Title :
Observing the movement of GAs on neutral networks
Author_Institution :
Sch. of Cognitive & Comput. Sci., Sussex Univ., Brighton, UK
Abstract :
This paper looks at the dynamics generated by running evolutionary search algorithms on real world problems (RWPs). First it is argued that to effectively use evolutionary search algorithms to solve engineering problems requires that we manage to create effective search dynamics and hence that the dynamics of search processes needs to be better understood. Next a new method of plotting the movement in the trace of a dynamical system is introduced. The resulting movement graphs for some random walkers and random search are plotted and compared to a movement graph of the Lorenz attractor. Then various movement graphs are plotted from a particular trace of a GA trying to solve a RWP that has been previously investigated in detail (Sharpe, 1999). The resulting movement graphs are used to observe the movement of the GA across large neutral networks in the problem space
Keywords :
genetic algorithms; graph theory; search problems; Lorenz attractor; evolutionary search algorithms; movement graph; neutral networks; random search; random walkers; real world problems; search dynamics; Electrostatic precipitators; Encoding; Heart; Navigation; Nonlinear dynamical systems; Visualization;
Conference_Titel :
Evolutionary Computation, 2000. Proceedings of the 2000 Congress on
Conference_Location :
La Jolla, CA
Print_ISBN :
0-7803-6375-2
DOI :
10.1109/CEC.2000.870288