A novel multiagent system based on dynamic fuzzy cognitive map approach

This paper proposes a novel multiagent system for complex system modeling based on a dynamic fuzzy cognitive map approach. It aims to represent the domain knowledge and carry out the inference process regarding the uncertainty, distribution and dynamism that exist in most of real world problems. The proposed multiagent system architecture is able to model dynamic real world problems that almost contain heterogenous distributed data resources and perhaps physical sensors. It hides the complexity of such data resources by a set of wrapping agents that collect the desired data periodically. An illustrative example shows robustly the ability of applying the proposed multiagent system to a wide spectrum of real world problems like products purchasing domain which is environment sensitive.