Simplex Differential Evolution

Differential evolution (DE) algorithms are commonly used metaheuristics for global optimization, but there has been very little research done on the generation of their initial population. The selection of the initial population in a population-based heuristic optimization method is important, since it affects the search for several iterations and often has an influence on the final solution. If no a priori information about the optima is available, the initial population is often selected randomly using pseudorandom numbers. In this paper, we have investigated the effect of generating the initial population without using the conventional methods like computer generated random numbers or quasi random sequences. We have applied non linear simplex method in conjugation of pseudorandom numbers to generate initial population for DE. Proposed algorithm is named as NSDE (using non linear simplex method), is tested on a set of 20 benchmark problems with box constraints, taken from literature and the numerical results are compared with results obtained by traditional DE and opposition based DE (ODE). Numerical results show that the proposed scheme considered by us for generating the random numbers significantly improves the performance of DE in terms of convergence rate and average CPU time.