Exploring and exploiting complex behavior in self-organizing multi-agent systems

Through experiments and analysis, we discover the phenomena of self-organized criticality (SOC) arising in self-organizing multi-agent systems (SOMAS) rather than in non-self-organizing MAS. Furthermore, we attempt to utilize the phenomena of SOC in local strategies of single agents and improve the performance of SOMAS. In addition, we examine how the global patterns evolve in SOMAS as local parameters of individual agents change, and attempt to explain the critical values of the parameters based on Shannon´s entropy theory.