Collaborative decision making in organic business environments

Abstract

This paper presents a multiagent architecture and algorithms for collaborative, self-organizing learning in distributed, heterogeneous and dynamic business systems, where the participating agents have local, incomplete knowledge about the whole business transaction. The agents are self-organized for integrating their individual learning based on the business context. This is illustrated by a supply chain scenario with proactive monitoring of logistics processes. Experiments run on a large real-world order data set indicate that our approach effectively improves the performance of decision making in distributed business systems.