A Novel Framework for Automated Warehouse Layout Generation

Atefeh Shahroudnejad,Payam Mousavi,Oleksii Perepelytsia,Sahir,David Staszak,Matthew E. Taylor,Brent Bawel

2024-07-13

Abstract:Optimizing warehouse layouts is crucial due to its significant impact on efficiency and productivity. We present an AI-driven framework for automated warehouse layout generation. This framework employs constrained beam search to derive optimal layouts within given spatial parameters, adhering to all functional requirements. The feasibility of the generated layouts is verified based on criteria such as item accessibility, required minimum clearances, and aisle connectivity. A scoring function is then used to evaluate the feasible layouts considering the number of storage locations, access points, and accessibility costs. We demonstrate our method's ability to produce feasible, optimal layouts for a variety of warehouse dimensions and shapes, diverse door placements, and interconnections. This approach, currently being prepared for deployment, will enable human designers to rapidly explore and confirm options, facilitating the selection of the most appropriate layout for their use-case.

Artificial Intelligence

What problem does this paper attempt to address?

The paper proposes an AI-based automated warehouse layout generation framework aimed at addressing the inefficiencies and error-prone nature of traditional manual warehouse layout design. Specifically, the paper attempts to solve the following key issues:

1. **Optimizing Warehouse Layout**: By optimizing the warehouse layout to improve operational efficiency, reduce labor costs, and ultimately enhance customer satisfaction.

2. **Automated Layout Generation**: Developing a method that can automatically generate multiple candidate optimal layouts to meet the needs and preferences of different users, while considering factors such as storage capacity and the number of access points.

3. **Multi-Objective Optimization**: Considering the need to balance the maximization of storage space and the convenience of access in practical applications, the paper proposes a multi-objective optimization method that can find an appropriate balance between these competing goals.

4. **Constraint Satisfaction**: Ensuring that the generated layouts meet various constraints in real-world scenarios, such as item accessibility, minimum aisle width requirements, and aisle connectivity.

5. **Evaluation and Selection**: To assist designers in making decisions from multiple feasible layout options, the paper also provides a scoring mechanism to evaluate the pros and cons of different layouts, thereby selecting the most suitable solution for specific needs.

In summary, this research aims to develop an efficient, flexible, and practical automated tool to support warehouse designers in quickly exploring and confirming the best warehouse layout options.