AMR-Based Automated Material Handling in Semiconductor Fabs: A Discrete Event Simulation Study

Given their capital intensity, advanced semiconductor fabs are accelerating automation to improve control efficiency and flexibility in complex production systems. Within this context, intralogistics is pivotal for enhancing high throughput and minimizing lead time. An industrial case study is presented on the design and planned implementation of an Automated Material Handling System (AMHS) based on Autonomous Mobile Robots (AMR) in a production fab. To support this transition, a data-driven sizing of transport demand was first performed through the integration of heterogeneous data sources, and the layout of pick-up/drop-off stations was defined; these elements are then fed directly into a discrete-event simulation (DES) model of the AMHS, which captures wafer-transport dynamics. The proposed methodology leverages a highly reconfigurable and fast-executing DES model, coupled with a Genetic Algorithm (GA), to jointly optimize operational parameters, configuration variables, and routing/dispatching policies with the objective of cost minimization. The GA–DES loop identifies the best combination of AMR fleet setup and control rules by maximizing transport completion rate. The optimization explicitly accounts for battery and buffer capacity constraints, fab traffic, and the stochastic nature of AMR-operator interactions. This results in a simulation-based decision-support framework for smart logistics management and reconfiguration in semiconductor fabs, enabling more effective AMR deployment and higher operational performance.