org.evosuite.ga.metaheuristics.MonotonicGA Maven / Gradle / Ivy
/**
* Copyright (C) 2010-2018 Gordon Fraser, Andrea Arcuri and EvoSuite
* contributors
*
* This file is part of EvoSuite.
*
* EvoSuite is free software: you can redistribute it and/or modify it
* under the terms of the GNU Lesser General Public License as published
* by the Free Software Foundation, either version 3.0 of the License, or
* (at your option) any later version.
*
* EvoSuite is distributed in the hope that it will be useful, but
* WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with EvoSuite. If not, see extends GeneticAlgorithm {
private static final long serialVersionUID = 7846967347821123201L;
protected ReplacementFunction replacementFunction;
private final Logger logger = LoggerFactory.getLogger(MonotonicGA.class);
/**
* Constructor
*
* @param factory
* a {@link org.evosuite.ga.ChromosomeFactory} object.
*/
public MonotonicGA(ChromosomeFactory factory) {
super(factory);
setReplacementFunction(new FitnessReplacementFunction());
}
/**
*
* keepOffspring
*
*
* @param parent1
* a {@link org.evosuite.ga.Chromosome} object.
* @param parent2
* a {@link org.evosuite.ga.Chromosome} object.
* @param offspring1
* a {@link org.evosuite.ga.Chromosome} object.
* @param offspring2
* a {@link org.evosuite.ga.Chromosome} object.
* @return a boolean.
*/
protected boolean keepOffspring(Chromosome parent1, Chromosome parent2, Chromosome offspring1,
Chromosome offspring2) {
return replacementFunction.keepOffspring(parent1, parent2, offspring1, offspring2);
}
/** {@inheritDoc} */
@SuppressWarnings("unchecked")
@Override
protected void evolve() {
List newGeneration = new ArrayList();
// Elitism
logger.debug("Elitism");
newGeneration.addAll(elitism());
// Add random elements
// new_generation.addAll(randomism());
while (!isNextPopulationFull(newGeneration) && !isFinished()) {
logger.debug("Generating offspring");
T parent1 = selectionFunction.select(population);
T parent2;
if (Properties.HEADLESS_CHICKEN_TEST)
parent2 = newRandomIndividual(); // crossover with new random
// individual
else
parent2 = selectionFunction.select(population); // crossover
// with existing
// individual
T offspring1 = (T) parent1.clone();
T offspring2 = (T) parent2.clone();
try {
// Crossover
if (Randomness.nextDouble() <= Properties.CROSSOVER_RATE) {
crossoverFunction.crossOver(offspring1, offspring2);
}
} catch (ConstructionFailedException e) {
logger.info("CrossOver failed");
continue;
}
// Mutation
notifyMutation(offspring1);
offspring1.mutate();
notifyMutation(offspring2);
offspring2.mutate();
if (offspring1.isChanged()) {
offspring1.updateAge(currentIteration);
}
if (offspring2.isChanged()) {
offspring2.updateAge(currentIteration);
}
// The two offspring replace the parents if and only if one of
// the offspring is not worse than the best parent.
for (FitnessFunction fitnessFunction : fitnessFunctions) {
fitnessFunction.getFitness(offspring1);
notifyEvaluation(offspring1);
fitnessFunction.getFitness(offspring2);
notifyEvaluation(offspring2);
}
if (keepOffspring(parent1, parent2, offspring1, offspring2)) {
logger.debug("Keeping offspring");
// Reject offspring straight away if it's too long
int rejected = 0;
if (isTooLong(offspring1) || offspring1.size() == 0) {
rejected++;
} else {
// if(Properties.ADAPTIVE_LOCAL_SEARCH ==
// AdaptiveLocalSearchTarget.ALL)
// applyAdaptiveLocalSearch(offspring1);
newGeneration.add(offspring1);
}
if (isTooLong(offspring2) || offspring2.size() == 0) {
rejected++;
} else {
// if(Properties.ADAPTIVE_LOCAL_SEARCH ==
// AdaptiveLocalSearchTarget.ALL)
// applyAdaptiveLocalSearch(offspring2);
newGeneration.add(offspring2);
}
if (rejected == 1)
newGeneration.add(Randomness.choice(parent1, parent2));
else if (rejected == 2) {
newGeneration.add(parent1);
newGeneration.add(parent2);
}
} else {
logger.debug("Keeping parents");
newGeneration.add(parent1);
newGeneration.add(parent2);
}
}
population = newGeneration;
// archive
updateFitnessFunctionsAndValues();
currentIteration++;
}
private T newRandomIndividual() {
T randomChromosome = chromosomeFactory.getChromosome();
for (FitnessFunction fitnessFunction : this.fitnessFunctions) {
randomChromosome.addFitness(fitnessFunction);
}
return randomChromosome;
}
/** {@inheritDoc} */
@Override
public void initializePopulation() {
notifySearchStarted();
currentIteration = 0;
// Set up initial population
generateInitialPopulation(Properties.POPULATION);
logger.debug("Calculating fitness of initial population");
calculateFitnessAndSortPopulation();
this.notifyIteration();
}
private static final double DELTA = 0.000000001; // it seems there is some
// rounding error in LS,
// but hard to debug :(
/** {@inheritDoc} */
@Override
public void generateSolution() {
if (Properties.ENABLE_SECONDARY_OBJECTIVE_AFTER > 0 || Properties.ENABLE_SECONDARY_OBJECTIVE_STARVATION) {
disableFirstSecondaryCriterion();
}
if (population.isEmpty()) {
initializePopulation();
assert!population.isEmpty() : "Could not create any test";
}
logger.debug("Starting evolution");
int starvationCounter = 0;
double bestFitness = Double.MAX_VALUE;
double lastBestFitness = Double.MAX_VALUE;
if (getFitnessFunction().isMaximizationFunction()) {
bestFitness = 0.0;
lastBestFitness = 0.0;
}
while (!isFinished()) {
logger.info("Population size before: " + population.size());
// related to Properties.ENABLE_SECONDARY_OBJECTIVE_AFTER;
// check the budget progress and activate a secondary criterion
// according to the property value.
{
double bestFitnessBeforeEvolution = getBestFitness();
evolve();
sortPopulation();
double bestFitnessAfterEvolution = getBestFitness();
if (getFitnessFunction().isMaximizationFunction())
assert(bestFitnessAfterEvolution >= (bestFitnessBeforeEvolution
- DELTA)) : "best fitness before evolve()/sortPopulation() was: " + bestFitnessBeforeEvolution
+ ", now best fitness is " + bestFitnessAfterEvolution;
else
assert(bestFitnessAfterEvolution <= (bestFitnessBeforeEvolution
+ DELTA)) : "best fitness before evolve()/sortPopulation() was: " + bestFitnessBeforeEvolution
+ ", now best fitness is " + bestFitnessAfterEvolution;
}
{
double bestFitnessBeforeLocalSearch = getBestFitness();
applyLocalSearch();
double bestFitnessAfterLocalSearch = getBestFitness();
if (getFitnessFunction().isMaximizationFunction())
assert(bestFitnessAfterLocalSearch >= (bestFitnessBeforeLocalSearch
- DELTA)) : "best fitness before applyLocalSearch() was: " + bestFitnessBeforeLocalSearch
+ ", now best fitness is " + bestFitnessAfterLocalSearch;
else
assert(bestFitnessAfterLocalSearch <= (bestFitnessBeforeLocalSearch
+ DELTA)) : "best fitness before applyLocalSearch() was: " + bestFitnessBeforeLocalSearch
+ ", now best fitness is " + bestFitnessAfterLocalSearch;
}
/*
* TODO: before explanation: due to static state handling, LS can
* worse individuals. so, need to re-sort.
*
* now: the system tests that were failing have no static state...
* so re-sorting does just hide the problem away, and reduce
* performance (likely significantly). it is definitively a bug
* somewhere...
*/
// sortPopulation();
double newFitness = getBestFitness();
if (getFitnessFunction().isMaximizationFunction())
assert(newFitness >= (bestFitness - DELTA)) : "best fitness was: " + bestFitness
+ ", now best fitness is " + newFitness;
else
assert(newFitness <= (bestFitness + DELTA)) : "best fitness was: " + bestFitness
+ ", now best fitness is " + newFitness;
bestFitness = newFitness;
if (Double.compare(bestFitness, lastBestFitness) == 0) {
starvationCounter++;
} else {
logger.info("reset starvationCounter after " + starvationCounter + " iterations");
starvationCounter = 0;
lastBestFitness = bestFitness;
}
updateSecondaryCriterion(starvationCounter);
logger.info("Current iteration: " + currentIteration);
this.notifyIteration();
logger.info("Population size: " + population.size());
logger.info("Best individual has fitness: " + population.get(0).getFitness());
logger.info("Worst individual has fitness: " + population.get(population.size() - 1).getFitness());
}
// archive
TimeController.execute(this::updateBestIndividualFromArchive, "update from archive", 5_000);
notifySearchFinished();
}
private double getBestFitness() {
T bestIndividual = getBestIndividual();
for (FitnessFunction ff : fitnessFunctions) {
ff.getFitness(bestIndividual);
}
return bestIndividual.getFitness();
}
/**
*
* setReplacementFunction
*
*
* @param replacement_function
* a {@link org.evosuite.ga.ReplacementFunction} object.
*/
public void setReplacementFunction(ReplacementFunction replacement_function) {
this.replacementFunction = replacement_function;
}
/**
*
* getReplacementFunction
*
*
* @return a {@link org.evosuite.ga.ReplacementFunction} object.
*/
public ReplacementFunction getReplacementFunction() {
return replacementFunction;
}
}