Modifications for the Differential Evolution Algorithm

Differential Evolution (DE) is a method of optimization used in symmetrical problems and also that are not even continuous, noisy change over time. DE optimizes problem with population candidate solutions creates new per generation combination existing rules according to discriminatory rules. The present work proposes two variations for this method. first significantly improves the termination by proposing an asymptotic rule, which based on differentiation average function values DE. second modification scheme critical parameter method, method’s ability better explore search space objective function. proposed have been tested number from current literature, experimental results, it appears modifications render quite robust faster large-scale problems.