Product Line Design and Scheduling at Intel
Product Line Design and Scheduling at Intel
复制标题
英特尔产品线设计与调度
DOI：
10.1287/inte.1120.0641
复制
发表时间：
2012
期刊：
影响因子：
--
通讯作者：
K. Kempf
中科院分区：
文献类型：
--
作者：
Evan Rash;K. Kempf
我们为产品线设计和调度问题开发了一个整体且连贯的模型。我们的模型结合了市场需求和财务、设计工程能力、制造成本和多次动态。该解决方案集成了最优集覆盖、资源受限作业调度、动态规划和投资组合优化等技术和概念，以实现总体利润最大化。关键概念是将问题分解为两层。外部遗传算法层处理资源约束、调度和财务优化。内部数学编程层像经典的集合覆盖一样优化产品组合。由此产生的算法有效地解决了比以前更大尺寸和更高复杂性的问题。在最初部署 21 个月后，代表大多数主要英特尔集团和许多不同工作职能的 250 多名人员正在使用围绕该算法的决策支持套件 (DSS)。该 DSS 用于集成一组以前独立的非通信业务流程。
We develop a holistic and coherent model for the product line design and scheduling problem. Our model incorporates market requirements and financials, design-engineering capabilities, manufacturing costs, and multiple-time dynamics. The solution integrates techniques and concepts from optimal set-covering, resource-constrained job scheduling, dynamic programming, and portfolio optimization to maximize overall profit. The key concept is the decomposition of the problem into two layers. The outer genetic algorithm layer handles resource constraints, scheduling, and financial optimization. The inner mathematical programming layer optimizes product composition as classical set covering. The resulting algorithm efficiently solves problems of larger size and higher complexity than previously possible. Over 250 personnel representing most major Intel groups and many distinct job functions are using the decision support suite (DSS) surrounding the algorithm 21 months after its initial deployment. This DSS serves to integrate a set of previously separate noncommunicating business processes.