Series of Dept Research Seminars - “A Multi-Source Inventory System with Stochastic Lead Times: Optimal Policy and Heuristics" (Date: 28 November 2016)

Our speaker Dr. Li XIAO would like to share with us on "A Multi-Source Inventory System with Stochastic Lead Times: Optimal Policy and Heuristics" :

Multi-source inventory systems provide many advantages and have been widely adopted in practice. In addition, several new tracking technologies, such as RFID and mobile devices, provide managers with real-time supply information. In this work, we study multi-source inventory system and investigate the optimal policy and several of the heuristics, using supply system’s real-time information.

We first consider a single-product, two-source inventory system. Inventory can be replenished through a normal supply source, which consists of a two-stage tandem queue with exponential production time at each stage. We can also place an emergency order by skipping the first stage, for a fee. We develop a fairly novel approach, to obtain optimal ordering policies for the discounted or long-run average cost. The basic idea of this approach is to reduce the original three-dimensional system to an equivalent two-dimensional system, by transforming the original problem to a new problem with service rate controls for a tandem queue with two servers.

Second, we consider a single-product, multi-source inventory system with expediting options. We develop performance evaluation tools for a family of reasonable ordering and expediting policies. Performance evaluation of such state-dependent policies is generally hard. The main thrust is to show that, under these policies, the supply system becomes a network of queues with a special routing mechanism. We show that under several alterations of the general policy, the solution has a simple product form and thus we obtain closed-form performance measures.

All are welcome and no registration is in need.