ECJ A Java-based evolutionary computation research system
Download

ECJ Ranking & Summary

ECJ Tags

evolution compute research evolution computation evolution research research evolution

ECJ Description

A Java-based evolutionary computation research system ECJ is a free research EC system written in Java that was designed to be highly flexible, with nearly all classes (and all of their settings) dynamically determined at runtime by a user-provided parameter file. All structures in the system are arranged to be easily modifiable. Even so, the system was designed with an eye toward efficiency.NOTE: ECJ is provided and licensed under the terms of the Academic Free License version 3.0. Here are some key features of "ECJ": General Features: · GUI with charting · Platform-independent checkpointing and logging · Hierarchical parameter files · Multithreading · Mersenne Twister Random Number Generators · Abstractions for implementing a variety of EC forms. EC Features: · Asynchronous island models over TCP/IP · Master/Slave evaluation over multiple processors, with support for generational, asynchronous steady-state, and coevolutionary distribution · Genetic Algorithms/Programming style Steady State and Generational evolution, with or without Elitism · Evolutionary-Strategies style (mu,lambda) and (mu+lambda) evolution · Very flexible breeding architecture · Many selection operators · Multiple subpopulations and species · Inter-subpopulation exchanges · Reading populations from files · Single- and Multi-population coevolution · SPEA2 multiobjective optimization · Particle Swarm Optimization · Differential Evolution · Spatially embedded evolutionary algorithms · Hooks for other multiobjective optimization methods · Packages for parsimony pressure GP Tree Representations: · Set-based Strongly-Typed Genetic Programming · Ephemeral Random Constants · Automatically-Defined Functions and Automatically Defined Macros · Multiple tree forests · Six tree-creation algorithms · Extensive set of GP breeding operators · Eight pre-done GP application problem domains (ant, regression, multiplexer, lawnmower, parity, two-box, edge, serengeti) Vector (GA/ES) Representations: · Fixed-Length and Variable-Length Genomes · Arbitrary representations · Ten pre-done vector application problem domains (rastrigin, sum, rosenbrock, sphere, step, noisy-quartic, booth, griewangk, nk, hiff) Other Representations: · Multiset-based genomes in the rule package, for evolving Pitt-approach rulesets or other set-based representations. Requirements: · Java

ECJ Related Software