Improvement And Application On Query Algorithm Of Distributed Database

Big data are flooding into our life, in recent decades.About research ondistributed database become a hot once again as era of big data is here.However,animportant indicator of measure a distributed database is query algorithm of distributeddatabase.Traditional query algorithm of distributed database can satisfy therequirements of big data query on single join,but query on multi join is unsatisfactorystill. Require more high efficiency on multi join as the era of big data is a collection ofvarious subjects.The traditional query algorithm of distributed database has beenimminent.This thesis firstly research the basic theory about query of distributeddatabase.Then,go on research detaile technology of the algorithm of distributedSDD-1query and genetic algorithm and about extend and improve the key technologyof them.This thesis,combined with the characteristics of multi query of distributeddatabase and distributed SDD-1query algorithm and genetic algorithm are proposedbased on parallel of SDD-1query algorithm and improved genetic algorithm.Finally,as a large number of experiments to verify, improved query algorithm can greatly fitthe multi join of distributed database,the query cost is greatly reduced.The main work of this thesis can be summarized as follows:(1) Research carefully the basis theory of query distributed database,somecommon query optimization technology and classification and application scenariosabout them.(2)Through the theory of query distributed database, traditional SDD-1queryalgorithm take a long time to generated query plans,this thesis proposed a kind ofadvanced SDD-1algorithm based on parallel to solve it.The new algorithm solvestage of the benefit evaluation and site assembly.The experimental data show that theimproved algorithm obviously reduces the generation time of optimal query,itimprove the efficiency of query.(3) Through the theory of query distributed database, traditional genetic queryalgorithm for generate query plan is not the actual optimal query plan. To solve thisproblem, proposed a improved genetic algorithm that multiple possibility of crossover and mutation based on k-means clustering algorithm. The experimental data show thatthe improved genetic algorithm for generate query plan is the actual optimal queryplan,it improve the efficiency of query.