中文摘要：

本文提出了以Hadoop为核心的大数据分析平台架构，并设计了基于MapReduce的分布式预测算法、动态票额预分、敏捷票额调整等流程，构建了动态票额智能预分系统，使得客票销售形成“预测、预分、监控、调整、再预测”的闭环流程，实验数据证明，该系统可有效提升铁路客运产品收益。

英文摘要：

This article proposed a architecture of big data analysis platform with Hadoop as the core, designed MapReduce based distributed prediction algorithm, the process of dynamic ticket pre-assignment and agile seats adjustment, implemented the Railway Intelligent Dynamic Ticket Pre-assignment System. The System made railway ticketing form a closed loop process, which concluded prediction, pre-assignment, monitoring, adjustment, and re-prediction.The experimental data showed the System could effectively enhance the revenue of railway passenger transport products.