Artificial Intelligence Support for Cooperative Decision Management

We propose an innovative architecture for a unique and important class of decision aids. Traditionally, decision aids have been used to help manage the vast amounts of information that may be pertinent to time-critical decisions. In contrast, our project will focus on developing a tool that advises human decision makers as to when the assumptions which form the basis of their plans are no longer valid. The Cooperative Decision Management Tool (CDMT) will track decision rationale and assumptions explicitly, identify invalid decisions, and efficiently propagate information corrections and updates to all affected individuals in a way not currently possible (effectively providing active information filtering, fusion, and prioritization). The CDMT would be interfaced to current planning, management, and decision support tools, and be useable over computer networks. The CDMT will include several AI concepts which have been combined into a system of theories, capabilities, and formalisms, explicitly to aid the management and coordination of complex cognitive tasks. This system (called Redux) has been extensively studied, extended, and tested with real domain problems by Stanford University's Dr. Charles Petrie, a leading researcher in the field of collaborative problem-solving. We will prove CDMT's feasibility by developing a proof-of-concept, limited prototype in Phase I. BENEFITS: The CDMT will dramatically reduce response time and errors, and will increase the quality of decisions for highly time-constrained, complex, and distributed battle management. Because this effort directly addresses perhaps the most important issues in military operations the potential military market opportunities are enormous.