Mobile agents for economic market simulation

This paper reports on the development of an economic market simulator using mobile agents. Agents are specialized software programs that can move autonomously between supporting platforms. Agents can be used to model the preferences of users wishing to buy or sell items on a global marketplace. Users can select a number of options that will guide the agents in their quest for an optimal trade. Besides the price range, users can determine the way in which the agents will bargain, thus fine-tuning the tradeoffs involved in deciding whether a price in the price range requires further negotiation or is acceptable. The simulation developed is realized entirely in Java, and takes advantage of TEEMA, an extensible, general-purpose mobile agent execution environment, dedicated to the purpose of exploring the capabilities of mobile agents. TEEMA has been realized to provide a testbed for agent applications without having to be limited by proprietary application suites or inaccessible low-level interfaces. It is completely configurable and easily extensible as it is built around a microkernel that supports primitive agent operations and a set of services that provide high-level agent support. The paper starts with the description of the infrastructure necessary to define a marketplace as a combination of service agents. Then, trader agents are explained. The simulation environment is then introduced, and analyzed. The architecture emerging from the simulation environment is then examined, together with its advantages and disadvantages. Finally, the simulation parameters are presented.