Integrated inventory pooling for firms servicing both on-line and store demand

The use of local store inventory for both in-store and on-line fulfillment can be preferable to the central warehousing of on-line stock in certain scenarios. Part of this appeal comes from the availability of various decentralized neighboring stores as sources for the on-line fulfillment of any given order. In this study, a range of substitute-inventory availability scenarios for these systems are simulated, under the assumption that transshipments to support in-store demand are not possible and the fulfillment of in-store demand is always a priority relative to on-line orders at a given site. Meta-models are used to summarize the dependency of performance measures on experimental factors and ultimately form the backbone for service constrained non-linear optimization. The findings show that there is a limit to which reductions in coordination costs for on-line source substitution and associated inventory availability decisions can provide benefit. This limit is influenced by both the proportion of the market managed through on-line sales and the service level constraints in place for both on-line and in-store channels.