Research in Mixed-Integer Programming

该项目的目标是开发求解大规模混合整数程序的算法和原型软件，并开发用于解决分布，制造，生产调度和资源分配中出现的问题的应用程序。本研究的动机是在上述领域对高效、鲁棒的混合整数规划算法的巨大需求。这些研究人员在之前的项目中开发了一个混合整数优化器（Minto）系统。MINTO系统允许研究人员通过应用程序例程定制通用混合整数优化代码，并专注于问题细节，而不是像线性规划这样的工具的实现细节。该项目将研究加强MINTO能力的方法。并将研究在配送、调度、制造和资源分配等通用问题的应用程序开发。实际应用数据将用于测试定制系统。

The goal of this project is to develop algorithms and prototype software for solving large-scale mixed-integer programs and to develop application routines for solving problems that arise in distribution, manufacturing, production scheduling and resource allocation. The motivation behind this research is the tremendous demand for efficient and robust mixed-integer programming algorithms in the preceding domains. A Mixed-Integer Optimizer (Minto) system was developed by these researchers in prior projects. The MINTO system allows a researcher to customize general mixed- integer optimization code through application routines and concentrate on problem specifics rather that implementation details of tools like linear programming. The project will research methods to strengthen the capabilities of MINTO. It will also research the development of the application routines for generic problems in distribution, scheduling, manufacturing and resource allocation. Real application data will be used to test the customized systems.