Enhanced Decision Modeling Using Multiagent System Simulation

摘要

Modeling and simulation is being used more and more to capture our physical world in a computational form. However, simulating the mental processes that go on within the mind is a complex task that has met with far less success. To improve on this situation, the author developed a computational model of the decision process used by experienced decision makers. This model has, as its basis, the cognitive decision theory entitled naturalistic decision making and its specific implementation, called recognition primed decision (RPD) making. Specifically, this research developed a model that mimicked the decision process of a senior military commander at the operational level of warfare. The model, called RPDAgent, was validated against decisions made by actual military officers. RPDAgent produced decisions that were equivalent to its human counterparts-they were not optimum decisions, but they reflected the variability inherent in those made by humans in an operational military environment.