A genetic algorithm based dynamic pricing for improving bi-directional interactions with reduced power imbalance

Abstract

Active supply-demand interactions in a smart grid are essential for reducing grid power imbalance which is important for the security and efficiency of power supply. A key element to the success of such interactions is the proper pricing strategy. The latest game-theory based dynamic pricing methods require information exchanges not only between the supply and demand sides, but also among individual buildings, since they make decisions for one building's demand response based on/influenced by operations of the others. However, in practical applications in which a number of buildings are considered, the latter information exchanges are extremely difficult due to the concerns of privacy, communication complexity and high computation load. Therefore, this study proposed a genetic algorithm based dynamic pricing method for improving bi-directional interactions with reduced power imbalance, which does not require information exchanges among individual buildings. In this study, at the demand side, targeting at minimizing daily electricity cost, a non-linear programming based demand response control is performed in individual buildings at a dynamic price given by the grid operator genetic algorithm optimizer. Targeting at reducing grid power imbalance, the genetic algorithm optimizer is used by the grid operator to search for a better dynamic price based on the aggregated demand response results. Such interaction will continue until the grid power imbalance cannot be further reduced. The impacts of demand elasticity are also investigated on performance improvements. The proposed pricing method can be used in practical applications to improve dynamic pricing of a smart grid for reduced grid power imbalance and thus increased operation efficiency.