Roche-Genentech is a global biotech company that operates three distribution centers in the US. Inventory levels for each SKU are set on a country level based on historic demand and epidemiological forecasts. However, there was no formal guidance for determining the product mix and inventory levels for regions within a country. This study introduces two models to support inventory placement decisions. A Multi-Criteria Optimization (MCO) model is developed to advise on which customers to serve from each stocking point, and which SKUs to stock at each location by considering the cost and emissions associated with serving each customer from each distribution center (DC). Additionally, we develop an inbound allocation model that recommends SKU replenishment quantities to maintain target inventory levels at each site given the quantity of inbound imports. The initial run of our models finds a 11% reduction in US network-wide transportation cost and emissions over the current state customer-DC and product mix allocations. We recommend a revision of company policies that constrain service of the highest value customers and highest revenue segment products exclusively from a single distribution node.
