimizer.class);
for (Class<? extends InterfaceOptimizer> optimizer : optimizers) {
// We only want instantiable classes
if (optimizer.isInterface() || Modifier.isAbstract(optimizer.getModifiers())) {
continue;
}
optimizerList.put(optimizer.getSimpleName(), optimizer);
}
return optimizerList;
}
public static Map<String, Class<? extends AbstractProblemDoubleOffset>> createProblemList() {
Map<String, Class<? extends AbstractProblemDoubleOffset>> problemList = new TreeMap<>();
Reflections reflections = new Reflections("eva2.problems");
Set<Class<? extends AbstractProblemDoubleOffset>> problems = reflections.getSubTypesOf(AbstractProblemDoubleOffset.class);
for (Class<? extends AbstractProblemDoubleOffset> problem : problems) {
// We only want instantiable classes
if (problem.isInterface() || Modifier.isAbstract(problem.getModifiers())) {
continue;
}
problemList.put(problem.getSimpleName(), problem);
}
return problemList;
}
public static Map<String, Class<? extends InterfaceMutation>> createMutatorList() {
Map<String, Class<? extends InterfaceMutation>> mutationList = new TreeMap<>();
Reflections reflections = new Reflections("eva2.optimization.operator.mutation");
Set<Class<? extends InterfaceMutation>> mutators = reflections.getSubTypesOf(InterfaceMutation.class);
for (Class<? extends InterfaceMutation> mutator : mutators) {
// We only want instantiable classes
if (mutator.isInterface() || Modifier.isAbstract(mutator.getModifiers())) {
continue;
}
mutationList.put(mutator.getSimpleName(), mutator);
}
return mutationList;
}
public static Map<String, Class<? extends InterfaceCrossover>> createCrossoverList() {
Map<String, Class<? extends InterfaceCrossover>> crossoverList = new TreeMap<>();
Reflections reflections = new Reflections("eva2.optimization.operator.crossover");
Set<Class<? extends InterfaceCrossover>> crossovers = reflections.getSubTypesOf(InterfaceCrossover.class);
for (Class<? extends InterfaceCrossover> crossover : crossovers) {
// We only want instantiable classes
if (crossover.isInterface() || Modifier.isAbstract(crossover.getModifiers())) {
continue;
}
crossoverList.put(crossover.getSimpleName(), crossover);
}
return crossoverList;
}
public static Map<String, Class<? extends InterfaceSelection>> createSelectionList() {
Map<String, Class<? extends InterfaceSelection>> selectionList = new TreeMap<>();
Reflections reflections = new Reflections("eva2.optimization.operator.selection");
Set<Class<? extends InterfaceSelection>> selections = reflections.getSubTypesOf(InterfaceSelection.class);
for (Class<? extends InterfaceSelection> selection : selections) {
// We only want instantiable classes
if (selection.isInterface() || Modifier.isAbstract(selection.getModifiers())) {
continue;
}
selectionList.put(selection.getSimpleName(), selection);
}
return selectionList;
}
public static void showHelp(Options options) {
HelpFormatter helpFormatter = new HelpFormatter();
helpFormatter.printHelp("java -jar EvA2.jar", "Global Parameters", options, "", true);
}
public static void main(String[] args) { public static void main(String[] args) {
Main optimizationMain = new Main(args); InterfaceOptimizationParameters parameters = OptimizationBuilder.parseArguments(args);
optimizationMain.runOptimization();
}
public Main(String[] args) {
int index = Arrays.asList(args).indexOf("--");
String[] programParams = args;
String[] optimizerParams;
if (index >= 0) {
programParams = Arrays.copyOfRange(args, 0, index);
optimizerParams = Arrays.copyOfRange(args, index + 1, args.length);
} else {
optimizerParams = new String[]{};
}
this.jsonObject = new JsonObject();
this.optimizationRuns = new JsonArray();
this.generationsArray = new JsonArray();
/**
* Default command line options only require help or optimizer.
* Later we build extended command line options depending on
* the selected optimizer.
*/
Options defaultOptions = this.createDefaultCommandLineOptions();
/**
* Parse default options.
*/
CommandLineParser cliParser = new BasicParser();
CommandLine commandLine = null;
try {
commandLine = cliParser.parse(defaultOptions, programParams);
} catch (ParseException e) {
showHelp(defaultOptions);
System.exit(-1);
}
/**
* Process help and help sub pages.
*/
if (commandLine.hasOption("help")) {
String helpOption = commandLine.getOptionValue("help");
if (helpOption == null) {
showHelp(defaultOptions);
} else {
switch (helpOption) {
case "optimizer":
showOptimizerHelp();
break;
case "problem":
listProblems();
break;
default:
showHelp(defaultOptions);
break;
}
}
System.exit(0);
}
// OK, so we've got valid parameters - let's setup the optimizer and problem
if (commandLine.hasOption("popsize")) {
this.populationSize = Integer.parseInt(commandLine.getOptionValue("popsize"));
}
if (commandLine.hasOption("runs")) {
this.numberOfRuns = Integer.parseInt(commandLine.getOptionValue("runs"));
}
if (commandLine.hasOption("seed")) {
this.seed = Long.parseLong(commandLine.getOptionValue("seed"));
}
if (commandLine.hasOption("dim")) {
this.dimension = Integer.parseInt(commandLine.getOptionValue("dim"));
}
if (commandLine.hasOption("problem")) {
String problemName = commandLine.getOptionValue("problem");
setProblemFromName(problemName);
this.problem.setProblemDimension(this.dimension);
} else {
LOGGER.severe("No problem specified. Please specify a problem with '--problem'.");
System.exit(-1);
}
if (commandLine.hasOption("mutator")) {
String mutatorName = commandLine.getOptionValue("mutator");
try {
setMutatorFromName(mutatorName);
} catch (Exception ex) {
System.out.println(ex.getMessage());
System.exit(-1);
}
} else {
this.mutator = new MutateDefault();
}
if (commandLine.hasOption("crossover")) {
String crossoverName = commandLine.getOptionValue("crossover");
try {
setCrossoverFromName(crossoverName);
} catch (Exception ex) {
System.out.println(ex.getMessage());
System.exit(-1);
}
} else {
this.crossover = new CrossoverESDefault();
}
if (commandLine.hasOption("selection")) {
String selectionName = commandLine.getOptionValue("selection");
try {
setSelectionFromName(selectionName);
} catch (Exception ex) {
System.out.println(ex.getMessage());
System.exit(-1);
}
} else {
this.selection = new SelectXProbRouletteWheel();
}
// Depends on mutator/crossover/selection being set
if (commandLine.hasOption("optimizer")) {
String optimizerName = commandLine.getOptionValue("optimizer");
try {
createOptimizerFromName(optimizerName, optimizerParams);
} catch(Exception ex) {
System.out.println(ex.getMessage());
System.exit(-1);
}
}
this.jsonObject.addProperty("population_size", this.populationSize);
this.jsonObject.addProperty("number_of_runs", this.numberOfRuns);
this.jsonObject.addProperty("dimension", this.dimension);
this.jsonObject.addProperty("seed", this.seed);
JsonObject problemObject = new JsonObject();
problemObject.addProperty("name", this.problem.getName());
problemObject.addProperty("dimension", 30);
this.jsonObject.add("problem", problemObject);
}
private void setMutatorFromName(String mutatorName) {
Map<String, Class<? extends InterfaceMutation>> mutatorList = createMutatorList();
Class<? extends InterfaceMutation> mutator = mutatorList.get(mutatorName);
try {
this.mutator = mutator.newInstance();
} catch (InstantiationException e) {
e.printStackTrace();
} catch (IllegalAccessException e) {
e.printStackTrace();
}
}
private void setCrossoverFromName(String crossoverName) {
Map<String, Class<? extends InterfaceCrossover>> crossoverList = createCrossoverList();
Class<? extends InterfaceCrossover> crossover = crossoverList.get(crossoverName);
try {
this.crossover = crossover.newInstance();
} catch (InstantiationException e) {
e.printStackTrace();
} catch (IllegalAccessException e) {
e.printStackTrace();
}
}
private void setSelectionFromName(String selectionName) {
Map<String, Class<? extends InterfaceSelection>> selectionList = createSelectionList();
Class<? extends InterfaceSelection> selection = selectionList.get(selectionName);
try {
this.selection = selection.newInstance();
} catch (InstantiationException e) {
e.printStackTrace();
} catch (IllegalAccessException e) {
e.printStackTrace();
}
}
/**
* This method will create the various optimizers that are supported on the CLI.
* It's a really messy process since neither Java nor args4j/apache-cli can handle
* complex object parameters. The trick here is that all parameters after the
* double-dash (--) are treated as parameters for the optimization algorithm.
*
* @param optimizerName The name of the optimizer.
* @param optimizerParams The remaining command line parameters.
* @throws Exception
*/
private void createOptimizerFromName(String optimizerName, String[] optimizerParams) throws Exception {
Options opt = new Options();
CommandLineParser cliParser = new BasicParser();
CommandLine commandLine = null;
switch(optimizerName) {
case "DifferentialEvolution": {
opt.addOption("F", true, "Differential Weight");
opt.addOption("CR", true, "Crossover Rate");
opt.addOption("DEType", true, "DE Type ()");
/**
* Parse default options.
*/
try {
commandLine = cliParser.parse(opt, optimizerParams);
} catch (ParseException e) {
showHelp(opt);
System.exit(-1);
}
double f = 0.8, lambda = 0.6, cr = 0.6;
if (commandLine.hasOption("F")) {
f = Double.parseDouble(commandLine.getOptionValue("F"));
}
if (commandLine.hasOption("CR")) {
cr = Double.parseDouble(commandLine.getOptionValue("CR"));
}
this.optimizer = OptimizerFactory.createDifferentialEvolution(this.problem, this.populationSize, f, lambda, cr, this);
if (commandLine.hasOption("DEType")) {
((DifferentialEvolution)this.optimizer).setDEType(
DEType.getFromId(
Integer.parseInt(commandLine.getOptionValue("DEType"))
)
);
}
break;
}
case "GeneticAlgorithm": {
double pm = 0.01, pc = 0.5;
opt.addOption("pm", true, "Mutation Probability");
opt.addOption("pc", true, "Crossover Probability");
/**
* Parse default options.
*/
try {
commandLine = cliParser.parse(opt, optimizerParams);
} catch (ParseException e) {
showHelp(opt);
System.exit(-1);
}
if (commandLine.hasOption("pm")) {
pm = Double.parseDouble(commandLine.getOptionValue("pm"));
}
if (commandLine.hasOption("pc")) {
pc = Double.parseDouble(commandLine.getOptionValue("pc"));
}
this.optimizer = OptimizerFactory.createGeneticAlgorithm(mutator, pm, crossover, pc, selection, this.populationSize, this.problem, this);
break;
}
case "ParticleSwarmOptimization": {
double phi1 = 2.05, phi2 = 2.05, speedLimit = 0.1;
int topoRange = 2;
PSOTopology selectedTopology = PSOTopology.star;
opt.addOption("speedLimit", true, "Speed Limit");
opt.addOption("topology", true, "Particle Swarm Topology (0-7)");
opt.addOption("phi1", true, "Phi 1");
opt.addOption("phi2", true, "Phi 2");
/**
* Parse default options.
*/
try {
commandLine = cliParser.parse(opt, optimizerParams);
} catch (ParseException e) {
showHelp(opt);
System.exit(-1);
}
if (commandLine.hasOption("phi1")) {
phi1 = Double.parseDouble(commandLine.getOptionValue("phi1"));
}
if (commandLine.hasOption("phi2")) {
phi2 = Double.parseDouble(commandLine.getOptionValue("phi2"));
}
if (commandLine.hasOption("topology")) {
selectedTopology = PSOTopology.getFromId(Integer.parseInt(commandLine.getOptionValue("topology")));
}
if (commandLine.hasOption("speedLimit")) {
speedLimit = Double.parseDouble(commandLine.getOptionValue("speedLimit"));
}
this.optimizer = OptimizerFactory.createParticleSwarmOptimization(problem, this.populationSize, phi1, phi2, speedLimit, selectedTopology, topoRange, this);
break;
}
case "EvolutionStrategies": {
double pm, pc;
int mu = 5, lambda = 20;
boolean plusStrategy = false;
opt.addOption("pm", true, "Mutation Probability");
opt.addOption("pc", true, "Crossover Probability");
opt.addOption("mu", true, "Mu");
opt.addOption("lambda", true, "Lambda");
opt.addOption("plusStrategy", true, "Whether to use the plus or comma strategy.");
/**
* Parse default options.
*/
try {
commandLine = cliParser.parse(opt, optimizerParams);
} catch (ParseException e) {
showHelp(opt);
System.exit(-1);
}
if (commandLine.hasOption("pm")) {
pm = Double.parseDouble(commandLine.getOptionValue("pm"));
} else {
pm = 0.01;
}
if (commandLine.hasOption("pc")) {
pc = Double.parseDouble(commandLine.getOptionValue("pc"));
} else {
pc = 0.9;
}
if (commandLine.hasOption("mu")) {
mu = Integer.parseInt(commandLine.getOptionValue("mu"));
}
if (commandLine.hasOption("lambda")) {
lambda = Integer.parseInt(commandLine.getOptionValue("lambda"));
}
if (commandLine.hasOption("plusStrategy")) {
plusStrategy = Boolean.parseBoolean(commandLine.getOptionValue("plusStrategy"));
}
this.optimizer = OptimizerFactory.createEvolutionStrategy(mu, lambda, plusStrategy, this.mutator, pm, this.crossover, pc, this.selection, problem, this);
break;
}
default:
throw new Exception("Unsupported Optimizer");
}
}
private void setProblemFromName(String problemName) {
Map<String, Class<? extends AbstractProblemDoubleOffset>> problemList = createProblemList();
Class<? extends AbstractProblemDoubleOffset> problem = problemList.get(problemName);
try {
this.problem = problem.newInstance();
} catch (InstantiationException e) {
e.printStackTrace();
} catch (IllegalAccessException e) {
e.printStackTrace();
}
//setCECDefaults(this.problem);
}
private void setCECDefaults(AbstractProblemDouble problem) {
switch(problem.getName()) {
case "F1-Problem": // F1: Shifted Sphere
this.problem.setDefaultRange(100);
this.problem.setYOffset(fBias[0]);
break;
case "F2-Problem": // F6: Shifted Rosenbrock's Function
this.problem.setDefaultRange(100);
this.problem.setYOffset(fBias[5]);
break;
case "F5-Problem": // F2: Schwefel's 1.2
this.problem.setDefaultRange(100);
this.problem.setYOffset(fBias[1]);
break;
case "F6-Problem": // F9: Shifted Rastrigin's Function
this.problem.setDefaultRange(5);
this.problem.setYOffset(fBias[8]);
break;
default:
LOGGER.info("No CEC'05 default parameters for this problem found.");
break;
}
}
/**
* Executes the optimization and outputs a JSON document to the command line
* with the statistics of the optimization run(s).
*/
private void runOptimization() {
for(int i = 0; i < this.numberOfRuns; i++) {
// Terminate after 10000 function evaluations OR after reaching a fitness < 0.1
OptimizerFactory.setEvaluationTerminator(500000);
//OptimizerFactory.setTerminator(new FitnessValueTerminator(new double[]{0.0001}));
OptimizerFactory.addTerminator(new FitnessValueTerminator(new double[]{0.0001}), CombinedTerminator.OR);
LOGGER.log(Level.INFO, "Running {0}", optimizer.getName());
OptimizationParameters params = OptimizerFactory.makeParams(optimizer, this.populationSize, this.problem, this.seed, OptimizerFactory.getTerminator());
double[] result = OptimizerFactory.optimizeToDouble(params);
JsonObject optimizationDetails = new JsonObject();
optimizationDetails.addProperty("total_time", 1.0);
optimizationDetails.addProperty("total_function_calls", optimizer.getPopulation().getFunctionCalls());
optimizationDetails.addProperty("termination_criteria", OptimizerFactory.terminatedBecause());
optimizationDetails.add("generations", this.generationsArray);
JsonArray solutionArray = new JsonArray();
for(double val : result) {
solutionArray.add(new JsonPrimitive(val));
}
optimizationDetails.add("solution", solutionArray);
this.optimizationRuns.add(optimizationDetails);
// Needs to be re-created here.
this.generationsArray = new JsonArray();
}
this.jsonObject.add("runs", this.optimizationRuns);
Gson gson = new GsonBuilder().setPrettyPrinting().create();
System.out.println(gson.toJson(this.jsonObject));
}
private static void showOptimizerHelp() {
Map<String, Class<? extends InterfaceOptimizer>> optimizerList = createOptimizerList();
System.out.println("Available Optimizers:");
for (String name : optimizerList.keySet()) {
System.out.printf("%s\n", name);
}
}
private static void listProblems() {
Map<String, Class<? extends AbstractProblemDoubleOffset>> problemList = createProblemList();
System.out.println("Available Problems:");
for (String name : problemList.keySet()) {
System.out.printf("%s\n", name);
}
}
@Override
public void registerPopulationStateChanged(Object source, String name) {
if (name.equals("NextGenerationPerformed")) {
InterfaceOptimizer optimizer = (InterfaceOptimizer)source;
Population population = optimizer.getPopulation();
JsonObject newGeneration = new JsonObject();
newGeneration.addProperty("generation", population.getGeneration());
newGeneration.addProperty("function_calls", population.getFunctionCalls());
JsonArray bestFitness = new JsonArray();
for(double val : population.getBestFitness()) {
bestFitness.add(new JsonPrimitive(val));
}
newGeneration.add("best_fitness", bestFitness);
JsonArray meanFitness = new JsonArray();
for(double val : population.getMeanFitness()) {
meanFitness.add(new JsonPrimitive(val));
}
newGeneration.add("mean_fitness", meanFitness);
//System.out.println(newGeneration.toString());
this.generationsArray.add(newGeneration);
}
} }
} }

208
src/eva2/cli/OptimizationBuilder.java Normal file
View File

@@ -0,0 +1,208 @@
package eva2.cli;
import eva2.gui.BeanInspector;
import eva2.optimization.go.InterfaceOptimizationParameters;
import eva2.optimization.modules.OptimizationParameters;
import eva2.tools.ReflectPackage;
import eva2.util.annotation.Hidden;
import eva2.util.annotation.Parameter;
import java.beans.BeanInfo;
import java.beans.IntrospectionException;
import java.beans.Introspector;
import java.beans.PropertyDescriptor;
import java.lang.reflect.Constructor;
import java.lang.reflect.InvocationTargetException;
import java.lang.reflect.Method;
import java.lang.reflect.Modifier;
import java.util.HashMap;
import java.util.LinkedHashMap;
class ArgumentTree extends LinkedHashMap<String, Object> {
private Object value;
public void setValue(Object value) {
this.value = value;
}
public Object getValue() {
return this.value;
}
@Override
public String toString() {
return ((value != null) ? value.toString() + ", " : "") + super.toString();
}
/**
* If there are no key, value pairs present and the value is unset,
* this tree belongs to a flag.
*
* @return
*/
public boolean isFlag() {
return this.size() == 0 && this.value == null;
}
}
/**
*
*/
public final class OptimizationBuilder {
private OptimizationBuilder() {}
public static InterfaceOptimizationParameters parseArguments(String[] args) {
HashMap<String, String> argumentMap = new HashMap<>(args.length/2);
int i = 0;
while (i < args.length) {
// Is it a parameter?
if (args[i].startsWith("--")) {
String key = args[i].substring(2);
String value = null;
// Is the next a value?
if (i < args.length - 1 && !args[i+1].startsWith("--")) {
value = args[i + 1];
argumentMap.put(key, value);
i = i + 2;
} else {
argumentMap.put(key, null);
i++;
}
}
}
System.out.println(argumentMap.toString());
ArgumentTree argumentTree = new ArgumentTree();
for (String key : argumentMap.keySet()) {
insertIntoArgumentTree(argumentTree, key, argumentMap.get(key));
}
System.out.println(argumentTree.toString());
return constructFromArgumentTree(OptimizationParameters.class, argumentTree);
}
private static void insertIntoArgumentTree(ArgumentTree tree, String key, String value) {
// Basic type?
if (!key.contains("-")) {
if (!tree.containsKey(key)) {
tree.put(key, new ArgumentTree());
}
((ArgumentTree)tree.get(key)).setValue(value);
} else {
String baseKey = key.substring(0, key.indexOf('-'));
String restKey = key.substring(key.indexOf('-') + 1);
if (!tree.containsKey(baseKey)) {
tree.put(baseKey, new ArgumentTree());
}
insertIntoArgumentTree((ArgumentTree)tree.get(baseKey), restKey, value);
}
}
/**
*
* @param clazz
* @param tree Tree containing key, value pairs
*/
private static <T> T constructFromArgumentTree(Class<T> clazz, ArgumentTree tree) {
T instance = null;
// Create new instance
if (clazz.isInterface() || Modifier.isAbstract(clazz.getModifiers())) {
// Find subclasses of clazz that match tree.getValue()
} else {
Class<?>[] params = new Class[0];
try {
Constructor constructor = clazz.getConstructor(params);
instance = (T)constructor.newInstance(new Object[]{});
} catch (NoSuchMethodException | InvocationTargetException | InstantiationException | IllegalAccessException ex) {
ex.printStackTrace();
}
}
/* No need to continue if there are no parameters to set */
if (tree.isEmpty()) {
return instance;
}
BeanInfo info;
try {
if (clazz.isInterface()) {
info = Introspector.getBeanInfo(clazz);
} else {
info = Introspector.getBeanInfo(clazz, Object.class);
}
PropertyDescriptor[] properties = info.getPropertyDescriptors();
int foundParameters = 0;
for (PropertyDescriptor pd : properties) {
String name = pd.getName();
Method getter = pd.getReadMethod();
Method setter = pd.getWriteMethod();
Class<?> type = pd.getPropertyType();
// We skip non-existing setters or setters that are hidden by annotation
if (setter == null || setter.isAnnotationPresent(Hidden.class)) {
continue;
}
System.out.println(name + " = " + " type = " + type);
// We use the name of the descriptor or if possible
// one that is given by the @Parameter annotation.
if (setter.isAnnotationPresent(Parameter.class)) {
Parameter param = setter.getAnnotation(Parameter.class);
if (!param.name().isEmpty()) {
name = param.name();
}
}
/**
* If the tree contains this property we try to set it on the object.
*/
if (tree.containsKey(name)) {
foundParameters++;
Object obj;
if (type.isPrimitive() && ((ArgumentTree)tree.get(name)).getValue() != null) {
obj = BeanInspector.stringToPrimitive((String)((ArgumentTree) tree.get(name)).getValue(), type);
} else {
// The subtree has the name of the class
String className = (String)((ArgumentTree)tree.get(name)).getValue();
Class subType;
if (className != null) {
// Try to get the actual class from its name
subType = getClassFromName(className, type);
} else {
subType = type;
}
// Here the recursion starts
obj = constructFromArgumentTree(subType, (ArgumentTree) tree.get(name));
}
// We preserve the default if obj is null
if (obj != null) {
BeanInspector.callIfAvailable(instance, setter.getName(), new Object[]{obj});
}
}
// If we configured all parameters in the tree we can break the loop
if (tree.size() == foundParameters) {
break;
}
}
} catch (IntrospectionException ex) {
ex.printStackTrace();
}
return instance;
}
private static Class<?> getClassFromName(String name, Class type) {
Class<?>[] classes = ReflectPackage.getAssignableClassesInPackage("eva2", type, true, true);
for (Class clazz : classes) {
// We allow both the fully qualified name (eva2.optimization.strategies.GeneticAlgorithm
// and the simple name (GeneticAlgorithm)
if (clazz.getName().equals(name) || clazz.getSimpleName().equals(name)) {
return clazz;
}
}
return null;
}
}

9
src/eva2/optimization/individuals/ESIndividualDoubleData.java
View File

@@ -105,9 +105,6 @@ public class ESIndividualDoubleData extends AbstractEAIndividual implements Inte
} }
} }
/************************************************************************************
* InterfaceDataTypeDouble methods
*/
/** /**
* This method allows you to request a certain amount of double data * This method allows you to request a certain amount of double data
* *
@@ -235,9 +232,6 @@ public class ESIndividualDoubleData extends AbstractEAIndividual implements Inte
System.arraycopy(doubleData, 0, this.genotype, 0, doubleData.length); System.arraycopy(doubleData, 0, this.genotype, 0, doubleData.length);
} }
/************************************************************************************
* AbstractEAIndividual methods
*/
/** /**
* This method will allow a default initialisation of the individual * This method will allow a default initialisation of the individual
* *
@@ -304,9 +298,6 @@ public class ESIndividualDoubleData extends AbstractEAIndividual implements Inte
return strB.toString(); return strB.toString();
} }
/************************************************************************************
* InterfaceESIndividual methods
*/
/** /**
* This method will allow the user to read the ES 'genotype' * This method will allow the user to read the ES 'genotype'
* *

15
src/eva2/optimization/individuals/GAIndividualDoubleData.java
View File

@@ -10,6 +10,7 @@ import eva2.optimization.operator.mutation.MutateGAUniform;
import eva2.problems.InterfaceOptimizationProblem; import eva2.problems.InterfaceOptimizationProblem;
import eva2.tools.math.RNG; import eva2.tools.math.RNG;
import eva2.util.annotation.Description; import eva2.util.annotation.Description;
import eva2.util.annotation.Parameter;
import java.util.BitSet; import java.util.BitSet;
@@ -229,10 +230,6 @@ public class GAIndividualDoubleData extends AbstractEAIndividual implements Inte
} }
} }
/************************************************************************************
* AbstractEAIndividual methods
*/
/** /**
* This method will initialize the individual with a given value for the * This method will initialize the individual with a given value for the
* phenotype. * phenotype.
@@ -373,6 +370,7 @@ public class GAIndividualDoubleData extends AbstractEAIndividual implements Inte
* *
* @param coding The used genotype coding method * @param coding The used genotype coding method
*/ */
@Parameter(name = "coding", description = "Choose the coding to use.")
public void setGACoding(InterfaceGADoubleCoding coding) { public void setGACoding(InterfaceGADoubleCoding coding) {
this.doubleCoding = coding; this.doubleCoding = coding;
} }
@@ -381,16 +379,13 @@ public class GAIndividualDoubleData extends AbstractEAIndividual implements Inte
return this.doubleCoding; return this.doubleCoding;
} }
public String gADoubleCodingTipText() {
return "Choose the coding to use.";
}
/** /**
* This method allows you to set the number of mulitruns that are to be performed, * This method allows you to set the number of mulitruns that are to be performed,
* necessary for stochastic optimizers to ensure reliable results. * necessary for stochastic optimizers to ensure reliable results.
* *
* @param precision The number of multiruns that are to be performed * @param precision The number of multiruns that are to be performed
*/ */
@Parameter(description = "Gives the number of bits to be used to code a double.")
public void setPrecision(int precision) { public void setPrecision(int precision) {
this.precision = precision; this.precision = precision;
} }
@@ -398,8 +393,4 @@ public class GAIndividualDoubleData extends AbstractEAIndividual implements Inte
public int getPrecision() { public int getPrecision() {
return this.precision; return this.precision;
} }
public String precisionTipText() {
return "Gives the number of bits to be used to code a double.";
}
} }

9
src/eva2/optimization/population/Population.java
View File

@@ -15,6 +15,7 @@ import eva2.tools.math.Mathematics;
import eva2.tools.math.RNG; import eva2.tools.math.RNG;
import eva2.tools.math.StatisticUtils; import eva2.tools.math.StatisticUtils;
import eva2.util.annotation.Description; import eva2.util.annotation.Description;
import eva2.util.annotation.Hidden;
import eva2.util.annotation.Parameter; import eva2.util.annotation.Parameter;
import java.util.*; import java.util.*;
@@ -602,6 +603,7 @@ public class Population extends ArrayList implements PopulationInterface, Clonea
return historyList; return historyList;
} }
@Hidden
public void setHistory(LinkedList<AbstractEAIndividual> theHist) { public void setHistory(LinkedList<AbstractEAIndividual> theHist) {
historyList = theHist; historyList = theHist;
} }
@@ -684,6 +686,7 @@ public class Population extends ArrayList implements PopulationInterface, Clonea
* *
* @param d The new number of functioncalls. * @param d The new number of functioncalls.
*/ */
@Hidden
public void setFunctionCalls(int d) { public void setFunctionCalls(int d) {
this.functionCallCount = d; this.functionCallCount = d;
} }
@@ -738,6 +741,7 @@ public class Population extends ArrayList implements PopulationInterface, Clonea
* *
* @param gen the value to set as new generation index * @param gen the value to set as new generation index
*/ */
@Hidden
public void setGeneration(int gen) { public void setGeneration(int gen) {
this.generationCount = gen; this.generationCount = gen;
} }
@@ -2519,11 +2523,8 @@ public class Population extends ArrayList implements PopulationInterface, Clonea
return seedCardinality; return seedCardinality;
} }
@Parameter(description = "The initial cardinality for binary genotype individuals, given as pair of mean and std.dev.")
public void setSeedCardinality(Pair<Integer, Integer> seedCardinality) { public void setSeedCardinality(Pair<Integer, Integer> seedCardinality) {
this.seedCardinality = seedCardinality; this.seedCardinality = seedCardinality;
} }
public String seedCardinalityTipText() {
return "The initial cardinality for binary genotype individuals, given as pair of mean and std.dev.";
}
} }

13
src/eva2/optimization/statistics/EvAStatisticalEvaluation.java
View File

@@ -6,11 +6,12 @@ import eva2.optimization.enums.StatisticsOnTwoSampledData;
import eva2.tools.ReflectPackage; import eva2.tools.ReflectPackage;
import eva2.tools.StringTools; import eva2.tools.StringTools;
import eva2.tools.math.Mathematics; import eva2.tools.math.Mathematics;
import org.apache.commons.math3.stat.ranking.NaNStrategy;
import org.apache.commons.math3.stat.ranking.TiesStrategy;
import java.text.DecimalFormat; import java.text.DecimalFormat;
import java.util.*; import java.util.ArrayList;
import java.util.Collections;
import java.util.LinkedList;
import java.util.List;
import java.util.logging.Logger; import java.util.logging.Logger;
/** /**
@@ -257,19 +258,15 @@ public class EvAStatisticalEvaluation {
private static String calculateMannWhitney(String field, OptimizationJob job1, OptimizationJob job2) { private static String calculateMannWhitney(String field, OptimizationJob job1, OptimizationJob job2) {
double[] dat1 = job1.getDoubleDataColumn(field); double[] dat1 = job1.getDoubleDataColumn(field);
double[] dat2 = job2.getDoubleDataColumn(field); double[] dat2 = job2.getDoubleDataColumn(field);
org.apache.commons.math3.stat.inference.MannWhitneyUTest mU = new org.apache.commons.math3.stat.inference.MannWhitneyUTest(NaNStrategy.FAILED, TiesStrategy.AVERAGE);
double t = Double.NaN; double t = Double.NaN;
if (dat1 != null && dat2 != null) { if (dat1 != null && dat2 != null) {
return "" + mU.mannWhitneyUTest(dat1, dat2);
/*
Object obj = ReflectPackage.instantiateWithParams("jsc.independentsamples.MannWhitneyTest", new Object[]{dat1, dat2}, null); Object obj = ReflectPackage.instantiateWithParams("jsc.independentsamples.MannWhitneyTest", new Object[]{dat1, dat2}, null);
if (obj != null) { if (obj != null) {
Object sp = BeanInspector.callIfAvailable(obj, "getSP", new Object[]{}); Object sp = BeanInspector.callIfAvailable(obj, "getSP", new Object[]{});
t = (Double) sp; t = (Double) sp;
} else { } else {
LOGGER.warning("For the MannWhitney test, the JSC package is required on the class path!"); LOGGER.warning("For the MannWhitney test, the JSC package is required on the class path!");
}*/ }
} }
return "" + t; return "" + t;
} }

2
src/eva2/optimization/strategies/MultiObjectiveCMAES.java
View File

@@ -100,7 +100,7 @@ public class MultiObjectiveCMAES extends AbstractOptimizer implements Serializab
/* /*
* (non-Javadoc) * (non-Javadoc)
* *
* @see eva2.optimization.strategies.InterfaceOptimizer#init() * @see eva2.optimization.strategies.InterfaceOptimizer#initialize()
*/ */
@Override @Override
public void initialize() { public void initialize() {

8
src/eva2/problems/AbstractOptimizationProblem.java
View File

@@ -87,13 +87,11 @@ public abstract class AbstractOptimizationProblem implements InterfaceOptimizati
return parallelThreads; return parallelThreads;
} }
@Parameter(name = "parallel", description = "Set the number of threaded parallel function evaluations - interesting for slow functions and generational optimizers.")
public void setParallelThreads(int parallelThreads) { public void setParallelThreads(int parallelThreads) {
this.parallelThreads = parallelThreads; this.parallelThreads = parallelThreads;
} }
public String parallelThreadsTipText() {
return "Set the number of threaded parallel function evaluations - interesting for slow functions and generational optimizers.";
}
/** /**
* This method initializes the problem instance. * This method initializes the problem instance.
@@ -406,10 +404,6 @@ public abstract class AbstractOptimizationProblem implements InterfaceOptimizati
return template; return template;
} }
public String individualTemplateTipText() {
return "Choose the individual representation to use.";
}
/** /**
* This method extracts the individuals from a given population that are assumed to correspond to local or global optima. * This method extracts the individuals from a given population that are assumed to correspond to local or global optima.
* Similar individuals are clustered together with a density based clustering method * Similar individuals are clustered together with a density based clustering method

16
src/eva2/problems/AbstractProblemDouble.java
View File

@@ -22,6 +22,7 @@ import eva2.tools.diagram.ColorBarCalculator;
import eva2.tools.math.Jama.Matrix; import eva2.tools.math.Jama.Matrix;
import eva2.tools.math.Mathematics; import eva2.tools.math.Mathematics;
import eva2.tools.math.RNG; import eva2.tools.math.RNG;
import eva2.util.annotation.Parameter;
/** /**
* For a double valued problem, there are two main methods to implement: * For a double valued problem, there are two main methods to implement:
@@ -285,6 +286,7 @@ public abstract class AbstractProblemDouble extends AbstractOptimizationProblem
* *
* @param noise The sigma for a gaussian random number. * @param noise The sigma for a gaussian random number.
*/ */
@Parameter(description = "Gaussian noise level on the fitness value.")
public void setNoise(double noise) { public void setNoise(double noise) {
if (noise < 0) { if (noise < 0) {
noise = 0; noise = 0;
@@ -301,15 +303,12 @@ public abstract class AbstractProblemDouble extends AbstractOptimizationProblem
return this.noise; return this.noise;
} }
public String noiseTipText() {
return "Gaussian noise level on the fitness value.";
}
/** /**
* This method allows you to choose the EA individual used by the problem. * This method allows you to choose the EA individual used by the problem.
* *
* @param indy The EAIndividual type * @param indy The EAIndividual type
*/ */
@Parameter(name = "individual", description = "Base individual type defining the data representation and mutation/crossover operators")
public void setEAIndividual(InterfaceDataTypeDouble indy) { public void setEAIndividual(InterfaceDataTypeDouble indy) {
this.template = (AbstractEAIndividual) indy; this.template = (AbstractEAIndividual) indy;
} }
@@ -324,10 +323,6 @@ public abstract class AbstractProblemDouble extends AbstractOptimizationProblem
return (InterfaceDataTypeDouble) this.template; return (InterfaceDataTypeDouble) this.template;
} }
public String EAIndividualTipText() {
return "Set the base individual type defining the data representation and mutation/crossover operators";
}
/** /**
* A (symmetric) absolute range limit. * A (symmetric) absolute range limit.
* *
@@ -351,6 +346,7 @@ public abstract class AbstractProblemDouble extends AbstractOptimizationProblem
return "Absolute limit for the symmetric range in any dimension"; return "Absolute limit for the symmetric range in any dimension";
} }
@Parameter(name = "rotate", description = "If marked, the function is rotated by 22.5 degrees along every axis.")
public void setDoRotation(boolean doRotation) { public void setDoRotation(boolean doRotation) {
this.doRotation = doRotation; this.doRotation = doRotation;
if (!doRotation) { if (!doRotation) {
@@ -362,10 +358,6 @@ public abstract class AbstractProblemDouble extends AbstractOptimizationProblem
return doRotation; return doRotation;
} }
public String doRotationTipText() {
return "If marked, the function is rotated by 22.5 degrees along every axis.";
}
/** /**
* ******************************************************************************************************************* * *******************************************************************************************************************
* These are for InterfaceParamControllable * These are for InterfaceParamControllable

26
src/eva2/problems/SimpleProblemWrapper.java
View File

@@ -12,6 +12,7 @@ import eva2.problems.simple.InterfaceSimpleProblem;
import eva2.problems.simple.SimpleF1; import eva2.problems.simple.SimpleF1;
import eva2.problems.simple.SimpleProblemBinary; import eva2.problems.simple.SimpleProblemBinary;
import eva2.problems.simple.SimpleProblemDouble; import eva2.problems.simple.SimpleProblemDouble;
import eva2.util.annotation.Parameter;
import java.util.BitSet; import java.util.BitSet;
@@ -164,6 +165,7 @@ public class SimpleProblemWrapper extends AbstractOptimizationProblem {
/** /**
* @param simProb the simProb to set * @param simProb the simProb to set
*/ */
@Parameter(description = "Set the simple problem class which is to be optimized")
public void setSimpleProblem(InterfaceSimpleProblem<?> simProb) { public void setSimpleProblem(InterfaceSimpleProblem<?> simProb) {
this.simProb = simProb; this.simProb = simProb;
initTemplate(); initTemplate();
@@ -186,19 +188,13 @@ public class SimpleProblemWrapper extends AbstractOptimizationProblem {
} }
} }
/**
*
*/
public String simpleProblemTipText() {
return "Set the simple problem class which is to be optimized";
}
/** /**
* This method allows you to choose how much noise is to be added to the * This method allows you to choose how much noise is to be added to the
* fitness. This can be used to make the optimization problem more difficult. * fitness. This can be used to make the optimization problem more difficult.
* *
* @param noise The sigma for a gaussian random number. * @param noise The sigma for a gaussian random number.
*/ */
@Parameter(description = "Gaussian noise level on the fitness value.")
public void setNoise(double noise) { public void setNoise(double noise) {
if (noise < 0) { if (noise < 0) {
noise = 0; noise = 0;
@@ -210,11 +206,6 @@ public class SimpleProblemWrapper extends AbstractOptimizationProblem {
return this.noise; return this.noise;
} }
public String noiseTipText() {
return "Gaussian noise level on the fitness value.";
}
/** /**
* A (symmetric) absolute range limit. * A (symmetric) absolute range limit.
* *
@@ -229,15 +220,12 @@ public class SimpleProblemWrapper extends AbstractOptimizationProblem {
* *
* @param defaultRange * @param defaultRange
*/ */
@Parameter(name = "range", description = "Absolute limit for the symmetric range in any dimension")
public void setDefaultRange(double defaultRange) { public void setDefaultRange(double defaultRange) {
this.defaultRange = defaultRange; this.defaultRange = defaultRange;
initTemplate(); initTemplate();
} }
public String defaultRangeTipText() {
return "Absolute limit for the symmetric range in any dimension";
}
/** /**
* Take care that all properties which may be hidden (and currently are) send a "hide" message to the Java Bean properties. * Take care that all properties which may be hidden (and currently are) send a "hide" message to the Java Bean properties.
* This is called by PropertySheetPanel in use with the GenericObjectEditor. * This is called by PropertySheetPanel in use with the GenericObjectEditor.
@@ -246,16 +234,12 @@ public class SimpleProblemWrapper extends AbstractOptimizationProblem {
setSimpleProblem(getSimpleProblem()); setSimpleProblem(getSimpleProblem());
} }
@Parameter(name = "individual", description = "Set the individual properties for the optimization")
public void setIndividualTemplate(AbstractEAIndividual indy) { public void setIndividualTemplate(AbstractEAIndividual indy) {
resetTemplate = false; resetTemplate = false;
template = indy; template = indy;
} }
@Override
public String individualTemplateTipText() {
return "Set the individual properties for the optimization";
}
/** /**
* This method returns a string describing the optimization problem. * This method returns a string describing the optimization problem.
* *

2
src/eva2/tools/Serializer.java
View File

@@ -1,7 +1,5 @@
package eva2.tools; package eva2.tools;
import com.google.gson.Gson;
import java.io.*; import java.io.*;
import java.util.logging.Level; import java.util.logging.Level;
import java.util.logging.Logger; import java.util.logging.Logger;