Date of Award
Master of Science
Electrical and Computer Engineering
The recent deregulation of the electric industry in the United States opened some sectors of the power market to competition. This work addresses a computational restructured wholesale electricity market. The goal of the study is to model agent driven bilateral power market auctions where the players are represented by autonomous intelligent agents. Different aspects of the market are considered. Some of them are studies on structural and strategic market power of buyers and sellers varies with changes in relative concentration and relative capacity. Others are cases where players attempt to benefit from causing instabilities like brownouts and blackouts, as well as economic instabilities by applying different gaming strategies. Agents are modeled using various evolutionary programming techniques, such as reinforced learning, genetic algorithms and genetic programming. The results suggest that some of the solutions are suitable for robust industrial applications.
Valentin Tzankov Petrov
Petrov, Valentin Tzankov, "Exploring computational power markets with evolutionary algorithms" (2002). Retrospective Theses and Dissertations. 20201.