Predicting the Performance of Distributed Knowledge-Based Systems: A Modeling Approach

Book One

Volume

Proceedings of the AAAI Conference on Artificial Intelligence, 3

Search

A model of a distributed knowledge-based system is presented. The model captures the features specific to those systems, such as alternative paths to the solution, utilization of inexact and/or incomplete knowledge and data, dynamic task creation, complex subproblem dependencies and focusing aspect of problem solving. The model is applied to the analysis of communication policies in a distributed interpretation system. The result of the analysis is the best policy for the given environment and system conditions. Another use of the model as a real-time simulation tool is suggested.