SBIR Phase I: Alpha Sigma Pi - A Method for Confident, Robust, and Optimal Process Control

这个小企业创新研究（SBIR）第一阶段项目提出了一个动态和定量的决策支持系统，以帮助制造商理解和优化具有多种工艺设置和多种质量要求的复杂工艺。本研究的智力优势在于利用广义单纯形法对可行性进行了解析求解。使用基于约束的方法，这种表示提供了过程的全局可行性和局部灵活性，给出了过程设置和质量需求之间的行为联系。遵循这些可行性图或“过程窗口”，过程工程师可以描述和理解制造过程的行为，评估单个过程条件的可行性，并发现任何和所有质量要求的潜在改进。初步结果表明，该接口是优化工艺参数和提高性能指标的有力工具。提出的研究将解决的目标包括：(1)过程行为和可行性的适应性改进；(2)工艺设置和质量要求的动态调整；(3)开发了一个易于使用的数据处理和回归系统。提出的研究将通过扩展和合理化六西格玛和其他目前在制造企业中使用的技术，对过程和质量控制产生广泛的影响。使用开发的工具，制造商可以综合和存档过程信息，深入了解其过程行为，并快速收敛到更优化的过程配置。这样一个直观的界面，考虑到多个工艺参数和质量要求的相互作用，以及变化和不确定性的影响，是目前任何其他可用的决策方法所不具备的。此外，从所提出的研究中衍生的决策支持系统可以吸引电子表格模型和其他形式的仿真的当前用户，并应用于工程设计，制造，金融，保险和军事行动。

This Small Business Innovation Research (SBIR) Phase I project proposes a dynamic and quantitative decision support system to assist manufacturers in understanding and optimizing complex processes with multiple process settings and multiple quality requirements. The intellectual merit of the proposed research stems from the analytical solution of the feasibility with the Extensive Simplex Method. Using a constraint based approach, this representation provides the global feasibility and the local flexibility of the process given the behavioral linkages between the process settings and the quality requirements. Following these feasibility maps or "process windows", the process engineer can characterize and understand the behavior of manufacturing process, evaluate the feasibility of individual process conditions, and discover potential improvements in any and all of the quality requirements. The preliminary results indicate that the proposed interface is a powerful tool in optimizing process parameters and tightening performance specifications. Objectives that the proposed research will resolve include: (1) adaptive refinement of process behavior and feasibility; (2) dynamic tuning of process settings and quality requirements; and (3) development of an easy to use system with data handling and regression. The proposed research will have a broad impact on process and quality control by extending and rationalizing Six Sigma and other techniques currently utilized in manufacturing enterprises. With the developed tool, manufacturers can synthesize and archive process information, gain insight into their process' behavior, and rapidly converge to more optimal process configurations. Such an intuitive interface, considering the interaction of multiple process parameters and quality requirements as well as the effects of variation and uncertainty, is not available in any other decision making approach available today. Furthermore, the decision support systems derived from the proposed research can appeal to current users of spreadsheet models and other forms of simulation with applications in engineering design, manufacturing, finance, insurance, and military operations.