Universidad Carlos III de Madrid, Spain
Complex problems can be encountered in real-life domains such as finance and economics, telecommunications, and industrial environments. Solving these problems with evolutionary computation (EC) techniques requires deciding on what the best representations, operators, evaluation function, and parameters are. Furthermore, some complex problems need long computation times to evaluate the solutions proposed by the technique applied or even interaction with an expert. In those cases, it is necessary to decrease the number of evaluations made. One approach commonly used with population-based approaches is to reduce the number of individuals in the population. Micropopulations, which means at most 10 individuals, reduce the genetic diversity dramatically, diminishing the overall performance of the technique applied. However, there are some alternatives to addressing this problem, such as a fitness prediction genetic algorithm (FPGA) or the chromosome appearance probability matrix (CAPM) algorithm.
EC encompasses computational models that follow a biological evolution metaphor. The success of these techniques is based on the maintenance of genetic diversity, for which it is necessary to work with large ...