Set query optimization in distributed database systems

Abstract

This paper addresses the problem of optimizing queries that involve set operations (set queries) in a distributed relational database system. A particular emphasis is put on the optimization of such queries in horizontally partitioned database systems. A mathematical programming model of the set query problem is developed and its NP-completeness is proved. Solution procedures are proposed and computational results presented. One of the main results of the computational experiments is that, for many queries, the solution procedures are not sensitive to errors in estimating the size of results of set operations.