SBIR Phase I: OptimizerAero - A robust production scheduling optimizer for aerospace manufacturers

这个小企业创新研究（SBIR）第一阶段项目的更广泛的影响将使美国制造业基地更具竞争力。航空航天供应链包含数千家数字化程度较低的中小企业（SME）制造商。中小企业在全国范围内构成了一个庞大的航空航天供应基地，并且在很大程度上没有受到运筹学进步的影响。因此，当地制造商与低成本国家的制造商进行不利的竞争。虽然优化技术的计算能力和速度已经增加到现在可以在真实的时间内解决大规模工业问题的程度，但是很少关注需要做出的许多建模决策，以准确地将真实的工厂物理的复杂性捕获到规定的数学模型中。 该项目将为中小企业开发一种即插即用软件，这一估计为8.5亿美元的市场将提高整个供应链沿着效率。由此产生的解决方案预计将是一个生产优化器，可以在不到两周的时间内在任何中小企业航空航天制造商使用其现有的数据流实施。 该项目中所体现的数字技术和运筹学进步的使用将最终在将外包制造带回美国方面发挥重要作用。拟议的项目将促进生产计划和执行中强大的优化工具的翻译。提出的贡献包括：1）一种分层方法，允许在不同的时间尺度上进行规划和调度，2）对经典运营模型进行计算改进，以纳入实时问题，如不完整的订单，生产结转和跨时期继续设置活动，3）算法设计，以适应不同的时间尺度，同时保持模型的准确性，4）分析规划期限、时段选择和目标系数的影响，以及创建系统方案以确定其满足用户需求的最佳价值，（5）发展快速再生产的物流技术，该奖项反映了NSF的法定使命，并被认为是值得通过使用基金会的知识价值和更广泛的影响审查标准。

The broader impact of this Small Business Innovation Research (SBIR) Phase I project will be in making US manufacturing base more competitive. The aerospace supply chain contains thousands of less digitally sophisticated Small and Medium Enterprise (SME) manufacturers. The SMEs constitute a vast aerospace supply base across the country and have largely remained unaffected by the advances in operations research. Consequently, local manufacturers compete unfavorably with those in low-cost countries. While the computing power and speed of optimization techniques have increased to a point where one could now solve large-scale industry problems in real time, little attention has been given to the many modeling decisions that need to be made to accurately capture the complexities of real factory physics into a prescriptive mathematical model. This project will develop of a plug-and-play software for SMEs this estimated $850 M market that will improve efficiency along the entire supply chain. The resulting solution is expected to be a production optimizer that can be implemented in less than two weeks at any SME aerospace manufacturer using their existing data streams. Use of digital technologies and operations research advances embodied in this project will ultimately play an appreciable role in bringing outsourced manufacturing back to the United States.The proposed project will advance translation of powerful optimization tools in production planning and execution. The proposed contributions include 1) a hierarchical approach that allows for planning and scheduling at different timescales, 2) computational improvements to classic operations models to incorporate real-time issues such as incomplete orders, carryover of production and continuation of setup activities across periods, 3) the design of algorithms to adapt the shop’s data to the different timescales while preserving model accuracy, 4) the analysis of the impact of planning horizon, time period choice and objective coefficients, as well as the creation of systematic schemes to determine their best value to meet user needs, and 5) the development of heuristics for quick re-optimization to respond to small deviations from the plan.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.