Optimal pricing and purchasing strategy in energy transaction based on reinforcement learning

Abstract

The invention discloses an optimal pricing and purchasing strategy in energy transaction based on reinforcement learning. A middleman firstly observes a market environment state and selects a pricing behavior through a middleman strategy network to determine an electric energy price; a micro-grid observes the market environment state and selects purchasing behaviors through a micro-grid strategy network to determine the electric energy purchasing quantity; the market receives behaviors of the middleman and the microgrid and interacts with the middleman and the microgrid, the market state, the behavior set, the reward function and the market state at the next moment jointly form a transition set, and the transition state at the previous moment is stored in an experience warehouse to be used for updating the strategy network; the network is learned and trained, and finally the best middleman pricing strategy and micro-net purchase strategy are acquired. According to the reinforcement learning method, the overall economic benefit of the market can be effectively improved, the demand response can be realized, and the pressure of electric energy supply and distribution can be reduced.

Description

technical field [0001] The invention belongs to the technical field of power system data security and control, and relates to optimal pricing and purchasing strategies in energy transactions based on reinforcement learning. Background technique [0002] Smart grid is a fully automated new power transmission network. Microgrid is an important part of smart grid. As a local power generation and distribution system, in a microgrid, local production units (such as wind energy, solar energy, etc.) can meet the requirements of the microgrid. Part of the power demand of the microgrid, but these unstable production capacity cannot meet all the needs of the microgrid, so the microgrid also needs to purchase power from the main grid. However, the centralized main grid cannot regulate the relationship between supply and demand, and cannot adjust the supply for the demand of the microgrid. In order to solve this problem, middlemen are introduced into the electric energy market. The mid...

AI Technical Summary

Problems solved by technology

2. In the electric energy trading market, the electricity price and the demand of the microgrid are unpredictable, which also makes it difficult to model the electric energy trading market. Therefore, the model-based reinforcement learning method is actually difficult to be applied to the actual In the electricity market

Images