Demand Side Management Pricing Method Based on LSTM and A3C in Cloud Environment

Abstract

Listen

Translate article

This electronic document is a "live" template and already defines the components of your paper [title, text, heads, etc.] in its style sheet. DSM can effectively reduce the peak load and fill the valley, and improve the stability and operation efficiency of power system. With the development of the power Internet of things, the control structure at the edge of cloud management is becoming more and more popular. For the protection of users' privacy, users' power consumption information can only be used locally after collection and cannot be uploaded further, which makes it difficult for load aggregators to make pricing decisions. This paper proposes a demand side management method based on asynchronous advanced actor critical algorithm (A3C) reinforcement learning algorithm and long short term memory (LSTM) network in cloud edge environment, and solves the problem of lack of foresight in demand side management decision-making through reinforcement learning process; Through the structure of distributed learning and centralized decision-making, the local utilization of user information is realized, and the confusion in actual implementation is reduced; The virtual environment based on LSTM network accelerates the learning process and reduces the cost of algorithm implementation. Through the example analysis, it can be seen that the decision-making method described in this paper can effectively speed up the learning process while ensuring users' privacy, and price decision-making can more accurately grasp the characteristics of users' response behavior and ensure the economy of decision-making.