芯片代工厂物料派单调度与运输路径规划联合优化

智能制造是我国制造强国战略的主攻方向之一，而集成电路制造作为一个极其复杂的系统工程，其智能化对制造强国战略意义重大。物料派单调度与运输路径规划是芯片制造过程中的重要环节，高效的物料派单调度与运输路径规划在提高产能的同时，还可以降低单位生产成本，助力绿色制造。本项目将对芯片代工厂中的真实需求进行系统建模，围绕产能优化对芯片代工厂物料派单调度与运输路径规划联合优化开展研究。面向其中的多层次复杂决策，建立运输能力感知的物料派单调度问题模型与智能拥塞控制使能的运输路径规划问题模型，探索模型间的交互与反馈机制。采用数学启发式算法实现物料派单调度智能决策，使用基于松弛-修复两阶段求解技术实现运输路径格局优化，结合时空状态图与弹射链等技术进行车流疏导，并探索多层次求解框架下的子算法联动协同。本项目在理论上探索求解相关核心问题的智能优化算法，在实践上为真实智能制造系统提供算法内核，为芯片制造的智能化赋能。

Intelligent manufacturing is one of the major objectives of China's manufacturing power strategy. As the manufacture of integrated circuits is a highly complex system engineering problem, its intelligence is very important to the manufacturing power strategy. Material dispatch scheduling and transferring path planning are important steps in semi-conductor manufacturing. Effective material dispatch scheduling and transferring path planning can not only increase the productivity, but also reduce the unit production costs which is benefit to green manufacturing. This project will systematically model the real-world demands in foundries, and study the joint optimization of material dispatch scheduling and transferring path planning in order to achieve productivity optimization. Aiming at the multi-level complex decision within it, we will construct a transporting capability-aware material dispatch scheduling model and an intelligent congestion control-enabled transferring path planning model, and then explore the interaction and feedback mechanism between the two models. Furthermore, we will adopt a matheuristic algorithm for the intelligent decision of material dispatch scheduling, a two-stage relax-repair framework for the structural route optimization, and a time expansion graph and ecjection chain technique for traffic grooming, and then investigate the collaboration among the sub-algorithms. In theory, this project studies the intelligent optimization algorithms for solving the corresponding key problems. In practice, it offers an algorithm core to real smart manufacturing systems, and enables the intelligence of chip manufacturing.