The electric utility industry is in the transition from operating in a monopolistic environment to one that is regulated as a competitive marketplace. At this point in time, over eight markets have failed. The objective of this research is to analyze this new emerging electricity marketplace from the viewpoint of including reliability as a product dffferentiator as solved by adaptive agents. Reliability is a key feature of electric energy delivery and has provided national economic growth for almost a century. This research is interested in how a company would compete effectively for long term growth potential and for continuous financial performance.
This is a computer simulation of the portfolio analysis of various electric contracts with various levels of reliability. All players must use complex decision making to properly bid and to make a positive profit. This effort proposes to use classical optimization techniques in conjunction with decision analysis for the buyers and sellers of electricity in such an electric marketplace. This project will investigate the steady state results for the application of forecasting and decision methods to determine if price discovery occurs using various learning algorithms. Previous research by this author and others has lead to new algorithms based on real option analysis.
This effort proposes computer-learning algorithms (Modified Roth-Erev, genetic algorithm, genetic programming, and neural networks). Agents will be implemented with these techniques to generate a sustainable profit from the various energy markets. The long-term goal is to develop adaptive agents to assist humans who play the electric marketplace.
