BOF Doctoral Grant Ruben D'Haen
In business-to-consumer (B2C) e-commerce sales predefined deadlines are typically used to separate order picking activities in the warehouse (i.e., retrieving goods from their storage locations) from delivery operations to customer locations, allowing both activities to be scheduled independently. Integrating these operational decisions can be seen as one of the key opportunities for improvement. However, it is unclear how both problems can be integrated, and what the benefits are, in complex and realistic settings. Therefore, the main objective of this project is to develop adequate models and algorithms to support integrated decision making in a realistic setting, and to analyse the benefit of integrated decision making under different problem characteristics.
Cluster SBO project DISpATch: Digital twin for synchromodal transport
The DISpATch project will focus on organizational and technical enablers for seamless synchromodal transport services in Flanders. Given the real-time dynamics and flexible nature of synchromodal transport, different actors and transport modalities need to work together and adapt according to unexpected events as well as contextual information that affect transport processes. These events and contextual information can be positive or negative perturbations that shape freight movement and transport mode selection, such as newly incoming orders, transport delays, cancellations, collaborative bundling opportunities, accidents, water levels, strikes and many more. The project will develop a platform represented by a Digital Twin component in order to provide a testbed for synchromodal opportunities within a risk-free environment. Such a risk free environment allows for analysis and evaluation of triggering events (e.g., new orders, disruptions, delays) which induce physical movements. It will measure the real-time synchromodal complexity and evaluate various decisions and offer alternatives by making use of mathematical, simulation and machine learning models.
FWO Postdoctoral fellowship Yves Molenbruch
In a traditional mobility policy, public transport is supplemented with (private) dial-a-ride services, providing demand-dependent door-to-door transport to people with reduced mobility. For efficiency reasons, many governments are currently implementing an innovative demand-driven mobility policy in which private dial-a-ride services also replace unprofitable public transport in rural areas. This project focuses on developing a dynamic matheuristic routing algorithm which provides integrated solutions, combining public transport and dial-a-ride services by ensuring synchronization between routes and modes.
FWO Strategic Basic Research ORDinL
The ORDinL project seeks to develop innovative methodologies for data-driven optimisation in logistics. Such an approach would enable the use of available data to learn and find patterns, thereby continuously and automatically adapting and improving logistics optimisation processes. The project is carried out in cooperation with partners from KU Leuven and VUB.
FWO PhD fellowship Lien Vanbrabant
Emergency departments constitute an important chain in a health care system. Due to a remarkable growth in demand and the ever tightening budgets, the need for services often exceeds the available resources. In this project, the aim is to analyse, optimise and manage emergency departments in order to reduce emergency department crowding and to make emergency departments work more efficiently by use of simulation and simulation-optimisation techniques.
FWO PhD fellowship Sebastian Rojas Gonzalez
The use of numerical models to simulate and analyse complex real world systems is now commonplace in many scientific and engineering domains. Depending on the system under study, and the assumptions of the modeller, the models can be deterministic (e.g., in the case of analytical functions) or stochastic (e.g., when Monte Carlo simulation or discrete-event simulation is used). Often, the goal of the modeller is to find the values of controllable parameters (i.e., decision variables) that optimize the performance measure(s) of interest. As the evaluation of the primary numerical model can be computationally expensive, different approaches have been developed to provide less expensive metamodels, also referred to as surrogate models. The goal of this research is to develop effective and efficient algorithms for multi-objective simulation optimization, using such metamodels, and to compare the performance of different algorithms using appropriate metrics. The challenge lies in the inherent randomness of the observed outputs, which complicates the search for the Pareto front, as well as the efficient identification of this front. Additionally, the simulation budget is typically limited, so a major question is how to allocate this budget optimally between the exploration and exploitation stages of the algorithms.