Operationalizing entropy in logistics to mitigate congestion: application to a semiconductor wafer fabrication facility layout
针对半导体晶圆厂布局设计，提出基于熵的“流量平衡”方法替代传统最小化加权距离，通过混合整数线性规划模型和离散事件仿真验证，可缓解运输拥堵并提升交付量达11%。
In a facility layout design problem for semiconductor fabrication facilities, the traditional approach has been to minimise the flow-weighted distance of material flow. Transports are conducted by an Automated Material Handling System on its internal transportation network. However, this approach often leads to significant traffic congestion, raising the question of whether minimising flow-weighted distance is truly an effective criterion for designing facility layouts. This study proposes a novel approach to layout design for semiconductor fabrication facilities, termed ‘flow-balancing’, to mitigate traffic congestion within the transportation system by leveraging the concept of entropy to quantify congestion in the transportation network. We present Mixed Integer Linear Programming models for designing the facility layout of semiconductor fabrication facilities under both the traditional and proposed approaches, and compare their performances using discrete event simulation. The result shows that the flow-balancing approach tends to outperform the traditional method in key performance measures of transportation system in semiconductor manufacturing facilities, with improvements of up to 11% in the total number of deliveries.