Dual Attractive Centers Optimization: A Simple and Efficient Approach for Real Function

A new global optimization called dual attractive centers optimization (DACO) for optimizing possibly nonlinear and non-differentiable continuous space function is proposed. DACO has two attractive centers, the best attractive center and a random attractive center. The best one is for local search, and the random one for global search. The algorithm´s performance was studied using a test bed of real valued functions with different degrees. In all cases studied, DACO performed very well. The new method requires only two control variables, population size and mutation probability, easy to use.