Tour planning of human order pickers and/or mobile transport robots in modern warehouses

• Applicants: Professor Dr. Stefan Bock; Professor Dr. Nils Boysen
• Subject Area: Management and Marketing
• Term: since 2022
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 502273254

The significant transformation, associated with the advance of e-commerce, has not spared distribution centers and warehouses, where large numbers of orders from customers have to be assembled under time pressure. Such locations are increasingly developing into highly automated distribution factories, in which human staff work in tandem with stationary machines and mobile robots. The proposed project aims to improve e-commerce processes in modern warehouses by optimizing the route planning of picking processes. Warehouse route planning, which refers to the determining of a path of a (typically) human order picker along rows of shelves to assemble a customer order, is generally-speaking one of the classic issues in the field of Operations Research. Modern warehouses, especially for e-commerce, are increasingly being confronted with innovative technical possibilities and processes which contribute important (and - for optimization - crucial) framework conditions to classical route planning. Examples range from the route planning of human shelf stockers in so-called scattered storage warehouses, and the cooperation between human order pickers and mobile transport robots, up to the coordination of autonomous picking robots under real-time conditions. The proposed project is aimed at the development of powerful optimization algorithms for six novel route planning problems in the context of modern warehouses. The project places a particular focus on the finding of exact solution methods to issues which represent a particular challenge in the warehousing sector. Warehouses are commonly constructed to a regular standard layout, consisting of parallel picking aisles and two or more cross aisles which connect them. The resulting special structure of distance matrices allows to solve exactly route planning problems in polynomial time, problems that cannot be solved efficiently in other domains (e.g., for general road networks). Accordingly, the development, implementation and thorough evaluation of the algorithms is always preceded by an in-depth examination of structural problem properties. In this way, our algorithms should provide efficient blueprints, which can be integrated - with little adaptation effort - into decision support and planning systems of real-world operations in modern distribution centers. By analyzing the more application-driven results of the six novel tour planning problems, in a more generalizing step, the research question of identifying main complexity drivers for solving routing problems in regular warehouses is addressed. This includes the derivation of adequate data structures and solution methods for defined cases.

• DFG Programme: Research Grants