MACC'97 Session: Complex Systems and Multiagent
TRURL is a simulation environment, which evolves artificial worlds of multiagents to socially interact with each other. The micro-level agent activities are determined by both pre-determined and acquired parameters. The formers have constant values during one simulation cycle, however, the latters changes during the interactions. Unlike conventional artificial society models, TRURL utilizes Genetic Algorithms to evolve the societies by changing the pre-determined parameters to optimize macro-level socio-metric measures, which can be observed in such real societies as e-mail oriented organizations and electronic commerce markets. Thus, using TRURL, we automatically tune the parameters up and observe both micro- and macro-level phenomena grounding on the activities of real worlds. This paper first describes basic principles, architecture, and mechanisms of TRURL. Then, to investigate the features of face-to-face-, e-mail-, news group-, and mass- communication-oriented societies, we have carried out intensive experiments. The results have suggested that each society is characterized by features of evolved agents.
Multiagent, Artificial Society, Genetic Algorithms, Social Interaction
PS
file(+gzip) (in Japanese)