An optimization algorithm based on chaotic behavior and fractal nature

In this paper, we propose a new optimization technique by modifying a chaos optimization algorithm (COA) based on the fractal theory. We first implement the weighted gradient direction-based chaos optimization in which the chaotic property is used to determine the initial choice of the optimization parameters both in the starting step and in the mutations applied when a convergence to local minima occurred. The algorithm is then improved by introducing a method to determine the optimal step size. This method is based on the fact that the sensitive dependence on the initial condition of a root finding technique (such as the Newton–Raphson search technique) has a fractal nature. From all roots (step sizes) found by the implemented technique, the one that most minimizes the cost function is employed in each iteration. Numerical simulation results are presented to evaluate the performance of the proposed algorithm.