This toolkit provides a platform for developing optimization algorithms with tools and evaluations included.
The software offers numerous optimization algorithms, including biologically inspired approaches such as genetic algorithms, swarm algorithms, and immune system algorithms. Similarly, there are more conventional approaches inspired by physics, including simulated annealing and extremal optimization. Problem domains include numerical function optimization, traveling salesman problems, and protein folding problems. The software also includes standard benchmark instances taken from the research literature.
A user-friendly graphical interface is provided to evaluate and compare algorithm and problem configurations, visualize algorithm behavior, and graph algorithm performance over time. The platform is structured with a robust, modular, and extensible framework meant to facilitate easy addition and modification of algorithms and problem domains. This framework makes advanced algorithm experimentation possible, and algorithm implementations are extensible and easily support modification and application to varied problem domains.
The software is open source, released under the GPL, so the source code is available. The software was compiled with Java 1.5 (update 9). This release features a significant restructuring of the API, bug fixes throughout, a new beta experimenter API and graphical user interface, and notably, the new experimenter includes the standard statistical hypothesis tests for normality and comparison of algorithm results.
If you encounter any bugs or wish to request new features, you can access the services on the project's website, and you can even add your algorithms to the software. In conclusion, Optimization Algorithm Toolkit is a useful tool for both research scientists and algorithm practitioners to improve optimization algorithms by using classical and state-of-the-art optimization algorithms.
Version 1.4: N/A