Optimization Algorithm Toolkit

Home | Screenshots | About | Support

Getting Started

  1. If you do not have Java, download and install the Java Runtime Environment
  2. Download the Optimization Algorithm Toolkit
  3. Unzip into a directory, example /oat
    1. Windows: Double click the jar file: optalgtoolkit.jar
    2. Command Line: In the installation directory, type: java -jar optalgtoolkit.jar

In addition to the simplest method for starting the application outlined above, there are many other entry points into the GUI's of the library. For example the experimenter may be launched from com.oat.experimenter.gui.entry.ExperimenterMain. The explorer may be launched from com.oat.explorer.gui.entry.ExplorerMain or as an applet from com.oat.explorer.gui.entry.ExplorerApplet, and domain-specific explorer applications and applets may be launched from com.oat.explorer.domains.*. The launcher may be invoked manually via com.oat.gui.launcher.LauncherMain.

Basic Support

Visit the OAT project home on sourceforge for useful things such as downloads, user forum, feature requests, bug reports, and CVS access.
  • Found a Bug?
    If you encounter something unexpected while using the OAT platform (a possible software bug) you have multiple options. The best option is to post a message on one of the project's message boards. As a last resort, you can contact one of the project's developers and mention 'OAT' in the subject of the email.
  • Desire a Feature?
    If you are an OAT user, there may be something you require or desire from the platform that is currently not offered. Examples may include small requests such as configuration options and interface behaviors, moderate requests such as implementation of an algorithm instance, problem instance or statistic, and finally large requests such as new domains, and new user interfaces for the platform. The best option is to post a message on one of the project's message boards. As a last resort, you can contact one of the project's developers and mention 'OAT' in the subject of the email.
  • Want to Contribute?
    Contributions to OAT are welcome with open arms! You may contribute to OAT in many different ways. If you are an interested user you can try and break OAT (find bugs), or create and document the results of standard experiments. If you are an experienced software developer you can contribute problem and algorithm implementations or join the project as a developer. If you are a research scientists you can contribute your algorithm and problem implementations as well as your experiment files. Other examples of contributions include documentation, examples, demonstrations and publications. If you feel that you have something to contribute to the OAT project the best option is to contact one of the project's developers and mention 'OAT' in the subject of the email.
  • Want to Contact the Lead Developer?
    The project lead is Jason Brownlee.

Developer Support

Software developers are directed primarily to the source code, which is quite self-explanatory. Source code is provided in the software download, and CVS access is provided on sourceforge. Access is provided to the source, although the comments may be less useful than actually eye-balling the code. Ideally we want to provide tutorials on implementing new algorithms, problems, domains, stop conditions, run probes, etc., although at this time this work has not been completed due to time pressures.

Simple examples of using the API algorithms and problem instances are provided in the package: com.oat.domains.examples. Simple test examples of managing and analyzing experiments are provided in the package: com.oat.experimenter.examples. Further examples may be drawn from the domain unit tests provided in the package: com.oat.junit.

Documentation

  • Jason Brownlee. OAT HowTo: High-level Domain, Problem, and Algorithm Implementation [Technical Report]. Victoria, Australia: Complex Intelligent Systems Laboratory (CIS), Centre for Information Technology Research (CITR), Faculty of Information and Communication Technologies (ICT), Swinburne University of Technology; 2007 December; Technical Report: 20071218A.
  • Jason Brownlee. OAT: The Optimization Algorithm Toolkit [Technical Report]. Victoria, Australia: Complex Intelligent Systems Laboratory (CIS), Centre for Information Technology Research (CITR), Faculty of Information and Communication Technologies (ICT), Swinburne University of Technology; 2007 December; Technical Report: 20071220A.
  • Jason Brownlee. OAT: A Quick Start Guide [Technical Report]. Victoria, Australia: Complex Intelligent Systems Laboratory (CIS), Centre for Information Technology Research (CITR), Faculty of Information and Communication Technologies (ICT), Swinburne University of Technology; 2007 December; Guide: 20071220A.

Research Support

Help for those interested in using OAT as a platform for research in any and all of the fields of Computational Intelligence, Machine Learning, Biologically-Inspired Computation, Metaheuristics, and Artificial Intelligence.

The vision for OAT in the context of computational intelligence is to provide standardised algorithm and problem implementations, a so called point-and-click reproduction of experiments from literature. With this as a base OAT seeks to provide a trusted and consistent platform for computational intelligence research including best practice exploration and experimentation methods and configurations. The future for OAT is an online computational intelligence knowledge repository to capture algorithms, problems, standard experimental results, such that algorithm-problem combinations may be simply referred to rather continuously re-implemented and re-experimented. Treating such combinations as a starting-point (a known) for research rather than an end-point.


SourceForge.net Logo
©2006-2008 Jason Brownlee