Materialized view selection algorithm—CSSA_VSP

In this paper, materialized view selection problem that can improve OLAP (online analytical processing) query efficiency is studied. In recent years, a new hotspot in artificial intelligence, namely clonal selection algorithm has good global search ability and special memory base that can accelerate searching speed. In order to overcome premature phenomenon, this paper presents Metropolis criterion of simulated annealing algorithm and proposes CSSA_VSP algorithm based on clonal selection algorithm and simulated annealing algorithm to solve materialized view selection problem. Theoretical analysis and experimental verification show: this algorithm has better solution quality and quicker convergence speed than genetic algorithm.

1. Introduction

In order to improve the response speed of OLAP systems, store some intermediate results of searching in the form of real tables in DW. The view that contains intermediate results for searching and is stored in the disk is called materialized view. In the real applications, the limited storage space and the cost to keep the view maintenance lead to impossibility that all the view of data warehouse can be materialized. But the need of high query efficiency requires materialized view as many as possible. How to find the balance point in the three factors is a materialized view selection.