An automatic and interactive evolutionary system with Grid Computing.

摘要

Evolutionary Computation (EC) is a powerful technique that when incorporated into a tool can allow for the exploration of solutions to complex problems. Based upon the Darwinian theory of natural selection, tools that employ EC have demonstrated the ability to propose amazing solutions, even rivaling the best solutions that human beings can come up with. This research investigates and presents a system that is a tool for exploring creative design problems and satisfying parameter constraint problems. Specifically, Evolvere, is an evolutionary design system that has the ability to interactively evolve creative design solutions and evolve parametric problems using a fitness criterion. The purpose is to demonstrate the effectiveness of using a Swarm Intelligence (SI) algorithm called Particle Swarm Optimization (PSO) and Grid Computing in order to evolve quickly and easily solutions to difficult problems. The system architecture of Evolvere will be presented, followed by two example applications. These two examples involve the evolution of fractal structures and a gene regulatory system, which both help to demonstrate Evolvere's flexibility and effectiveness as an exploration tool.