An Adaptive Genetic Algorithm with Parallel Mutation and Its Performance Evaluation

To solve the problems of slow convergent speed and low convergent precision in the genetic algorithm based on real coding, we define a new index to describe population evolution, population vigor, and present a revised adaptive genetic algorithm. Using the population vigor index, we consider the diversity of a population and the similarity of adjacent populations in a unified frame, and adaptively adjust the probabilities of crossover and mutation. In addition, we reset the reference value of population fitness using mode fitness instead of average fitness and improve the mutation operator using a parallel mechanism. Stimulation results show that the algorithm has a quicker convergent speed and better convergent precision. As an application example, we also employed the algorithm to solve gasoline blending recipe optimization.