Consolidation effects: Whether and how inventories should be pooled

Center for Logistics Studies, The COPPEAD Graduate Business School, Federal University of Rio de Janeiro, Rua Paschoal Lemme, 355, Cidade Universitária, Rio de Janeiro, CEP. 21.949-900, Brazil
Transportation Research Part E: Logistics and Transportation Review 01/2009; DOI: 10.1016/j.tre.2009.01.006

ABSTRACT This paper presents a framework for deciding whether and how inventories should be pooled, using the consolidation effect as a cornerstone tool to measure inventory costs, service levels, and total costs. Based on the random generation of different scenarios, it is indicated the adequacy of inventory centralization, regular transshipments, and independent systems to a given set of demand, lead time, and holding costs characteristics. Sensitivity analyses on mathematical expressions are performed to determine when one alternative is preferable in terms of total costs. Real settings are also presented in light of the framework developed.

  • [Show abstract] [Hide abstract]
    ABSTRACT: Distribution network design is about recommending long-term network structures in an environment where logistic variables like transportation costs or retailer order sizes dynamically change over time. The challenge for management is to recommend an optimal network configuration that will allow for longer term optimal results despite of environmental turbulences. This paper studies the robustness of cost-optimized FMCG (fast-moving consumer goods) distribution networks. It aims at observing the impact of changing variables/conditions on optimized logistic structures in terms of the optimal number and geographical locations of existing distribution centers. Five variables have been identified as relevant to the network structure. A case study approach is applied to study the robustness of an existing, typical, and optimized FMCG network. First, distribution network data of a German manufacturer of FMCG are recorded and analyzed. A quantitative model is set up to reflect the actual cost structure. Second, a cost optimal network configuration is determined as a benchmark for further analysis. Third, the variables investigated are altered to represent changes, both isolated (ceteris paribus) and in combination (scenario analysis). Each one of the variables investigated proves to be fundamentally able to suggest a change of the optimal network structure. However, the scenario analysis indicates that the expected changes will by and large compensate each other, leaving the network in near optimal condition over an extended period of time.
    Logistics Research