Research Statement 1. Previous and Current Research Activities 1.1 Phd Dissertation Research: a Multi-agent Intelligent System for Analysis of Us Wholesale Electricity Markets 1.2 Ms Thesis Research: Robust Steganography Using Animations Motivation

Motivation During the summer of 2000, wholesale electricity prices in California were approximately 500% higher than those during the same months in 1998-1999. There is a need for a model that could systematically explain the anomalous event. Key contributions a. A formal model of a market structure One of the main deterrents in constructing a large complex system is the lack of a formal representation. I proposed a novel mathematical model to represent a market structure (environment) and a trader's (agent's) bidding strategy. [1] A simulation of the MAIS identified an agent's optimal policy to enhance its reward in a two-settlement market. [2] Funding agency: Central Research Institute of Electric Power Industry (2004) b. Multi–agent learning algorithms Based on the two laws of human learning, I developed two types of probabilistic learning algorithms to learn a bidding strategy. [3] The algorithms were validated using PJM market data. [4] The proposed algorithms performed as well as other machine learning algorithms in terms of forecasting for the PJM market price. [5][6] c. A complex software system I extended the MAIS to incorporate multiple two-settlement markets, where the agents can participate in multiple markets. The extended learning algorithm included a transmission flow constraint between any two markets. [7] The complex software system was validated with the California market data. [8] Funding agency: The Telecommunication Advancement Foundation (2005) d. Systematic analysis techniques I formulated a variance decomposition algorithm to perform sensitivity analysis of the various market fundamentals on the California market. [9]