Research On Energy Management Scheduling Algorithm And System Based On Reinforcement Learning

In recent years,the green development of ecological economy is generally attached importance by countries all over the world.In order to be able to complete the carbon peak in 2030 and achieve the vision of carbon neutrality in 2060,China has put forward higher requirements on the production methods of manufacturing enterprises.In order to realize the digital transformation of manufacturing industry,scientific and reliable intelligent production scheduling,smart energy management and massive production equipment and data management based on IoT have become important technologies and methods in modern smart manufacturing.To address these issues,this thesis realizes multi-objective flexible job scheduling in the workshop to reduce total production energy consumption based on an improved Q-learning algorithm,and develops a smart energy management system based on the IoT architecture.First,an improved Q-learning multi-objective flexible job shop scheduling algorithm is proposed by combining the knowledge of reinforcement learning theory,and the state space and action space of the Q-learning algorithm are designed according to the characteristics of the Flexible Job Shop Scheduling Problem(FJSP),in which the maximum With the maximum completion time and total energy consumption as the optimization objectives,we use two ways to reduce the probability of the algorithm falling into local optimum,namely,random greedy strategy and randomly generated initial process code.Then,based on the IoT architecture,we design a smart energy dispatch management system BT-IoT for industrial production environment,and design its overall architecture model and functional modules.In the equipment sensing layer,different types of data are collected and aggregated in the workshop using the communication interface of the factory production bus;the network transmission layer realizes the interconnection of sensor network and BT-IoT network through IoT fusion gateway or transmission protocols such as MQTT,HTTP,COAP,etc.;the application service layer uses data visualization technology and IoT technology to establish the real-time monitoring and management cloud platform for energy consumption data,workshop equipment and asset safety.Monitoring and management cloud platform.Finally,the intelligent energy dispatch management system was implemented and tested on the basis of the BT-IoT platform.The data was uploaded and sent to the BT-IoT platform by simulating intelligent devices,and the real-time energy consumption data and production situation of the devices were remotely monitored and managed by the visualization function module of BT-IoT,and then the calculation results of the improved Q-learning algorithm to solve the multi-objective flexible job shop scheduling algorithm were uploaded to the BT-IoT platform devices through the MQTT protocol.The test results show that the smart energy scheduling management system can meet the demand of manufacturing enterprises for supervision of massive data in the production process,and also provide scientific,energy-saving and high-efficiency production planning solutions for enterprises.Thus,the smart energy scheduling management system provides a scientific and reliable reference solution for the digital transformation of the manufacturing industry.