Research On Combined Heat And Power Economic Emission Dispatch System Based On Reinforcement Learning Multi-Objective Differential Evolution Algorithm

With the development of economy,human’s demand for energy is increasing,leading to the problem of energy shortage and environmental pollution is becoming more and more urgent,energy saving and emission reduction has become a hot issue in the current development.The energy conversion efficiency of traditional power generation is low.The energy conversion efficiency of traditional power generation is low,but combined heat and power technology can improve the energy conversion efficiency and play an important role in saving energy and reducing emissions.With the research of combined heat and power technology,the research of combined heat and power technology is gradually changing from the traditional single objective static dispatch to the complex multi-objective dynamic dispatch.This paper mainly studies the problems of the combined heat and power dynamic economic dispatch(CHPDED),combined heat and power economic emission dispatch(CHPEED)and combined heat and power dynamic economic emission dispatch(CHPDEED).The main research contents are as follows:1.The mathematical model of CHPDED problems is established,and a reinforcement learning differential evolution algorithm is proposed to solve CHPDED problems.This algorithm divides the population into elite individuals and ordinary individuals based on their fitness information.Elite-guided mutation operator,crossover operator and selection operator are used to update individuals to improve the convergence performance of the algorithm.In addition,the reinforcement learning differential evolution algorithm uses reinforcement learning strategy to update the scaling factor,further improving the optimization ability of the reinforcement learning differential evolution algorithm.In the simulation experiments,the reinforcement learning differential evolution algorithm is used to solve the CHPDED problem considering 24 scheduling periods.The experimental results show that the algorithm is more economically efficient than other mature single-objective optimization algorithms,verifying the feasibility of the algorithm in solving the CHPDED problem.2.The mathematical model of CHPEED problem is established,and a reinforcement learning multi-objective differential evolution algorithm is proposed to solve the CHPEED problem.In the reinforcement learning multi-objective differential evolution algorithm,the population individuals are divided into several frontiers by fast non-dominated sorting,and the individuals with better diversity in each frontier are screened by the crowding distance calculating.The population individuals are ranked according to the above two parameters,and the population is divided into elite individuals and ordinary individuals according to the ranking.The elite individual guided mutation and reinforcement learning strategy were used to update the ratio factor and improve the optimization ability of the algorithm.The simulation is carried out on two small scale and two large scale CHPEED systems show that the Pareto front obtained by the reinforcement learning multi-objective differential evolution algorithm is more extensive and uniform,and close to the real Pareto front.The performance index also shows that the reinforcement learning multi-objective differential evolution algorithm is better.The feasibility of reinforcement learning multi-objective differential evolution algorithm for solving CHPEED problem is verified.3.The mathematical model of CHPDEED problem is established,and the reinforcement learning multi-objective differential evolution algorithm is used to solve the CHPDEED problem.Compared with the problem of CHPEED,the dynamic demand change of 24 scheduling periods is added,which is more suitable for the real production situation,but the complexity of solving the problem is further increased.The simulation is carried out on three small scale and one large scale CHPDEED system,which fully considers various constraints.The experimental results show that the Pareto front obtained by reinforcement learning multi-objective differential evolution algorithm is more extensive and uniform,and close to the real Pareto front of the problem,performance indicators also show that reinforcement learning multi-objective differential evolution algorithm is better.The feasibility of reinforcement learning multi-objective differential evolution algorithm in solving the CHPDEED problem is verified.