Agent-oriented metasynthetic engineering for decision making

摘要

Decision making plays a more and more important role in present society, especially in dealing with open complex giant systems, such as macroeconomic decision making. A feasible and effective approach for dealing with the increasingly remarkable decision making problems, in particular open complex giant systems, is to construct a Hall for Workshop of Metasynthetic Engineering (HWME), which is essentially an open giant intelligent decision support system in itself. On the other hand, the emergent agent technology is proposed and popularized as a new computing paradigm for building complex software systems. In this paper, we first briefly introduce the principle of metasynthetic engineering and the HWME in terms of decision support, and propose the network-based hierarchical model of an HWME. Furthermore, we discuss the design idea of agent-oriented metasynthetic engineering for decision support systems. We discuss the design scheme, the framework and the working mechanism of an HWME based on intelligent agents in Java which took the macroeconomic decision making as an instance. In addition, we further examine the fundamental problem of designing agent-oriented open giant intelligent decision support systems. The suitability and challenges of the present agent-oriented abstractions for open giant intelligent systems are further analyzed. Then, we propose the concept of societal abstraction for agent-oriented open giant intelligent systems, which is based on the following hypothesis: a unified mechanism societal interactions, organizational schemes and component patterns, and multi-agent society. Our preliminary study reveals that the agent-oriented metasynthetic engineering is valid and powerful for decision making of open complex giant systems.