Evolutionary Dynamics of Four Pricebot Algorithms

Amy Greenwald and Victoria Manfredi

July 31, 2001

The work presented in this paper is a continuation of work done by Amy Greenwald and Sze Lok Audrey Yau during Fall 2000 and Spring 2001.

Abstract:

This paper looks at different pricebot algorithms, compares how well they do against each other and determines which algorithm does the best. Pricebots are agents that determine the best price at which to sell an item [1]. In particular this paper compares the game-theoretic, myopically-optimal, derivative-following, and Q-learning algorithms [1,2,3]. We look at the effects of evolution on a population of pricebots using these various dynamic pricing algorithms in an evolutionary simulation; a sort of ``survival of the fittest.''