An Adaptive Learning Automata for Genetic Operators Allocation Probabilities

The conventional Genetic algorithms (GAs) use a single mutation operator for whole population, It means that all solutions in population apply same leaning strategy. This property may cause lack of intelligence for specific individual, which is difficult to deal with complex situation. Different mutation operators have been suggested in GAs, but it is difficult to select which mutation operator should be used in the evolutionary process of GAs. In this paper, the fast learning automata is applied in GAs to automatically choose the most optimal strategy while solving the problem. Experimental results on different benchmark problems determines that the proposed method obtains the fast convergence speed and improve the performance of GAs.