A stochastic model for constructing an access policy for multiple knowledge systems

Abstract

The paper describes a paradigm in which the Dempster-shafer method of combining evidence from independent sources is used for two distinct but closely related purposes. The first of these is to obtain the answer to a specific query, while the second is to define a dynamic policy for parallel accessing of relevant knowledge systems. The access policy is determined by a sequence of goals which are themselves subdivided into control characteristics of potential knowledge systems Basic probability assignments (bpa), conditioned on specific performance objectives and potential feature values, are the basic tool that drives the inference in both instances. These conditional bpa's are weighted by different probabilities depending on the knowledge sources accessed. Information from independent sources is combined via the Dempster-Shafer Rule of Combination, and belief functions are employed to make decisions