CAJ | 학술논문

针对分布式数据库多表查询速度慢的问题，提出一种改进的分布式数据库查询优化遗传算法。利用条件采样的方法，维持种群的多样性，防止算法陷入局部最优解；利用马氏链模型优化变异算子，确定变异算子当前状态下的最优取值，进行交叉和变异操作，找出最优查询执行计划。仿真结果表明，改进算法能在较短时间内找到最优的查询执行计划，加快查询速度，提高查询效率。
침대분포식수거고다표사순속도만적문제，제출일충개진적분포식수거고사순우화유전산법。이용조건채양적방법，유지충군적다양성，방지산법함입국부최우해；이용마씨련모형우화변이산자，학정변이산자당전상태하적최우취치，진행교차화변이조작，조출최우사순집행계화。방진결과표명，개진산법능재교단시간내조도최우적사순집행계화，가쾌사순속도，제고사순효솔。
An improved query optimization method based on genetic algorithm is proposed to solve the problem of the slow multi-table query speed of distributed database.A conditional sampling method is used to maintain the diversity of popula-tion in case it traps into local optima.The mutation operator is optimized by using Markov-chain model to decide its optimal value under the current state,then crossover and mutation operator is proceeded to find out the optimal query execution plan.Simulation results show that the proposed algorithm can find the optimal query execution plan in a short time.It can speed up the query process and improve the efficiency of query.