A New Genetic-Based Approach for Function Optimization

Abstract

We present a novel genetic-based algorithm for optimizing n-D simple-bounded continuous functions. In this paper, we propose a new mutation operator, called rotational mutation. The proposed approach starts from the vertices of the polytope created by the simple bounds, as the initial population. Similar to the conventional genetic algorithm, we calculate the optimum point of each population based on its cost value using the elitism mechanism. Then, we create the new generations based on the proposed rotational mutation and the conventional crossover operators. We have evaluated the algorithm on the two well-known test problems. Experimental results showed that the proposed approach outperforms the conventional genetic algorithm, in terms of the number of generations.