Learning to Optimize Capacity Planning in Semiconductor Manufacturing

  • 2025-09-19 08:51:10
  • Philipp Andelfinger, Jieyi Bi, Qiuyu Zhu, Jianan Zhou, Bo Zhang, Fei Fei Zhang, Chew Wye Chan, Boon Ping Gan, Wentong Cai, Jie Zhang
  • 0

Abstract

In manufacturing, capacity planning is the process of allocating productionresources in accordance with variable demand. The current industry practice insemiconductor manufacturing typically applies heuristic rules to prioritizeactions, such as future change lists that account for incoming machine andrecipe dedications. However, while offering interpretability, heuristics cannoteasily account for the complex interactions along the process flow that cangradually lead to the formation of bottlenecks. Here, we present a neuralnetwork-based model for capacity planning on the level of individual machines,trained using deep reinforcement learning. By representing the policy using aheterogeneous graph neural network, the model directly captures the diverserelationships among machines and processing steps, allowing for proactivedecision-making. We describe several measures taken to achieve sufficientscalability to tackle the vast space of possible machine-level actions. Our evaluation results cover Intel's small-scale Minifab model andpreliminary experiments using the popular SMT2020 testbed. In the largesttested scenario, our trained policy increases throughput and decreases cycletime by about 1.8% each.

 

Quick Read (beta)

loading the full paper ...