On 29 October 2021, PHD Melvin Drent has successfully defended his thesis « Stochastic Models of Critical Operations » (under the supervision of Prof. Joachim Arts) and has been granted the title of Doctor. His excellent work has also qualified his thesis for the “Best Thesis Award” 2021/2022.
We warmly congratulate him for the great work he has accomplished during his PHD and wish him an as brilliant career.
Melvin’s Thesis Abstract:
Companies can often replenish their inventories through a regular supply mode and an expedited supply mode, the latter having a shorter lead time than the former. The flexibility of the expedited mode – albeit at the expense of a premium in terms of a cost or additional resource usage – can help companies to cope with uncertainties.
The first part of this thesis studies such inventory systems with two supply modes, also called dual-sourcing inventory systems. Since it is well-known that optimal policies for dual-sourcing inventory systems have complex structures (e.g., Whittemore and Saunders, 1977; Feng et al., 2006), we mostly focus on the development and application of heuristic policies.
Chapter 2 studies a two-echelon distribution network for repairable spare parts consisting of one central warehouse and multiple local warehouses. Each warehouse keeps multiple types of repairable parts to maintain several types of capital goods, such as aircraft, rolling stock, and lithography systems. The repair shop at the central warehouse has two repair options for each failed part: a regular repair option and an expedited repair option. The latter is faster than the former but it comes with a higher repair workload. In the design of these spare parts inventory systems, companies need to decide on stocking levels and expedite thresholds such that stock investments are minimized while satisfying asset availability and repair shop workload constraints. We use queueing theory to model the dynamics of the dual-index policy in the repair shop, and we rely on Dantzig-Wolfe decomposition to develop an effective and efficient solution algorithm for the decision problem. A key insight of this chapter is that anticipating expediting decisions that will be made later can lead to substantial reductions in stock investments required to meet customer service levels. Based on a case study at Netherlands Railways, we show how managers can significantly reduce the investment in repairable spare parts when dynamic expediting policies are leveraged to prioritize repair of parts whose inventory is critically low. The contents of Chapter 2 are based on Drent and Arts (2020).
Chapter 3 studies the inbound transport and inventory management decision making for a company that sells an assortment of products sourced from outside suppliers. The inbound transport is outsourced to a third party logistics provider that offers two distinct transport modes for each product. These modes differ in terms of their carbon emissions, speed, and costs. The company needs to decide periodically how much it wants to ship with each transport mode such that total inventory costs are minimized while keeping the total carbon emissions from transportation for the entire assortment below a certain target level. Such assortment-wide constraints will be increasingly prevalent, either voluntarily or enforced by government regulation. Assuming that shipment decisions are governed by the dual-index policy of Veeraraghavan and Scheller-Wolf (2008), we formulate the decision problem as a mixed integer linear program that we solve through Dantzig-Wolfe decomposition. We benchmark our decision model against two state-of-the art approaches in a large testbed based on real-life carbon emissions data. Relative to our decision model, the first benchmark lacks the flexibility to dynamically ship products with two transport modes while the second benchmark makes transport decisions for each product individually rather than holistically for the entire assortment. Our computational experiment shows that our decision model can significantly outperform both benchmarks for realistic targets for carbon emission reduction and furthermore indicates that dynamic mode selection, as opposed to static and blanket mode selection, has great potential to efficiently curb carbon emissions from transportation at relatively little additional costs. The contents of Chapter 3 are based on Drent et al. (2021b).
The first two chapters apply well-known dual-sourcing inventory policies to multiitem settings and subsequently focus on generating valuable managerial insights. By contrast, Chapter 4 is more fundamental as it focuses on devising a new dual-sourcing inventory policy. In particular, we study the canonical single-echelon single-item inventory system with two suppliers under periodic review facing stochastic demand where excess demand is backlogged. The expedited supplier has a shorter lead time than the regular supplier but charges a higher price. We introduce the Projected Expedited Inventory Position (PEIP) policy to control this inventory system. Under this policy, the expedited supplier is operated according to an order-up-to rule that keeps the expedited inventory position at (or above) a certain target level. Regular orders placed in the past can cause the expedited inventory position to exceed this target level. The PEIP policy accounts explicitly for this overshoot by placing regular orders such that the projected expedited inventory position is kept at a target level. We show that the relative difference between the long run average cost per period of the PEIP policy and the optimal policy converges to zero when both the shortage cost and the cost premium for expedited units become large, with their ratio held constant. A corollary of this result is that several existing heuristics are also asymptotically optimal in this non-trivial regime. We show through an extensive numerical investigation that the PEIP policy outperforms the current best performing heuristic policies. The contents of Chapter 4 are based on Drent and Arts (2021)