Special track on Applications of Evolutionary Computation
Bryant A. Julstrom · 2008
Evolutionary algorithms (EAs) are inspired by the mechanisms that underlie biological evolution: reproduction with variation, selection according to fitness, and repetition. An EA maintains a population of data structures that encode candidate solutions to the target problem instance. These data structures are usually strings of symbols and are called chromosomes. Associated with each chromosome is a numerical fitness that indicates the quality of the solution it represents, and chromosomes of better fitness are selected to be parents. Operators abstracted from genetic crossover and mutation, and bearing those names, generate novel chromosomes from the selected parents. As generations of chromosomes follow each other, representations of better solutions evolve. The algorithm halts and returns the best solution represented in its population after a specified number of generations or when the algorithm identifies a solution of adequate fitness.