Vector Evaluated Genetic Algorithm-Based Distributed Query Plan Generation in Distributed Database

Distributed query processing (DQP) determines an optimal query plan, which generates user query results in efficient manner by selecting optimal set of database sites. Multi-objective DQP problems become more complex because a query optimizer has to select optimal, non-dominated QEP’s, query equivalent plans, based on conflicting objective values. In past few years, evolutionary techniques are employed on such problems, although they are unable to get a good balance between efficacy and efficiency in all attempts. A meta-heuristic-based algorithm is presented which determines the combinations of database sites, in response to a query or group of queries. In this paper a technique is proposed for the optimal query plan generation, based on the meta-heuristics, modelled for distributed query processing, through an improved vector evaluated genetic algorithm for generation and selection of optimal query plans on distributed database. The algorithm’s optimization performance is evaluated with other approaches and optimization reliability along with efficiency is benchmarked using performance graphs; comparisons indicate that the vector evaluated genetic algorithm (VEGA) converges better than aggregation-based method (weighted-sum approach). Top-K query plans, average query cost and number of generations are the parameters used for the comparative analysis.