Software Agents and the Information Economy

Monday, November 3, 2003 - 11:35am - 12:10pm
Keller 3-180
Jeffrey Kephart (IBM)
Humans are on the verge of losing their status as the sole economic species on the planet. In private laboratories and in the Internet laboratory, researchers and developers are creating a variety of autonomous, economically-motivated software agents endowed with algorithms for maximizing profit or utility. Many economic software agents will function as miniature businesses, purchasing information inputs from other agents, combining and refining them into information goods and services, and selling them to humans or other agents. Their mutual interactions will form the information economy: a complex economic web of information goods and services that will adapt to the ever-changing needs of people and agents. The information economy will be the largest multi-agent system ever conceived, and an integral part of the world's economy.

One cannot predict how this new world economy will behave simply by extrapolating from hundreds of years of economies in which humans have been the only players. Economic software agents differ from their human counterparts in several ways. They operate more quickly on more up-to-date and accurate information, yet on the other hand they have much less world experience and common sense. In an effort to both understand and design the macroeconomic behavior of the future information economy, we have simulated several different markets and economies populated with economic software agents employing a variety of plausible pricing and bidding algorithms. I will present several interesting macroeconomic behaviors that we have observed and analyzed, including cyclical price wars and complex strategic interactions that are reminiscent of the prisoner's dilemma. I will then discuss how insights gained from our studies can be used to design not just market mechanisms, but the agents themselves -- an opportunity that traditionally has not been available to economists.