Reinforcement-Learning-Based Multi-Objective Differential Evolution Algorithm for Large-Scale Combined Heat and Power Economic Emission Dispatch

As social and environmental issues become increasingly serious, both fuel costs impacts should be considered in the cogeneration process. In recent years, combined heat power economic emission dispatch (CHPEED) has a crucial optimization problem system management. this paper, novel reinforcement-learning-based multi-objective differential evolution (RLMODE) algorithm is suggested to deal with CHPEED considering large-scale systems. RLMODE, Q-learning-based technique adopted automatically adjust control parameters of algorithm. Specifically, Pareto domination relationship between offspring solution parent used determine action reward, most-suitable parameter values for environment model are adjusted through Q-learning The proposed RLMODE was applied solve four problems: 5, 7, 100, 140 generating units. simulation results showed that, compared well-established algorithms, achieved smallest cost all problems. addition, acquired better Pareto-optimal frontiers terms convergence diversity. superiority particularly significant two