Retailers whose sales volumes have changed noticeably over time or who want to expand their business to new geographical areas are faced with the question of how to improve or suitably restructure their current logistics network. These include new developments in warehouse automation or design measures to improve supply chain resilience. We examine a distribution network design problem that integrates the strategic decision of determining the number of distribution centers (DC) and the number of DC types anticipating the impact of these decisions on subsequent tactical planning. We propose a MIP modeling approach that factors in costs that result from setting up and operating the warehouses, inbound and outbound transportation, inventory holding at the DCs and instore logistics. The resulting distribution network potentially consists of several types of warehouses that store a distinct set of products each, and corresponds to structures commonly applied in grocery retailing, for example. To solve the problem we present an iterated local search that is based on a decomposition approach to facilitate practical applicability of the model. The application of our solution approach to a real-life case of a major European grocery chain highlights the improvement potential that is achievable by restructuring existing retail networks. We found that up to 15% of decision-relevant supply chain costs and even 30% to 50% of the decision-relevant instore costs can be reduced compared to the original network structure. Further sensitivity analyses provide additional insights into the problem structure, cost tradeoffs and robustness of the solutions generated.
展开▼
