ABSTRACT
Genetic algorithms (GAs) were devised by John Holland at the University of Michigan in the early seventies and were detailed in his pioneering work Adaptation in Natural and Artificial Systems. An important difference between GAs and the other evolutionary optimisation algorithms is that GAs operate on the genotypic representation while the evolutionary strategies and evolutionary programming (EP) methods operate on a phenotypic representation. The main differences between GAs and Evolutionary Strategies (ES) arise either directly or indirectly from the representations used by the algorithm. In GAs, the primary search operator is the crossover operation and serves to enlarge the search space. The January 1994 IEEE Transactions on Neural Networks was a special issue on evolutionary computation, and included papers detailing both the theoretical analysis and application areas for evolutionary optimisation algorithms. EP and ES are two paradigms of simulated evolution which are very similar in structure and operation.
