A novel coral reefs optimization algorithm for materialized view selection in data warehouse environments

Abstract

High response time of analytical queries is one of the most challenging issues of data warehouses. Complicated nature of analytical queries and enormous volume of data are the most important reasons of this high response time. The aim of materialized view selection is to reduce the response time of these analytical queries. For this purpose, the search space is firstly constructed by producing the set of all possible views based on given queries and then, the (semi-) optimal set of materialized views will be selected so that the queries can be answered at the lowest cost using them. Various materialized view selection methods have been proposed in the literature, most of which are randomized methods due to the time-consuming nature of this problem. Randomized view selection methods choose a semi-optimal set of proper views for materialization in an appropriate time using one or a combination of some meta-heuristic(s). In this paper, a novel coral reefs optimization-based method is introduced for materialized view selection in a data warehouse. Coral reefs optimization algorithm is an optimization method that solves problems by simulating the coral behaviors for placement and growth in reefs. In the proposed method, each solution of the problem is considered as a coral, which is always trying to be placed and grow in the reefs. In each step, special operators of the coral reefs optimization algorithm are applied on the solutions. After several steps, better solutions are more likely to survive and grow on the reefs. The best solution is finally chosen as the final solution of the problem. The practical evaluations of the proposed method show that this method offers higher quality solutions than other similar random methods in terms of coverage rate of queries.