Evolution has created a wealth of species that are well adapted to their environments. Genetic algorithms (GAs) are artificial procedures seek to achieve similar success in a wide variety of optimization problems--design problems where the best possible solution is sought--by mimicking the principles behind biological evolution.

A GA searches a potentially huge space of solutions by maintaining a population of "individuals," that is, candidate solutions to a particular optimization problem. A crucial ingredient in any GA algorithm is the fitness function it uses to rate individuals' success. The optimization goal is to find the individual with the maximum possible "fitness"; this is accomplished by taking some fraction of the most fit individuals in the current population and discarding the rest. The remainder of relatively fit individuals is then modified to create a new generation by mutation and crossover. Mutation modifies a single individual randomly. Crossover involves...