Supply Chain Management and Business Analytics
Research activities in the fields of supply chain management (SCM) and business analytics (BA) are focused on analyzing data collected from a variety of sources and decision making informed by data analysis and supported by optimization techniques, with the aim of enabling enterprises and organizations to improve their service levels and operational efficiency, in particular in the context of supply chain and logistics management.
Research activities in those fields are mostly based in the research centre QuantOM (Quantitative methods and Operations Management), which is the association of around twenty researchers. In 2016, the research field SCM&BA was nominated as the expert field of HEC Liège in the category of international quality research, thanks to the numerous high quality scientific publications of its members.
The following is a collection of the PhD theses defended related to this field during the last five years:
Stéphanie Aerts defended her PhD thesis entitled “Robust Multivariate Dispersion Measures” in 2018. Her dissertation is devoted to the study of theoretical and empirical properties of robust multivariate dispersion measures. Multivariate dispersion measures, like covariance matrices, are involved in almost all multivariate statistical procedures but are however known to become unreliable in presence of atypical data. Robust measures, on the contrary, aim to limit the influence of such outliers while still sharing the attractive properties of classical measures. The dissertation investigates their use for the comparison of multivariate populations; for prediction via linear regression; as well as for high-dimensional automatic classification. Beyond the theoretical aspects, financial and quality control applications are discussed.
Véronique François defended her PhD thesis entitled “Neighborhood Search Algorithms for Multi-Trip Vehicle Routing” in 2018. Her dissertation is devoted to heuristic solution approaches for different variants of the multi-trip vehicle routing problem, where each delivery truck may perform more than one trip during the planning period, trips being separated by loading operations at the depot. Such problems are especially relevant in the field of city logistics. The dissertation provides insights about the behavior of several algorithmic components. Moreover, it questions algorithmic development methodologies and benchmark instances commonly used, and it proposes related methodological contributions. Véronique François is currently working as a Senior Project Manager at Bpost.
Stefano Michelini defended his PhD thesis entitled “A Comparative Study of Labeling Algorithms within the Branch-and-Price Framework for Vehicle Routing with Time Windows” in 2018. His dissertation is devoted to exact solution approaches for a vehicle routing problem with time windows, where the aim is to minimize the total working duration of the delivery trucks, instead of just the total travel time. Thus, the waiting times that are incurred by vehicles before they can service a customer are taken into account in the objective function of the problem. This aspect is relevant in applications where waiting times bear an implicit cost, such as when the vehicles consume energy during idle times, as in the case of refrigerated trucks delivering perishable goods.
Elisabeth Rodriguez-Heck defended her PhD thesis “Linear and Quadratic Reformulations of Nonlinear Optimization Problems in Binary Variables” in 2018. The thesis addresses a generic class of optimization problems which can be used to model a broad variety of situations arising in production management, logistics, finance, engineering, or computer science. It focuses on methodological aspects of the solution of such problems. More precisely, it investigates various approaches based on reformulations of the problems as linear or quadratic ones, for which efficient algorithms have been developed by previous researchers. It derives new theoretical properties of the reformulations and demonstrates the applicability of the approach through a large number of computational tests. Elisabeth Rodriguez-Heck is now working as a postdoctoral researcher at RWTH Aachen University.
Christine Tawfik defended her PhD thesis entitled “Tactical Planning on Freight Transport Networks: Service Design and Pricing” in 2018. She proposes a service network model in the context of intermodal transport by considering path-based formulations with scheduled services, minimum utilization, and resource-balancing constraints. Afterwards, the pricing decisions are simultaneously integrated by extending a bilevel programming formulation. Exact numerical tests are performed on real-world data to draw meaningful insights about the modal split and intermodal revenues’ breakdown. As the algorithmic aspects of the developed bilevel model for joint design and pricing represent new computational challenges, two heuristic approaches are also designed: a Lagrangian relaxation-based algorithm and an iterative heuristic algorithm. Christine Tawfik is now working as a post-doctoral researcher at the Zuse Institute Berlin (ZIB).
Marc Jamoulle, MD, defended his PhD thesis entitled “Multilingual Indexing in General Practice/ Family Medicine in the Era of Semantic Web” in 2017. His main contribution was the development of methods for indexing medical literature in ontologies. The ontologies were implemented (in collaboration with the CHU Rouen) in RDF/OWL, a state of the art semantic web technology. The ontologies aim at facilitating access to medical literature via advanced features, such as semantic and concept search, and entity linking (i.e., integrating information from different sources).
Martine Mostert defended her PhD thesis entitled “Design and Management of Freight Transport Networks: Intermodal Transport and Externalities” in 2017. She analyzes the viability of intermodal transport in terms of economic and environmental perspectives. The dissertation provides location-allocation models for road and intermodal freight transport and managerial insights for the different stakeholders of the transportation chain. The models are based on a bi-objective formulation which evaluates the trade-offs between an economic and an environmental optimization. The models are applied on case studies and the intervention of public authorities through a taxation policy for trucks is also studied. Martine Mostert is currently working as a Lecturer at Haute Ecole de la Province de Liège.
Célia Paquay defended her PhD thesis entitled “The Three-Dimensional Rectangular Multiple Bin Size Bin Packing Problem with Transportation Constraints, A Case Study in the Field of Air Transportation” in 2017. She analyzes a three-dimensional rectangular multiple bin size bin packing problem with specific orientation, bin shape, fragility, and weight distribution constraints that are faced by air transport companies when loading packages into the containers. The dissertation provides a complete mathematical formulation of the studied problem, three different mixed integer programming based constructive matheuristics and a tailored constructive heuristic. The utilization of the configuration tool irace and of the statistical methodology designed to compare the algorithms enhances scientific contributions of the dissertation. Célia Paquay will shortly start to work as an Assistant Professor at Maastricht University.
Tarik Bazgour defended his PhD thesis entitled “Three Essays on Market Liquidity in Equity Markets” in 2016. Over the past decades, market liquidity has been a major issue for stock market participants and even becomes one of the most important challenges since the recent financial crises. In this context, this dissertation is centered on this topic and presents three empirical essays incorporating liquidity considerations in three different subfields of Finance. The first essay demonstrates that liquidity and quality are priced characteristics in the stock market, but their pricing is conditional on market volatility conditions. The second essay is devoted to portfolio allocation and shows that aggregate market liquidity shocks have a significant influence on optimal portfolio allocations. The third essay focuses rather on mutual fund timing and examines what style liquidity timing skills may mutual fund managers possess.
Virginie Lurkin defended her PhD thesis entitled “Modeling in Air Transportation: Cargo Loading and Itinerary Choice” in 2016. The dissertation firstly analyzes a cargo loading problem, where the aim is to load a set of containers and pallets into a cargo aircraft that serves multiple airports. This work is the first to model cargo transport as a series of trips consisting of several legs at the end of which pickup and delivery operations might occur. The second problem involves the estimation of itinerary choice models that include price variables and correct for price endogeneity using a control function that uses several types of instrumental variables. The motivation is to demonstrate the importance of accounting for price endogeneity and to estimate different price sensitivities as a function of advance purchase periods. Virginie Lurkin is now working as an Assistant Professor at Eindhoven University of Technology.
Mahmood Rezaei Sadrabadi defended his PhD thesis entitled “Inventory Routing for Perishable Products” in 2016. His dissertation deals with the joint optimization of replenishment and distribution decisions in a framework where a single warehouse delivers perishable products to a network of spatially dispersed stores. The objective pursued by the optimization procedures is to balance the benefits derived from larger delivery quantities with the potential costs implied by the perishability of the products. The thesis proposes several optimization models and algorithms to deal with such problems, and validates their use through numerical experiments. Mahmood Rezaei is now working as EMEA Demand & Supply Planner at Johnson & Johnson Belgium.
Julien Hambuckers defended his PhD thesis entitled “Nonparametric and Bootstrap Techniques Ap-plied to Financial Risk Modeling” in 2015. In this thesis, we develop statistical procedures that tackle three financial risk modelling applications. In the first application, we are interested in selecting the error distribution in a multiplicative heteroscedastic model without relying on a parametric volatility assumption. In the second application, we are concerned by modeling the tail of the operational losses severity distribution, conditionally to several covariates. We develop a conditional generalized Pareto distribution model, where the shape parameter is an unspecified link function of a linear combination of covariates (single index part). In the last application, we provide a bootstrap efficient way to estimate the predictive ability of trading algorithm. Julien Hambuckers is now working as an Assistant Professor at HEC Liège.
Geraldine Laurent defended her PhD thesis entitled “From Censored to Cross-Sectional Data: Non and Semiparametric New Developments” in 2014. This thesis deals with semi-parametric analysis of duration data and especially censored and cross-sectional data. First, we present a new estimation procedure for the parameters of the parametric conditional variance in a general heteroscedastic regression model. This procedure constructs artificial data to replace censored data by exploiting the structure of the model. Next, we estimate the conditional mean and variance in nonparametric regression for cross-sectional data where the conditional truncation distribution function is supposed to be known. Finally, the semi-parametric estimation of the time of interest distribution is studied; the truncation distribution is then assumed to belong to a parametric family.
Thierry Pironet defended his PhD thesis entitled “Multi-period Stochastic Optimization Problems in Transportation Management” in 2014. His dissertation is devoted to propose a methodological framework and algorithms to tackle dynamic optimization problems based on confirmed and projected information. As these multi-period stochastic problems are hard to solve to optimality, a scenario based approach was developed to compare performances of several policies and to estimate several values of the information pertaining to the horizon. Several bounds and theoretical concepts have been defined to support the experimental framework. The proposed methodology was tested on an industrial application of vehicle loading and on a more generic framework of vehicle dispatching. Thierry Pironet is now working as an Assistant Professor at HEC Liège.