Evolutionary optimization

Simulated evolution can be used as an effective numerical optimization procedure. The robust nature of stochastic search can be applied to general problem solving. Recent research in simulated evolution has been applied to neural network design and training, automatic control, system identification and other combinatorial problems. A brief review of these methods is offered. Some mathematical properties of specific evolutionary techniques, such as evolutionary programming and genetic algorithms, are detailed. Function optimization experiments are conducted to illustrate the mathematical procedures.<
>