Building software agents for power trading agent competition

Abstract

Traditional grids have been used to distributed electricity to the consumers, however the electric supply is steadily becoming unstable as the electricity demand increases. Hence, this issue brings the need of a smart grid due to its ability to tackle this issue. The role of the smart grid is to be able to shift electricity consumption with various methods implemented by the agent. This paper will discuss on the role of the agent in the context of the smart grid, how the agent uses reinforcement learning for decision making, and how reinforcement learning is implemented.