Distributed Genetic Algorithm-based Multi-join Query Optimization System Design And Implementation

In recent years, as more and more distributed database applications appear, distributed database query optimization research became a hot research field of distributed database. However, Multi-join query optimization problem has not been well resolved, and traditional database query optimization techniques for distributed multi-join query optimization problem also appears to do nothing, with the growing size of distributed database, Multi-join queries of distributed database increasingly affect the overall efficiency of a distributed database system.This paper introduces the research background and significance of multi-join query optimization, and in-depth study the genetic algorithm and distributed multi-join query optimization theories, the key technologies has been discussed and improved. On this basis, this paper combines the characteristics of genetic algorithms and the special application direction of distributed multi-join queries, and based on object-oriented approach, XML and UML technologies, and designs and implements a distributed multi-join query optimization system GABQO. This major work done can be summarized as:(1) In-depth study of the genetic algorithm and the distributed multi-join query optimization theory, the key technologies are discussed, and demonstrated a distributed multi-join query optimization is necessary.(2) Design a query optimization system GABQO based on genetic algorithm, GABQO system includes three modules: GABQO-Framework System Framework, GABQO-Lib library of genetic algorithm development module and GABQO-Query distributed database access module.(3) Improved genetic algorithm in distributed multi-join query optimization. GABQO system for left-deep linear tree search space to propose new encoding method, and crossover and mutation operator, the mutation operator has added the "only allows better mutation" mechanism, to speed up the convergence and improve the efficiency of query. Added to the fitness function of "Bonus-Penalty" mechanism, accelerated the emergence of good chromosome and the elimination of poor chromosome.(4) In the simulation environment, through trial and get a best set of parameters for our GABQO system, and with this group of parameter values for distributed multi-join query optimization, achieved a desired result.