An agent-based proxy cache cleanup model using fuzzy logic

Abstract

This paper describes the use of fuzzy logic to improve the performance of the proxy cache. A multi-agent system has been developed to control the cache cleanup process on both parent and child sides. Fuzzy logic is used to combine LFU and LRU caching replacement policies on the child cache. Size, LFU, and LRU policies are used on the parent cache side. Agents use fuzzy logic to make an intelligent decision about clean up priority. Simulation results achieved a hit ratio of 86.36% in the best case and 73.33% in the worst case, and a byte hit ratio of 86.62% in the best case and 64.46% in the worst case on the parent cache side. Results for the child cache side are a hit ratio of 87.50% in the best case and 76.47% in the worst case, and a byte hit ratio of 89.04% in the best case and 64.76% in he worst case.