| Sumario: | The article presents information on John Holland's ideas about robust adaptive systems that led to the design and implementation of genetic algorithms (GAs). Since then, GAs have been demonstrated to be an effective problem-solving tool for tackling complex optimization and machine learning problems. The GA exhibits global search capabilities by simultaneously evaluating performances at multiple points in the solution space. The fundamental approach of reducing computational workload is to develop more efficient algorithms based on a sound theoretical understanding of GAs. The P2P GA environment provides a friendly user interface and a uniform GA executor capable of utilizing the resources of all participating computers on the Internet, while hiding the complexity of network programming from the end users. The P2P GA infrastructure is built on top of a class library that makes it easy to use or to expand with advanced features for solving sophisticated problems posted by the users. The P2P GA framework covers issues such as the choice of migration strategies, the use of communication mechanisms, and the installation of user-defined evaluators.
|
|---|