Opt4J

Version 2.7 changelog: - update to SAT4J 2.3.2 -> fixed Java 7 compile error - improved documentation - minor code fixes (especially, method visibility)

Version 2.7 changelog:- update to SAT4J 2.3.2 -> fixed Java 7 compile error

Version 2.5 changelog: - added Knapsack problem - added SMS selector (based on hypervolume)

Version 2.5 changelog:- added Knapsack problem

Version 2.4 changelog:- Decoupled optimization and GUI

Version 2.4 changelog: - Decoupled optimization and GUI - Revised package structure: - Moved Archive, Population, and IncompatibilityException to core.optimizer - Moved Phenotype and Genotype to core - Using Guice 3.0 - Revised Evaluator interface ... [More] (removed need for #getObjectives()) - GUI improvements (AutoZoom, ConvergencePlot) - Revised SAT package - Revised operator structure - Logger improvements - Documentation/javadoc improvements - Several minor bugfixes - Renamed IndividualCreator to IndivdiualFactory - Changed mutate interface from mutate(Genotype) to mutate(Genotype,double) - Split the Mating into an interface and implementation [Less]

Fixed several minor bugs Several minor improvements Added multi-modules (allows multiple instances of the same module) Added xml viewer in GUI Improved documentation (now based on docbook)

Fixed several minor bugsSeveral minor improvementsAdded multi-modules (allows multiple instances of the same module)

Thumbs up: Opt4J is a very intuitive optimization framework based on Java. Thanks to the detailed Javadocs and a comprehensive tutorial, it is easy to understand the basics of the framework and use it for own optimization problems. The Library ... [More] also offers a GUI to (a) configure and (b) visualize the optimization process. The programmers apply Google Guice to resolve dependencies at run-time. In my opinion, the latter is the only point where understanding Opt4J gets a little more tough. However, after some happy coding based on trial-and-error, one gets comfortable with this quite useful Guice-stuff. The project already contains several state-of-the-art optimizers, such as NSGA-II, SPEA2, Simulated Annealing etc. So, it is easy to run one problem with different Optimizers and look which one performs best. Personally, I am using Opt4J to perform design space exploration for application mapping in NoC-based systems. Though I am not an expert in EA-based optimization, it was quite easy for me to get first results. Consequently, a definitive „thumbs up“! [Less]

Significantly improved SPEA2 performance. Some bugfixes. Improved error messages and error handling.