Genetic algorithm-based simulation of electric power markets

The electric power industry has attracted much attention in the past decade following the movement toward deregulation. This movement has the potential to lead to greater profits for electricity producers and consumers. It requires a shift in thinking, however, as companies used to a regulated industry learn to treat electricity as an economic commodity. Economic markets are a complex area of study. Due to incomplete information and occasional irrationality on the part of market participants, they have the potential to careen wildly away form theoretical predictions. Electric markets in particular, having been regulated for so long, have had a bumpy re-entry into the atmosphere of de-regulated capitalism. For all entities vested in the electric power industry, with this re-entry comes the need to protect themselves from risk as well as new opportunities for profit. This thesis explores the use of genetic algorithms to learn profit-maximizing strategies in a variety of simulated electric markets