A genetic algorithm for set query optimization in distributed database systems

The problem of query optimization that involves set operations (set queries) to achieve minimum communication costs in a distributed database system is NP complete. The problem is formalized as a constraint problem. In this paper, we present genetic algorithms to minimize the data transmission costs required for a distributed set query processing. Few approaches to the constraints problem in genetic algorithms have previously been proposed. One of them uses penalty functions as an adjustment to the optimized problem. However, our approach is based on appropriate data structures and incorporates the constraints into chromosomes. Simulations have been performed on a distributed database system with up to 700 nodes. The performance evaluation is carried out and compared among different algorithms. Experimental results show that our genetic approach performed well both on the computational effort and on the quality of the solutions.