为解决小规模航班串编制问题,提出一种简单的非分布式算法,并在单机运行平台进行测试.然而,随着民航企业的迅速发展,航班数量不断增加,非分布式的航班串编制算法已经无法满足实际生产需求.为解决大规模航班串编制问题,提出另外两种基于MapReduce框架的分布式航班串编制算法.第一种算法将简单的非分布式算法扩展到MapRe-duce框架,解决大规模航班串编制问题;第二种算法在第一种算法的基础上进一步改进,优化Map和Reduce的处理流程,删除第一种算法中的迭代过程,充分发挥MapReduce框架的批处理优势.搭建Hadoop平台进行验证,实验结果表明,提出的两种分布式算法中,第二种算法即改进后的分布式算法,较之简单的非分布式算法和第一种分布式算法,能够有效提高大规模航班串编制效率.%A simple algorithm without distribution is proposed to solve the small scale flight string compilation problem and tested on stand-alone operation platform. However,with the rapid development of civil aviation enterprises and the rising number of flights,the simple algorithm has been unable to meet the requirement of practical production. Two new distributed flight string compilation algorithms based on MapReduce frame are proposed. The first one is extended from the simple algorithm to solve the large scale flight string compil-ation problem. And the second is made further improvements on the basis of the former where the processes of Map and Reduce are sim-plified and the iteration is deleted. A Hadoop platform is constructed to verify these algorithms. Results shows that compared to the simple algorithm and the first distributed algorithm,the second improved algorithm could effectively improve the efficiency of compiling flight string with large scale flights.
展开▼
