A Dynamical Characterization of Case-Based Reasoning Systems for Improving Its Performance in Highly Dynamic Environments

摘要

In this paper a mathematical technique is presented for modelling the generation of solutions in a standard-CBR methodology adapted to highly dynamic environments. In recent years, much research has focused on exploring how to improve CBR-systems to deal with dynamic environments where changes are difficult to model or predict and, consequently, the performance of the CBRs gets worse with time. High-level planning, reactive strategies or hybrid alternatives have been proposed to face this problem, but this contribution will not be related on particular techniques. We simply concentrate on formal aspects without establishing which should be the most adequate procedure in a subsequent implementation stage. The advantage of the presented scheme is that it does not depend on neither the problem nor the model of the environment. It consists in a formal approach that only requires, local information about the averaged-time spent by the system in obtaining a solution and an estimated measure about the dynamism of the environment.