PERFORMANCE ASSESSMENT OF DISTRIBUTED INVENTORY IN PHYSICAL INTERNET

Classical supply chain design relies on a hierarchical organization to store and distribute products over a given geographical area. Within this framework a shortage in a stock affects the whole downstream of the supply chain regardless of the inventory kept in others locations. Within the Physical Internet approach, inventories are distributed in hubs towards the market and source substitution is allowed. The Physical Internet aims to integrate logistics networks into a universal system of interconnected services through the development of protocols and standards for the routing of smart containers of various sizes. This organization enables a distributed storage of goods in hubs thanks to containerization, thus the feasibility of multi-sourcing to one ordering point. This contribution measures the impact of such an organization on stock levels and inventory costs with service level set as a constraint. The analysis focuses on the resources levels (transportation and inventory) needed by the current supply model and by the Physical Internet in order to serve a market with a (Q,R) stock policy. Starting with two supply models and with the definition of cost models as well as inventory policy, the work is based on computer simulation. The analysis tested 3 different families of criterion in order to select dynamically the source when an order is requested: Source Substitution, Minimum Ratio and Minimum Sum. The source substitution, one of the simplest, was found the more efficient and stable according to different scenarios.