An Adaptive Multiobjective Evolutionary Algorithm for Economic Emission Dispatch

This paper addresses the economic emission dispatch (EED) problem where the goal is to allocate the power output of power generation units to satisfy power demand and minimize the cost and emissions simultaneously. We propose a multiobjective differential evolution algorithm and a reinforcement learning technique to adaptively control the parameters of differential evolution. Moreover, the proposed approach utilizes mating restriction and preferences in mating selection to improve search effectiveness and a dynamically controlled mutation to increase the exploration ability. The proposed ideas and algorithm were examined using four EED test cases. Experimental results showed positive effects of our proposed methods and the competitive performance of our algorithm.