INVENTORY CONSIDERATIONS IN NETWORK DESIGN
Optimization is accepted by logisticians as the appropriate approach to network design. While network design is used to make decisions regarding many aspects of logistics networks, one set of important decisions relates to the number, location, and size of warehouses in the network. Managers must also determine the customers to be supplied from each warehouse. Because of the costs involved in operating a network, managers often achieve significant cost reductions through network design (Jimenez, Brown, and Jordan 1998).
The logistics network decision is usually modeled as a trade-off between transportation and fixed warehousing costs. When there are few warehouses in the network, fixed warehousing costs are low, but transportation costs are high. As you add warehouses to the network, the fixed costs increase, but the transportation costs are reduced due to two factors. First, adding warehouses generally decreases the number of miles traveled because the total distance a unit is shipped, from the supplier to the warehouse and then to the customer, gets closer to a straight-line. In addition, the most expensive portion of that journey is often from the warehouse to the customer as those are usually less-than-truckload shipments. The average distance of this leg of the journey is reduced when there are more warehouses in the network.
What is missing in this standard trade-off is the inclusion of inventory cost. It is well known from the square root law and portfolio effect theory that inventory levels increase as the number of warehouses in the system increases (Maister 1976;Zinn, Levy, and Bowersox 1989). Accordingly, the main purpose of this research is to show that including inventory often significantly affects the optimal design of the network and that consequently managers should account for inventory when designing logistics networks.