University of Twente Student Theses

Login
As of Friday, 8 August 2025, the current Student Theses repository is no longer available for thesis uploads. A new Student Theses repository will be available starting Friday, 15 August 2025.

Optimising early design decisions for SBS/RS warehouses through a Deep Q-Learning model integrated with simulation software at Vanderlande

Eldik, J.E. van (2025) Optimising early design decisions for SBS/RS warehouses through a Deep Q-Learning model integrated with simulation software at Vanderlande.

[img] PDF
6MB
Abstract:This thesis explores developing a Deep Q-Learning model to optimize early design decisions in Shuttle-Based Storage and Retrieval systems at Vanderlande warehouses. The model automates configuration through simulation feedback on throughput and costs, aiming to determine optimal designs for ADAPTO warehouses. Trained to interact with simulation software, the DQL agent proposes configurations and adjusts policies based on feedback to approach an optimal strategy. However, it struggled to consistently yield optimal configurations or converge reliably. Insights into reward structures, exploration strategies, and DQL variants highlight the benefit of simple reward structures penalizing goal-reaching failures and rewarding cost-effective solutions. Despite efforts like Prioritized Experience Replay and Double DQN, performance did not significantly improve. Tuning hyperparameters, especially for Double DQN, may enhance future performance. In conclusion, while the model didn't generalize well across complex reward environments with shifting throughput goals, it offers foundational insights. Future work could focus on refining hyperparameters and extending architectures for better convergence and scalability.
Item Type:Essay (Master)
Faculty:BMS: Behavioural, Management and Social Sciences
Subject:50 technical science in general
Programme:Industrial Engineering and Management MSc (60029)
Link to this item:https://purl.utwente.nl/essays/106354
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page