Distributed Query Plan Generation using Ant Colony Optimization

Query processing is a critical performance evaluation parameter and has received a considerable amount of attention especially in the context of distributed database systems. The aim of distributed query processing is to effectively and efficiently process the query. This entails laying down an optimal distributed query processing strategy that generates efficient query plans Since in distributed database systems, the data is distributed and replicated at multiple sites, the number of query plans increases exponentially with increase in the number of relations accessed by the query along with increase in the number of sites containing these relations. Thus, from amongst these query plans, there is a need to generate optimal query plans involving lesser number of sites which, in turn, would entail lower site-to-site communication cost leading to faster query response times. In this paper, an attempt has been made to generate such query plans for a distributed query using Ant Colony Optimization (ACO). This ACO based distributed query plan generation (DQPG) algorithm, when compared with the GA based DQPG algorithm, is able to generate comparatively better quality Top-K query plans for a given distributed query.