Conference Agenda

As space for storing items is frequently scarce, it is a common approach to store items directly on top of each other within stacks, e. g., in a container terminal. Such a storing system is space-saving, but at the expense of the drawback that not every item is accessible at any time. Whenever a required item is not the topmost of its stack, all items located above it must be relocated first by means of time-consuming operations.

In this talk, we address three typical combinatorial optimization problems associated with this type of storing system: the premarshalling problem which deals with presorting operations of the stored items, the block relocation problem and the stack loading problem which deal with unloading and loading operations, respectively. A novel mixed integer linear model formulation is proposed that is applicable to all three selected optimization problems with only minor adjustments. In addition to its flexibility, the proposed model contains significantly less decision variables than existing model formulations for the addressed problems. The results of computational experiments are given.

In the era of digital commerce, the surge in online shopping and the expectation for rapid delivery have placed unprecedented demands on warehouse operations. The traditional method of order fulfilment, where human order pickers traverse large storage areas to pick items, has become a bottleneck, consuming valuable time and resources. Robotic Mobile Fulfilment Systems (RMFS) offer a solution by using robots to transport storage racks directly to human-operated picking stations, eliminating the need for pickers to travel.

We present a new type of station, called combi-station, that enables both item picking and replenishment, as opposed to traditional separate stations. To analyse the efficiency of using combi-stations, we model the RMFS as a semi-open queueing network and apply approximation methods.

Optimizing retrieval requests in warehouses is essential for maintaining a smooth flow of products. Most studies on warehouse retrieval optimization have considered no more than two input/output-points for product retrieval. In this talk, we consider different variants of a new stacker crane scheduling problem, where pallets have to be retrieved in a warehouse with multiple input/output-points. The goal is to minimize the total travel time of the stacker crane to perform all retrievals. The problem variants we consider require determining either the pallet retrieval sequence, the assignment of pallets to input/output-points, or both. We prove NP-hardness results and identify cases that can be solved in strongly polynomial time. Additionally, we propose transformations to the traveling salesman problem, enabling the application of a vast collection of existing solution techniques. Finally, in an extensive computational study, we compare different problem variants, assess their gain of optimization, and experimentally analyze the impact of various instance parameters.

To ensure quick access to stored items, there are different types of storage systems. A number of factors, including the physical size and weight of the items to be stored, the frequency of use and the available resources (such as space), determine which type of system is most suitable for a particular warehouse. In this presentation, we examine a new type of storage system that has received little scientific attention to date: Compact storage systems. Compact storage systems aim to achieve the highest possible space utilisation with limited storage space. Although such a warehouse saves space, the removal of items from compact storage systems is complex. Due to the dense arrangement, the storage units are often not directly accessible as they are covered by other units. These must first be moved or repositioned to create enough space to access the desired item. This leads to complex decision-making processes. As the repositioning of storage units consumes energy, efficient decision-making contributes directly to an energy-efficient storage system. This requires the development of algorithmic support systems that are able to calculate optimal retrieval strategies in a reasonable amount of time.

In this presentation, we provide an overview of the combinatorial optimisation problems underlying these extraction strategies. In addition, we provide insights into the computability and algorithmic approaches used to derive efficient solutions.