Transferable belief model for decision making in the valuation-based systems

摘要

We present a decision support system which is based on the transferable belief model (TBM), a model to represent someone's degree of belief based on belief functions. The system performs reasoning and decision making by integrating a system for belief propagation and a system for Bayesian decision analysis. The two subsystems are developed within the framework of the valuation-based systems. They are connected through the pignistic transformation as described in the context of the TBM. The system takes as inputs the user's beliefs and utilities, and suggests either the optimal decision or the optimal sequence of decisions. An example concerning a nuclear waste disposal problem is given to demonstrate the applicability of the system in a real-world domain.