Applying Ant-based Multi-Agent Systems to Query Routing in Distributed Environments

Abstract

This paper presents SemAnt, a novel ant-based multi-agent system designed for distributed query routing. While the ant metaphor has been successfully applied to network routing both in wireless and fixed networks, little is yet known about its applicability to the task of query routing in distributed environments. We point out the similarities and dissimilarities between routing of data packets and routing of queries, and we present the design of SemAnt, which is based on the ant colony optimization meta-heuristic. For experimental evaluation, we deploy the algorithm in a peer-to-peer environment with a real-world application scenario and compare its performance against the well-known k-random walker approach. As we show, the benefits of SemAnt are that the routes for queries are optimized according to their popularity, and that the algorithm is highly suitable for volatile environments