Adaptation using constraint satisfaction techniques

Abstract

Case adaptation, a central component of case-based reasoning, is often considered to be the most difficult part of a case-based reasoning system. The difficulties arise from the fact that adaptation often does not converge, especially if it is not done in a systematic way. This problem, sometimes termed the assimilation problem, is especially pronounced in the case-based design problem solving domain where a large set of constraints and features are processed. Furthermore, in the design domain, multiple cases must be considered in conjunction in order to solve the new problem, resulting in the difficulty of how to efficiently combine the cases into a global solution for the new problem.

In order to achieve case combination, we investigate a methodology which formalizes the process using constraint satisfaction techniques. We represent each case as a primitive constraint satisfaction problem (CSP) and apply an existing repair-based CSP algorithm to combine these primitive CSPs into a globally consistent solution for the new problem. The run time is satisfactory for providing a quick and explicable answer to whether existing cases can be adapted or if new cases would have to be created.

We have tested our methodology in the configuration design and assembly sequence generation domains.

This research is sponsored by the National Science Foundation grant IRI-9208429.