PhD

Consumer Healthcare Supply Chain Visibility and Multi-Echelon Inventory Optimisation

My PhD project is in partnership with Haleon and explores multi-echelon inventory optimisation (MEIO) in complex supply chains in the consumer healthcare industry. While there is plenty of literature on MEIO, there are certain features and requirements in healthcare supply chains which have been sparsely studied. The aim of this project is to model and solve large-scale MEIO problems with unique features and requirements that are relevant to pharmaceutical and healthcare supply chains. We use Markov Decision Processes (MDPs) as a framework to model such problems. However, due to the curse of dimensionality, exact solution methods for MDPs are infeasible for solving large-scale problems that are relevant to this project. To mitigate this, we aim to design and develop scalable methods by leveraging techniques from approximate dynamic programming and reinforcement learning.

Supervisors: Dr. Anna-Lena Sachs (Lancaster University), Dr. Robert Shone (Lancaster University) and Dr. Gueorgui Mihaylov (Haleon UK plc)

MRes

I completed the Master of Research in Statistics and Operational Research programme within the STOR-i CDT at Lancaster University in 2024-25. In this programme, I was exposed to a range of research topics, programming assignments, group assignments, problem solving days, taught modules, and presentations, which equipped me with appropriate skills and the confidence to begin a PhD programme!

Dissertation

The 3-month dissertation in this programme was a PhD Research Proposal on my PhD project above. In this dissertation, we establish the motivation for multi-echelon inventory optimisation in the consumer healthcare industry and discuss some basic terminology relevant to inventory optimisation. We also discuss the methodology of Markov Decision Processes (MDPs) which form the framework of our inventory optimisation models. Using MDP frameworks, we model single- and multi-echelon inventory problems where we wish to find when and how much sites in a supply chain should order to ensure that products reach customers on time and efficiently. We explore the differences between policies and costs under centralised (headquartered) and decentralised (local) decision making, as well as the impacts of logistical constraints on such policies. Finally, we identify research gaps and further research directions which could be explored in the course of a PhD project.

Supervisors: Dr. Anna-Lena Sachs, Dr. Robert Shone, Dr. Gueorgui Mihaylov and Leonidas Tsaprounis

Research Project on the Maximum Flow Problem

Network flow problems, such as the maximum flow problem (MFP), are studied extensively due to their variety of applications in several domains. The MFP, which can be formulated as a linear program (LP), aims to maximise the flow of a commodity from a source to a sink in a network while adhering to capacity constraints along each arc in the network. In this project, we discuss tailored solution methods for the maximum flow problem, from the pioneer Ford-Fulkerson algorithm to recent Interior Point Methods as opposed to general LP solution approaches such as the simplex method. We discuss three distinct applications of the maximum flow problem in healthcare management, homeland security and river network evaluation. We also compare the general LP solution approaches with a maximum flow solution method in a case study. The aim of this study is to maximise flow in the emergency evacuation of people from Lancaster University to the Royal Lancaster Infirmary in the event. It is found that maximum flow solution approaches outperform general LP methods, thus emphasising the effectiveness of the specialised maximum flow algorithms.

Supervisor: Dr. Adam Letchford