Adaptation of Genetic Algorithms for Optimization Problem Solving

Since genetic algorithms (GAs) are inspired from the idea of evolution, it is natural to expect that besides original problem conversion to suit GA, also adaptation will be used for tuning internal parameters of genetic algorithms to speed up the iterative solution of the given problem. In this paper we present an approach for modifying crossover and mutation probability rates based on generation index. The crossover probability rate decreases, and mutation rate increases linearly with the generation index. We call it Linear Adapted Genetic Algorithm (LAGA) and we describe different modifications of genetic algorithms for multivariable function optimization. It is essential to have two characteristics in GAs for optimizing multivariable functions. There is always a problem of balancing global and local search in the solution space to speed up the optimization algorithm. The balance might be efficiently controlled by modification of internal parameters crossover rate value, Pc and mutation rate, Pm. The proposed approaches will control these parameters in order to get better performance of a GAs for multivariable functions optimization. These approaches are compared with a standard GA in solving unconstrained optimization problem while minimizing or maximizing several multivariable functions with varying degree of complexity.