Collaborative and scalable financial analysis with multi-agent technology

During the past two years, business world started to use clusters of PCs and workstations, multi-agent technology, Java programming, Intranet, and Internet to support financial computing, decision analysis, electronic business, etc. In this paper, we integrated these technologies and present a distributed multi-agent model for network-based financial computing or economic analysis. In particular, we focus on the potential to apply cluster computing to finance, business, investment, insurance, and socioeconomic application domains. These commercial domains demand special computing characteristics, which differ in many ways from the scientific computing requirements. A prototype of the distributed financial computing has been implemented to illustrate the idea of multiagent computing on workstation clusters. The cluster platform and the multi-agent model has been proved to be applicable in other domains such as electronic business, market analysis, financial decision making, and economic crisis management. A three-tier multiagent architecture has been designed to capture the structure of Hong Kong economic activities. The system calculates risks and returns based on more than thirty financial indicators that cover major business activities in Hong Kong. One more application of using such multi-agent system is to generate a composite economic index to measurement the economic performance.