Adaptive genetic algorithm with a cooperative mode

Adaptation of parameters and operators is one of the most important and promising areas of research in evolutionary computations such as GA. Up to now most studies considered the adaptation of one parameter or one operator only. In this paper, we attempt to combine the advantages of different adaptive mechanisms without falling into the complex interactions on each other that may trigger additional problems. It is achieved by using a number of populations with different adaptive mechanisms. The method also provides a new way for maintaining diversity in a GA: we can achieve diversity not only in the level of individuals, but also in the level of genetic operators, control parameters and even populations. The work is achieved in a serial computer. The proposed method is called CAGA (cooperative adaptive genetic algorithm). It has been tested on a chillers´ optimal scheduling problem