Java Evolutionary Computation Toolkit

ECJ is a freeware <a href="/facts/Evolutionary_computation/qXyxPI8j">evolutionary computation</a> research system written in Java. It is a framework that supports a variety of evolutionary computation techniques, such as <a href="/facts/Genetic_algorithms/WP2AFWuW">genetic algorithms</a>, <a href="/facts/Genetic_programming/kGa2jr0w">genetic programming</a>, <a href="/facts/Evolution_strategies/88T0MMkv">evolution strategies</a>, <a href="/facts/Coevolution/r7w308za">coevolution</a>, particle swarm optimization, and <a href="/facts/Differential_evolution/mcMagaB2">differential evolution</a>. The framework models iterative evolutionary processes using a series of pipelines arranged to connect one or more subpopulations of individuals with <a href="/facts/Selection_(genetic_algorithm)/YrIlmEBy">selection</a>, breeding (such as <a href="/facts/Crossover_(genetic_algorithm)/fRBnRV7r">crossover</a>, and <a href="/facts/Mutation_(genetic_algorithm)/3NprRzPI">mutation</a> operators that produce new individuals. The framework is open source and is distributed under the <a href="/facts/Academic_Free_License/kbPBSo3W">Academic Free License</a>. ECJ was created by <a href="http://www.cs.gmu.edu/~sean/">Sean Luke</a>, a computer science professor at George Mason University, and is maintained by Sean Luke and a variety of contributors.
Features (listed from <a href="http://cs.gmu.edu/~eclab/projects/ecj/">ECJ's project page</a>):
General Features:

<ul><li>GUI with charting</li>
<li>Platform-independent checkpointing and logging</li>
<li>Hierarchical parameter files</li>
<li>Multithreading</li>
<li>Mersenne Twister Random Number Generators</li>
<li>Abstractions for implementing a variety of EC forms.</li></ul>
EC Features:

<ul><li>Asynchronous island models over TCP/IP</li>
<li>Master/Slave evaluation over multiple processors</li>
<li>Genetic Algorithms/Programming style Steady State and Generational evolution, with or without Elitism</li>
<li>Evolutionary-Strategies style (mu, lambda) and (mu+lambda) evolution</li>
<li>Very flexible breeding architecture</li>
<li>Many selection operators</li>
<li>Multiple subpopulations and species</li>
<li>Inter-subpopulation exchanges</li>
<li>Reading populations from files</li>
<li>Single- and Multi-population coevolution</li>
<li>SPEA2 multiobjective optimization</li>
<li>Particle Swarm Optimization</li>
<li>Differential Evolution</li>
<li>Spatially embedded evolutionary algorithms</li>
<li>Hooks for other multiobjective optimization methods</li>
<li>Packages for parsimony pressure</li></ul>
GP Tree Representations:

<ul><li>Set-based Strongly Typed Genetic Programming</li>
<li>Ephemeral Random Constants</li>
<li>Automatically Defined Functions and Automatically Defined Macros</li>
<li>Multiple tree forests</li>
<li>Six tree-creation algorithms</li>
<li>Extensive set of GP breeding operators</li>
<li>Seven pre-done GP application problem domains (ant, regression, multiplexer, lawnmower, parity, two-box, edge)</li></ul>
Vector (GA/ES) Representations:

<ul><li>Fixed-Length and Variable-Length Genomes</li>
<li>Arbitrary representations</li>
<li>Five pre-done vector application problem domains (sum, rosenbrock, sphere, step, noisy-quartic)</li></ul>
Other Representations:

<ul><li>NEAT</li>
<li>Multiset-based genomes in the rule package, for evolving Pitt-approach rulesets or other set-based representations.</li></ul>

Java Evolutionary Computation Toolkit open-in-new

Java Evolutionary Computation Toolkit