T Mullen and MP Wellman
In a computational market, distributed market agents interact with other agents primarily through the exchange of goods and services. Thanks to a well-developed underlying economic framework, we can draw on a rich source of analytic tools and theoretical techniques for designing individual agents and predicting aggregate behavior. For many narrowly scoped static problems, design of a computational market is relatively straightforward. We consider some issues that arise in attempting to design computational economies for broadly scoped, dynamic environments. These issues include how to specify the goods and services being exchanged, how these market-oriented agents should set their exchange policies, and how computational market mechanisms appropriate for idealized environments can be adapted to work in a larger class of non-ideal environments.