Selection of materialized views using stochastic ranking based Backtracking Search Optimization Algorithm

Abstract

Selection of materialized view plays an important part in structuring decisions effectively in datawarehouse. Materialized view selection (MVS) is recognized as NP-hard and optimization problem, involving disk space and cost constraints. Numerous algorithms exist in literature for selection of materialized views. In this study, authors have proposed stochastic ranking (SR) method, together with Backtracking Search Optimization Algorithm (BSA) for solving MVS problem. The faster exploration and exploitation capabilities of BSA and the ranking method of SR technique for handling constraints are the motivating factors for proposing these two together for MVS problem. Authors have compared results with the constrained evolutionary optimization algorithm proposed by Yu et al. (IEEE Trans Syst Man Cybernet Part C Appl Rev 33(4):458–467, 2003). The proposed method handles the constraints effectively, lessens the total processing cost of query and scales well with problem size.