SBIR Phase II: A Decision Support System for the Railroad Blocking Problem

这个小企业创新研究计划(SBIR)第二阶段项目需要为铁路阻塞问题开发一个决策支持系统，这是货运铁路运输最重要的优化问题之一。铁路运输问题的数学复杂性阻碍了解决这些问题的优化算法的开发，使铁路无法从优化领域取得的进展中受益；它们仍然依赖于人工决策过程来满足其大部分计划和调度需求。在第一阶段，该公司为铁路阻塞问题开发了原型软件，并使用三家美国主要铁路公司提供的数据对其进行了测试：CSX运输公司、BNSF铁路公司和诺福克南方公司。在这个第二阶段的项目中，该公司将通过将最先进的运筹学技术与最新的信息技术工具相结合，为铁路阻塞问题开发一个商业决策支持系统的原型。该项目将利用网络优化、启发式优化、数据结构和软件工程方面的前沿思想来增强核心优化引擎和算法。还将提供数据库连接。这个第二阶段的项目将把铁路阻塞问题的算法扩展到邮政/包裹递送服务设计和开发原型软件时出现的类似问题。目前，铁路需要几个月的团队努力来确定阻塞计划，并每隔几年进行一次这项工作，并断断续续地进行微小调整，以适应交通模式的季节性变化。拟议的决策支持系统将允许铁路在几个小时内确定阻塞计划，并产生远远优于人工获得的解决方案。拟议的解决方案将使大型货运铁路能够经常优化其阻塞计划，并每年为美国和加拿大的铁路公司减少至少1000万美元的成本，并在几年内为其减少数亿美元的成本。这项研究将确立网络优化和启发式方法在解决铁路计划和调度问题中的有效性。该产品的成功将使优化模型和基于优化的软件在铁路行业得到更大的接受。

This Small Business Innovation Research Program (SBIR) Phase II project entails developing a decision support system for the railroad-blocking problem, one of freight railroad transportation's most significant optimization problems. The mathematical complexity of railroad transportation problems has precluded the development of optimization algorithms for solving them preventing railroads from benefiting from the advances taking place in the field of optimization; they still rely on manual decision-making processes for most of their planning and scheduling needs. During Phase I, the company developed prototype software for the railroad blocking problem and tested it on the data provided by three major US railroads: CSX Transportation, BNSF Railway, and Norfolk Southern Corporation. In this Phase II project the company will develop a prototype for a commercial decision support system for the railroad-blocking problem by combining state-of-the-art operations research techniques with latest information technology tools. This project will enhance core optimization engines and algorithms using cutting-edge ideas in network optimization, heuristic optimization, data structures, and software engineering. Database connectivity will also be provided. This Phase II project will extend algorithms for the railroad-blocking problem to similar problems arising in postal/package delivery service design and developing prototype software. Currently, railroads takes months of team effort to determine a blocking plan and undertake this exercise once in several years with intermittent periods of minor adjustments to account for seasonal variations in the traffic pattern. The proposed decision support system would allow a railroad to determine a blocking plan in a matter of a few hours and produce solutions far superior than those obtained manually. The proposed solution will enable a large freight railroad to optimize its blocking plans frequently and reduce cost by at least $10 million annually and hundreds of millions of dollars for railroads companies in the USA and Canada over a few years. The research will establish the efficacy of network optimization and heuristic methodology in solving railroad planning and scheduling problems. The success of this product will lead to a greater acceptance of optimization models and optimization-based software in the railroad industry.