Research On Multi Join Query Optimization Of Distributed Database Based On Improved Fish Swarm Algorithm

In recent years,distributed database applications become wider and wider,but the problem of optimizing a multi-join query is not solved well in distributed database.The problem of the multi-join query optimization largely influence the efficiency of the data query as the extending of the distributed database,but traditional optimization techniques for multi-join query are inadequate to solve the problem of multi-join queries optimization in distributed database.The complexity of multi-join optimization in distributed database has risen greatly as the increasing of relations which greatly affects the efficiency of searching for the optimal execution plan,so as to impede the further optimization of multi-join queries.Accordingly,we proposes an improved fish swarm algorithm based on strategy selection to solve the problem of further optimizing the multi join query in distributed database.Firstly,the connection set is used to select the strategies to reduce the search space before applying optimization algorithm.Secondly,according to the characteristics of multi join query in distributed database,join trees in the reduced search space are encoded to form one-to-one corresponding coding trees.In order to facilitate the calculation of fitness function of artificial fish,values that store the details of the congruent relations is given to leaf nodes of a join tree respectively.Thirdly,the artificial fish swarm algorithm is used to accomplish the following query optimization,which is insensitive to the initial state,easy to quickly converge,and strong ability to jump out of local extremisms.At the same time,we improve the deficiencies of fish swarm algorithm to search for the high-quality and global optimal execution plan quickly.Finally,the first stage of strategy selection and the second stage of improved fish swarm algorithm are combined to form the final optimization strategy.The simulation experiments are carried out in four steps under different relations:1 selecting best values of fish swarm algorithm parameters,2 validating the effectiveness of strategy selection and the improved fish swarm algorithm,3 comparing the optimization strategy with other stochastic algorithms,4 testing the convergence of the improved fish swarm algorithm.The experimental results show that:1 optimization time efficiency of the one using the strategy selection is better than that not using;optimization time efficiency of the one based on improved artificial fish swarm algorithm is superior to that based on original fish swarm algorithm;2 the optimization strategy we purposed is superior to other algorithms at the aspect of optimization time and execution time;3 the convergence of the improved fish swarm algorithm shows that the optimal execution plan we obtained is of high quality,which can effectively reduce the execution time and make the query execution time the shortest.