摘要： 针对遴选算法处理信息量大、条件复杂、运行时间长等特性，研究算法准并行化的方法，运用全局约束量主动推送和全局记录加锁策略对全局条件进行解耦，在此基础上建立一种基于客户端/服务器模式的多线程算法结构，并对2种解决并行同步问题的加锁策略进行比较分析。实验结果表明，优化实现后算法的运行速度有明显提升。

关键词: 遴选算法, 全局条件, 解耦, 多线程

Abstract: Expert-chosen algorithm need process a large amount of data under complex constraints, it is a time consuming work. This paper examines the quasi-parallel approach of the algorithm. The strategies of global constraints status automatic delivering and global record locking are used to decouple the global constraints. Based on this, this paper proposes a client/server multi-threading architecture and conducts a comparative analysis for two locking methods which handles the synchronization. Experimental results show that the running speed of the optimizing algorithm is significantly improved.

Key words: expert-chosen algorithm, global constraints, decouple, multi-thread

中图分类号: 

• TP391