Genetic Algorithms and Optimizing Chemical Oxygen-iodine Lasers

This paper presents results from the first known application of the genetic algorithm (GA) technique for optimizing the performance of a laser system (chemical, solid-state, or gaseous). The effects of elitism, single point and uniform crossover, creep mutation, different random number seeds, population size, niching and the number of children per pair of parents on the performance of the GA for this problem were studied. Micro-GAs (μGA) were also tested. The best overall performer was the uniform crossover μGA with a population size of 5. The uniform crossover μGA was also able to find the global maximum of an order-3 deceptive function which the other tested GAs failed to optimize.