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

这个小企业创新研究的第一阶段项目涉及开发一个铁路阻塞问题的决策支持系统，铁路阻塞问题是铁路网络上车辆运输中最重要的优化问题之一。该公司将利用最先进的优化技术、先进的数据结构和信息技术工具开发一系列软件产品，使其决策过程自动化，从而提高其运营效率并降低成本。铁路阻塞问题的决策支持系统是实现这一目标的第一步。铁路并没有从优化领域的进步中受益，他们依靠人工决策过程来满足大部分的计划和调度需求。铁路阻塞问题是一个非常大规模的优化问题，不能用最优的算法思想来解决。提议的工作将使用PI开发的启发式算法来解决这个问题，使用一种新兴技术，称为超大规模邻居（VLSN）搜索。提议的第一阶段研究包括将几个额外的特征纳入该算法，并对软件进行彻底的计算测试。该项目的成功将导致铁路行业对优化模型和基于优化的软件的更大接受，并将为解决其他几个同样重要的问题（包括块到列车分配、列车调度、机车调度和机组调度）的新软件产品铺平道路。从长远来看，这将提高美国铁路的效率和盈利能力。预计该软件的使用将为美国任何主要铁路公司每年节省数千万美元。

This Small Business Innovation Research Phase I project concerns developing a decision support system for the railroad-blocking problem, one of the most important optimization problems arising in the shipment of cars over the railroad network. It will develop a series of software products using state-of-the-art optimization techniques, advanced data structures, and information technology tools to automate their decision-making process, thereby improving their operational efficiency and reducing costs. A decision support system for the railroad-blocking problem is the first step towards this goal. Railroads have not benefited from the advances taking place in the field of optimization and they rely on manual decision-making process for most of their planning and scheduling needs.The railroad-blocking problem is a very large-scale optimization problem and cannot be solved to optimality using the state-of-the-art algorithmic ideas. The proposed work will use a heuristic algorithm developed by the PI to solve this problem using an emerging technique, known as the Very Large-Scale Neighborhood (VLSN) Search The proposed Phase I research consists of incorporating several additional features to this algorithm and performing a thorough computational testing of the software. The success of this project will lead to greater acceptance of the optimization models and optimization based software in the railroad industry and will pave the way for new software products for several other equally important problems including the block-to-train assignment, train scheduling, locomotive scheduling, and crew scheduling. In the long run, this will lead to more efficient US railroads with improved profitability. It is anticipated that the use of this software will result in an annual savings in tens of millions of dollars for any major US railroad.