PhD Seminars - Sandra Rauscher and Yifei Zhao
Friday 15 March 2013, 11:45
LT2, Management School
The two seminars for this session are: Sandra Rauscher (supervisor: Kevin Glazebrook) and Yifei Zhao (supervisor: Stein W Wallace). Franklin will be the first speaker followed by Yifei. There will be a buffet lunch at about 1pm following the seminars.
Sandra's title and abstract are:
Hybrid lateral transshipments in multi-item inventory networks
Saving costs in inventory systems can often only be accomplished by reducing service levels. Allowing lateral transshipments permits lower levels of safety stock thereby cutting costs while maintaining or even improving service levels. Usually, such movements of stock are either carried out in a reactive manner responding to a stock-out in the system, or preventively to rebalance inventory levels. Recent results show that using a combined approach can yield further improvements. Our model describes a multi-location inventory network where each location faces compound, non-homogeneous Poisson demand for several item types. We derive a quasi-myopic policy by applying a dynamic programming policy improvement step to a no-transshipment policy. Transshipment costs are modelled with a knapsack-like structure to accommodate different types of items. An extensive simulation study demonstrates the improvement of the derived policy over existing approaches and shows how this is amplified in a multi-item setup. In a second step the model is extended to a case where customers are impatient and may only accept transshipments with limited lead times. The willingness-to-wait is modelled using a parametric family of distributions given certain customer classes. We show that under this new assumption transshipments should not only be triggered by occurred shortages and that integrating an early warning signal inducing the redistribution of inventory among locations in the network is beneficial.
Yifei’s title and abstract are:
Integrated facility layout design and flow assignment problem under uncertainty
Facility layout problems (FLPs) are to assign physical facilities (such as machine tools, work centres, manufacturing cells, departments, and warehouses) to locations for a production or a delivery system according to various constraints on the flows between all available facilities. Therefore solutions of integrated facility layout design and flow assignment problems (IFLFAP) simultaneously determine the layout and flows between facilities. Stochastic IFLFAPs arise when jobs and demand levels are uncertain over production periods. A new heuristic is introduced in this research to solve two-stage stochastic IFLFAPs by reducing such problems to standard Quadratic Assignment Problems (QAPs). Numerical experiments show that major advantages of the proposed heuristic include improving quality of the final results and reducing computational efforts.