A New Self-adaptative Crossover Operator for Real-Coded Evolutionary Algorithms

摘要: In this paper we propose a new self-adaptative crossover operator for real coded evolutionary algorithms. This has the capacity to simulate other real-coded operators dynamically and, therefore, it achieve exploration and exploitation during process according best individuals. words, proposed may handle generational diversity of population in such way that either generate additional from current one, allowing take effect, or use previously generated exploit better solutions. In order test performance crossover, have used set functions made comparative study against classic operators. The analysis results allows us affirm very suitable behavior; although, should be noted offers behavior applied complex search spaces than simple ones.