Maximizing Profit via Assortment and Shelf‐Space Optimization for Two‐Dimensional Shelves

Product proliferation and changes in demand require that retailers regularly determine how items should be allocated to retail shelves. The existing shelf-space literature mainly deals with regular retail shelves onto which customers only have a frontal perspective. This study provides a modeling and solution approach for two-dimensional shelves, e.g., for meat, bread, fish, cheese, or clothes. These are categories that are kept on tilted shelves. Customers have a total perspective on these shelves and can observe units of one particular item horizontally and vertically instead of just seeing the foremost unit of an item, as is the case of regular shelves. We develop a decision model that optimizes the two-dimensional shelf-space assignment of items to a restricted, tilted shelf. We contribute to current literature by integrating the assortment decision and accounting for stochastic demand, space elasticity and substitution effects in the setting of such self types. To solve the model, we implement a specialized heuristic that is based on a genetic algorithm (GA). By comparing it to an exact approach and other benchmarks, we prove its efficiency and demonstrate that results are near-optimal with an average solution quality of above 99% in terms of profit. Based on a numerical study with data from one of Germany’s largest retailers, we were able to show within the scope of a case study that our approach can lead to an increase in profits of up to 15%. We demonstrate with further simulated data that integration of stochastic demand, substitution, and space elasticity results in up to 80% higher profits.