About the speakers
Title: "Demand and fleet management for on-demand delivery services”
Abstract: On-demand delivery systems operate in a very uncertain and dynamic environment. Customers request service spontaneously at any time and place in the city and expect a high service availability as well as fast, cost-efficient, and high-quality service. Service providers control a fleet of delivery vehicles in real time in reaction to the new orders and in anticipation of future demand. Further, they can control demand in terms of service availability, prices, and the offered service itself to stay flexible for future demand. In this talk, we discuss selected means of fleet and demand management for on-demand services and their impact on service providers, customers, and drivers. We will set a special focus on instant delivery, restaurant meal delivery, and crowdsourced delivery concepts.
Bio: Marlin Ulmer is a full professor for Management Science at the Otto-von-Guericke-Universität Magdeburg. He studied Mathematics at the University of Göttingen and received his PhD in Business Administration from the Technische Universität Braunschweig. Before moving to Magdeburg, Marlin was a Junior-Professor at TU Braunschweig and a DFG Emmy Noether Fellow at TU München. He also spends one semester each at the University of Iowa and at Georgia Tech. Marlin’s research focuses on data-driven decision support for urban transportation and mobility. To this end, he combines tools from Operations Research and Machine Learning.
Title: The steering problem: a meal delivery problem with advanced-scheduled drivers
Abstract: One of the main challenges of on-demand meal delivery platforms is to provide an on-time fulfillment service, mainly when the delivery resources are predetermined for the service period. In this case, the real-time capacity resizing strategies against spatial and temporal demand variations throughout the day are impossible. Hence, this study proposes a dynamic supply and demand guidance to the commonly used driver-order assignment mechanisms to manage meal delivery operations. We introduce steering actions, in which drivers are coordinated concerning future order arrivals, or upcoming customers' delivery promises are adjusted before they place their orders. The extensive numerical experiments reveal that the supply and demand steering to the dispatching decisions decreases the customers' quality loss via fever late deliveries.
Bio: Alp Arslan is a Postdoctoral Researcher in the Faculty of Technology, Policy, and Management. He primarily works for the CUSTOMIZE project at Transportation and Logistic Section. He received his Ph.D. degree in Operations Management from the Rotterdam School of Management, Erasmus University
Title: Building innovative data-driven logistics solutions at Just Eat Takeaway.com
Abstract: Just Eat Takeaway.com is a multinational online food ordering and delivery company. At JET, we aim to further boost the operational performance by leveraging the data-driven logistics solutions. One of the examples is the collaboration project with TU Delft on demand and supply steering.
Bio: Yihong Wang is a senior data scientist at Just Eat Takeaway.com. Before joining JET, he completed his PhD in Transport & Planning at Delft University of Technology.
Title: 'Look' ahead to plan ahead: short-term forecasting in on-demand delivery service
Abstract: For the business of online meal delivery, good courier management is the key to improve overall delivery efficiency and customer satisfaction. In order to optimally allocate couriers for the upcoming deliveries, the forecasted demands of different neighborhoods are useful indicators. Combining classic time series forecasting and machine learning techniques, we can predict the near-future demands for each service zone based on current information, such as weather, seasonality, and recent number of orders.
Bio: Jingyi Cheng obtained her BSc degree in Econometrics and Operation Research at Erasmus University Rotterdam in 2021. Her research interests are related to the application of AI, machine learning and network science to the field of optimization and operation research. She is a current graduate student in Computational Science at the University of Amsterdam.
Dolores Romero Morales
Title: Mathematical Optimization for Explainable-by-Design Machine Learning
Abstract: There is a pressing need to make Data Science tools more transparent. Despite excellent accuracy, state-of-the-art Data Science models effectively work as black boxes, which hinders model validation and may hide unfair outcomes for risk groups. Transparency is of particular importance for high stakes decisions, is required by regulators for models aiding, for instance, credit scoring, and since 2018 the EU has extended this requirement by imposing the so-called right-to-explanation in algorithmic decision-making. From the Mathematical Optimization perspective, this means that we need to strike a balance between several objectives, namely accuracy, transparency, and fairness. In this presentation, we will navigate through some novel techniques that embed explainability and fairness in the construction of Data Science models. This includes the ability to provide global, local and counterfactual explanations, as well as model cost-sensitivity and fairness requirements. We will show the versatility of our methodology when applied to more complex data types such as functional data.
Bio: Dolores Romero Morales is a Professor in Operations Research at Copenhagen Business School. Her areas of expertise include Data Science, Supply Chain Optimization and Revenue Management. In Data Science she investigates interpretability and visualization matters. In Supply Chain Optimization she works on environmental issues and robustness. In Revenue Management she works on large-scale network models. Her work has appeared in a variety of leading scholarly journals, including European Journal of Operational Research, Management Science, Mathematical Programming and Operations Research, and has received various distinctions. Currently, she is Editor-in-Chief to TOP, the Operations Research journal of the Spanish Society of Statistics and Operations Research, and an Associate Editor of Omega and the INFORMS Journal on Data Science.
Title : Machine Learning-Based Capacity Checks for Dynamic Time Slot Management
Abstract: Online grocers typically let customers choose a delivery time slot to receive their goods. To ensure a reliable service, the retailer may want to close time slots as capacity fills up. The number of customers that can be served per slot largely depends on the specific order sizes and delivery locations. Conceptually, checking whether it is possible to serve a certain customer in a certain time slot given a set of already accepted customers involves solving a vehicle routing problem with time windows. This is challenging in practice as there is little time available and not all relevant information is known in advance. We explore the use of machine learning to support time slot decisions in this context.
Bio: Niels Agatz is an Associate Professor of Last-mile Supply Chain Operations at the Rotterdam School of Management, Erasmus University, where he also serves as the Academic Director of the MSc program in Supply Chain Management. Dr. Agatz’ research focuses on developing quantitative models and optimization methods to support sustainable last-mile logistics systems. He is particularly interested in technology-driven innovations in urban delivery, shared mobility, and omni-channel retail operations. Niels currently leads a work package on last-mile delivery operations in the City Logistics Lab project – a collaboration between various cities and company stakeholders.
Title: Great customer experiences don’t happen by accident
Abstract: 8.1 million letters and 1.1 million parcels: the average numbers of daily shipments by PostNL in the Benelux. E-commerce is here to stay, and PostNL aspires to (stay to) be the favorite delivery service. To realize this ambition, the company transforms into a customer driven organization. As a part of this transition, PostNL introduced the Customer Journey Factory, where multidisciplinary teams (radically) redesign and optimize customer and consumer journeys. Merel Koppen, Program Manager within the Journey Factory, will be talking about the challenges and opportunities of this transition within the 200-year-old company.
Bio: Merel Koppen is an experienced project and program manager. In her current role as Program Manager Customer Excellence, she oversees the customer centricity transition within PostNL. With a background in Strategy & Innovation Management, she has been able to put theory into practice within companies such as ABN AMRO and Staples. She is drive by a need for continuous improvement, focus on excellent results and the belief that there is a solution for every problem.
Title: A Bayesian approach to identify the delivery preferences of medical recipients.
Abstract: Over the past years, Logistic Service Providers have seen an increase in demand regarding the home delivery of medical products. In order to offer medical recipients the quality delivery services they desire, Logistic Service Providers are constantly seeking to gain a better understanding of its recipients and tailor the services to their needs. In this paper, we quantify the delivery preferences of medical recipients using a Hierarchical Bayesian modelling procedure leveraging data provided by PostNL. We find that the choice of delivery can to a great extent be explained and predicted by introducing personalized variables such as loyalty and previous choice. In addition, the results show a great amount of heterogeneity between and within the delivery preferences of medical recipients, which confirms that assuming homogeneous delivery preferences across and among recipients is not reasonable.
Bio: Bram Mosterd is a Data Scientist at MetrixLab. He finished my Master in Econometrics at Erasmus University Rotterdam in Quantitative Marketing and Business Analytics specialization.
Title: Robust supply chain: from historical data to decision making under uncertainty
Abstract: We are living in an exciting period of developments in data-driven decision-making. On one hand, there is the availability of all sorts of data. On the other hand, we see that having more data results in seeing more uncertainty in many processes. In this talk, we focus on two examples of supply chains and how the use of data can help decision-makers. Because of all the uncertainty, we are interested in making robust strategic decisions. In the first example, we construct a hypothetical situation and investigate how robust optimization can help design the supply chain network. In the other example, we focus on a practical case and check the practicality of robust solutions, the advantageous and also disadvantageous.
Bio: Ahmadreza Marandi is an assistant professor at Industrial Engineering and Innovation Sciences of the Eindhoven University of Engineering. His research focuses on decision-making under uncertainty in industrial problems. More specifically, he is doing research on the usability of historical data in strategic decision-making using analytical tools. With a background in Math, Ahmadreza is focusing on developing algorithms that can help practitioners better make decisions while using historical data.