Toward a Framework for Evaluation of Cache-Based Decision Making Process

Abstract

While willing to implement any self-learning mechanisms, the decision making process must be partially supported by recalling some previous decisions rather than triggering the reasoning process every time a decision is needed. For relatively simple systems storing previous decisions does not need any sophisticated mechanism but for some complex systems, when there are multiple decisions and many environmental information to be analysed while making those decisions, the whole context information (information about the environment state and accompanying decision) may need to be supported on one hand, by a database system (providing efficient search mechanisms) and on the other hand, by a cache-like system that would even improve decision making process by providing quick access mechanism in the most typical situations. The aim of this paper is to propose a framework for benchmarking real cache management algorithms within embedded systems.