This study develops a synchromodal transport model with real-time replanning using a Double-Dueling Deep Q-Network (D3QN) agent. The hinterland port-hinterland leg dominates supply-chain costs and remains vulnerable to disruptions. We simulate realistic disruptions using varying occurrence and severity levels and train a D3QN agent to decide whether an affected shipment should wait or reassign to another mode. Model performance is evaluated in a Rhine-Alpine corridor case study against two baselines strategies (Always Wait, Always Reassign) and a tabular Q-learning benchmark. The results show that the D3QN policy achieves the lowest costs on a combined set of disruptions, outperforming alternatives overall. In particular, D3QN excels under high-occurrence, low-severity disruptions but appears less effective under rare, high-severity events. The integration of the D3QN with parallel simulation greatly sped up training (roughly 4 times faster convergence) without degrading policy quality. It is discussed how classifying disruptions by occurrence and severity helps tractability, and show that the D3QN learns to reassign shipments more effectively. The findings indicate that deep reinforcement learning can improve synchromodal resilience, particularly in large-scale settings where conventional RL or heuristics fail.

This study proposes an integrated mode choice and assignment model that combines the final steps in the classical four-step transport modelling approach. Using multi-objective optimisation, the model minimises service costs while simultaneously maximising the probability of mode selection. Compared to traditional sequential models, the integrated model prevents over-assignment to modes exceeding capacity. The model was then used to asses three key transport mode attributes—cost, capacity, and frequency—to evaluate their effect on the modal split, with a focus on increasing Inland Waterway Transport (IWT). Results showed that increasing road costs and IWT service frequency (with associated capacity growth) had the most impact on shifting modal share towards IWT. These findings demonstrate the potential of this integrated model to better evaluate policy impacts and inform transport planning.

Shared Automated Vehicles (SAVs) hold significant potential to redefine urban mobility services. However, the applicability and optimization of their operational strategies in diverse urban contexts remain unclear, particularly concerning the complex interactions with existing public transport systems. This study systematically investigates how urban heterogeneity, including city scale, network topology, and demand patterns, modulates SAV fleet management strategies to balance operator costs and user utility within a multi-modal transportation context.

To this end, this study develops a comprehensive simulation framework that combines an advanced ride-pooling candidate generation algorithm (ExMAS) with an innovative two-stage vehicle assignment optimization process, and integrates a nested Logit model to quantify the competition structure with public transport. This framework is applied to 37 Dutch cities of varying scales to derive generalizable findings.

The study finds that population scale is a fundamental determinant of the required fleet size, exhibiting a strong linear relationship, especially in large cities. The key to enhancing operational efficiency, however, lies in more nuanced urban structure metrics. In large cities, longer commuting distances combined with more complex networks (higher node degrees) foster ride-pooling potential. In contrast, for small cities, the local compactness of the network (higher clustering coefficient) and demand density are critical for improving vehicle turnover efficiency. Furthermore, the research confirms that a uniform pricing strategies are unlikely to achieve best performance on pooling efficiency, highlighting the necessity of implementing differentiated pricing based on user preferences and city scale. In competition with public transport, the study's Public Transport Competitiveness Index reveals a non-monotonic relationship with travel distance. A state of competitive balance is observed for short-distance trips, while PT holds a distinct advantage in the medium-distance range. For long-distance trips, the inherent speed advantage of SAVs allows them to become the more competitive option. This external competition also structurally increases the internal pooling rate of the SAV system.

The findings of this study provide critical strategic insights for SAV operators and policymakers, emphasizing the critical importance of adjusting fleet management, service design, and pricing strategies according to specific urban characteristics and the competitive environment, thereby providing decision support for achieving an efficient and sustainable urban transportation system.

In vendor-to-retailer systems where retailers with limited inventory capacity place orders and vendors dispatch trucks within soft time windows, trucks may arrive before sufficient space is available, requiring them to wait on site resulting in inventory-driven delays. Since no existing formulation incorporates inventory-dependent service times, this study models the dispatch problem as a discrete-time Markov Decision Process (MDP) with inventory-dependent service times, solved using Approximate Dynamic Programming (ADP) with a Partial Value Function Approximation (PVFA) that ranks outstanding orders based on urgency. The model is implemented at an Ethiopian brewery and benchmarked against a First-In First-Out (FIFO) policy across three replenishment strategies: Retailer-Managed Inventory (RMI) with Poisson-generated orders, threshold-based replenishment with a fixed lead time, and Vendor-Managed Inventory (VMI) where order timing is optimized. Results show that, depending on the replenishment strategy, the PVFA reduces unmet demand by 53–98%, lowers extended service times by 65–93%, and frees 8–16% of fleet capacity to be repurposed, with the total number of orders served ranging from a 5% decrease to a 22% increase. These findings demonstrate that inventory-driven delays can be effectively incorporated into a dispatch model, enabling soft handling of inventory constraints. The PVFA consistently outperforms the FIFO benchmark across all scenarios, demonstrating its ability to anticipate even in reactive systems. It also exhibits lower variance on Key Performance Indicators (KPIs), indicating robustness under different replenishment settings. As such, the model is well suited for environments where order policies vary or inventory-driven delays are common. Future work may address dynamic order generation, parameter uncertainty, and extensions toward the general Inventory Routing Problem (IRP).

Online grocery delivery has seen accelerated growth in the last decade. This growth introduced new challenges in operations, such as damaged groceries upon delivery. Companies have started collecting relevant data, yet operations remain largely disconnected from the data. Motivated by this gap in the industry, we formulated the main research question of this study: "How to improve quality of service by integrating customer feedback in a bin-packing model in the context of online groceries?". A review of recent literature revealed two complementary gaps: limited applications of machine learning in offline one-dimensional bin packing where ML directly assists with the packing process, and the lack of customer feedback integration in bin packing models to improve quality of service.

Therefore, we propose a novel two-part theoretical framework to process historical data of customer damage reports into quantifiable parameters that can be used by a bin packing model. In the first part, we formulate a predictive task and propose a classification machine learning model which provides a probability of damage for the bag, given a set of bag characteristics obtained from customer data. In the second part, we propose to integrate the machine learning component into a bin packing model as a weighted term in its objective function. This integrated model comprises our damage-aware bin packing model.

The proposed methodology was implemented and evaluated through a case study using real data from the daily operations of the online supermarket Picnic. As part of the experiments, we analysed over 60 million articles across 2.2 million deliveries. We started with a comprehensive data analysis to explore the relationships in the data and identify trends. We, then, developed and trained two machine learning variants, a logistic regression and an ensemble extreme gradient boost model (XGBoost) to fulfil the predicting task of bag-damage probability estimation. We applied random undersampling to the training dataset to mitigate the extreme class imbalance (0.41% damage rate). Then, we used the logistic regression model as the ML component and implemented a damage-aware bin packing model. We defined strategic empirical metrics to measure its performance and constructed an evaluation framework using counterfactual analysis on 6000 representative deliveries.

Our findings validate the technical feasibility of integrating customer feedback data into a bin packing algorithm. The damage-aware variant recorded a 14.1% relative reduction in the average probability of damage across all bags of the deliveries tested. On the other hand, extreme class imbalance severely limited the performance of the ML models trained (1% precision score in real operational conditions). As a result, a meaningful review of the economic impact of the model is not possible. The experiments illustrated sensible item movements across the bags tested, measured with some empirical key risk parameters, such as item categories, packaging types, and bag density. The implementation showed a tolerable computational overhead of around 18%, indicating that it is realistic to deploy an efficient model to real operations.

Public transportation systems operate in dynamic and unpredictable environments, necessitating close monitoring of real-world operations. Integrating data into route design, scheduling, and policy formulation is essential for creating an efficient and sustainable transit network. This thesis utilises historical bus travel time data to predict future bus travel times, enhancing the analysis of network reachability.

A literature review identified four prediction models to be investigated: Historical Average, Vector AutoRegression, Random Forest, and Long Short-Term Memory deep neural network. A case study involving six bus lines in Groningen was formulated, providing the necessary KoppelVlak 6 and GTFS schedule datasets to be used as inputs for the prediction models. During model development, patterns in historical travel time data were identified, and prediction accuracy was evaluated. The output from the most accurate prediction model was then utilised as input for the reachability analysis.

The analysis demonstrated that complex machine learning models, such as Random Forest and Long Short-Term Memory deep neural networks, yielded the most accurate predictions. The time of day when the journey occurs is particularly predictive of travel and dwell times. Integrating these predictions into a reachability analysis revealed instances of increased reachability, decreased reachability, and missed transfers compared to the original schedule.

The findings demonstrate that travel time prediction models can significantly enhance reachability analyses. This more accurate representation of reachability can be used to identify systemic issues in the design of the public transportation network, allowing for interventions to improve performance.

Recent years show an increasing number of people who use regional public transport, resulting in higher pressure on the operators. This increased pressure results in more disruption for passengers. This paper aims to develop a methodology for estimating the criticality of assets in a regional public transport network, taking the Rotterdam tram system as an example. The consequences of disruptions of each asset in the network are analysed by calculating the effect the disruption has on the average path length for all passengers. Each asset in the network is disconnected one by one to simulate this breakdown. The lines used in Rotterdam are analysed using real passenger data. Results show that 'Erasmus brug' is the most critical asset in the network, as expected by experts from RET. Aside from this, the model presents that the assets around 'Rotterdam Centraal' are less critical than expected because of the alternative routing options for the trams and passengers. This highlights that using this model can give new insights into critical assets in networks.

Commercial waste collection in Dutch cities is a free market in which many waste collectors
operate independently. Collaboration among waste collectors is a relatively recent phenomenon
that is still in its infancy. This study develops a Python-based simulation model using OSMnx and
Google OR-tools to assess the potential benefits of collaboration in waste collection in Rotterdam. Four collaborative scenarios are simulated, three of them in a collaborative setting. Results show that collaboration can significantly reduce the distance traveled by vehicles and the costs for waste collectors, with smaller waste collectors benefiting the most. While full collaboration offers the greatest efficiency gains, even constrained partnerships improve operations. These findings highlight the potential for collaboration to enhance urban logistics and sustainability.

Urban waterways play a crucial role in transportation, recreation, and logistics, yet their traffic capacity remains understudied compared to, for example, road and pedestrian traffic. This paper presents a novel methodology for estimating the capacity of urban waterway networks using empirical data, with a case study focusing on Amsterdam. Existing theoretical capacity models require calibration with real-world observations, which are often lacking. This study uses empirical data to develop a macroscopic capacity estimation model based on intensity-density relationships. Due to the distinct characteristics of urban water traffic, such as reaction times, risk adversity, regulations, and the dominant use case, traditional fundamental diagram models are not trivial to apply. Instead, an intensity percentile-based approach is introduced to define capacity, complemented by an important capacity marker that determines whether the estimated capacity represents the likelihood of the capacity being the actual capacity. The methodology is calibrated using expert-identified bottlenecks and validated through expert validation. Results indicate that the developed methodology shows promise for the capacity estimation of urban canals from dynamic data. However, further refinements are needed in the input data, the calibration of the model, and improving the methodology to account for missing key influencing factors to the traffic characteristics. Still this research provides a foundation for empirical data-driven waterway traffic research, offering policymakers insights for optimizing urban water transport operations.

Increasing congestion and environmental pressures in urban logistics, alongside growing population demands, necessitate innovative solutions. This thesis tackles synchronized two-echelon routing problems within multimodal logistics, specifically exploring the potential of integrating waterborne transport. By developing optimization models and solution approaches for the comprehensive design and management of these systems—encompassing strategic location, tactical allocation, and operational routing and synchronization—the results promote a transition toward more sustainable and efficient urban logistics, ultimately fostering healthier, more livable, and economically vibrant cities.

The increasing volume of global freight trade, coupled with economic growth, necessitates ongoing innovation in optimizing freight operations. Over the past decade, the concept of synchromodality has been explored to encourage a modal shift from unimodal to multimodal transport. Synchromodality, with its flexibility feature, can create more resilient freight transport systems. Various models employing different techniques have been proposed to establish a resilient synchromodal framework capable of reacting to disruptions. However, there are only few studies addressing the unknown duration of disruptions. This research proposes a learning-based modular framework comprising to capture the dynamics of disruptions in multimodal transport and learn to make more effective decisions, thus addressing the challenge of limited prior knowledge about disruptions and enabling fast responses to disruptions.

This project presents a comprehensive study on the development of Integrated Water and Land-based Transportation (IWLT) Systems for city logistics. The research addresses the growing challenges of urban traffic by proposing a decision model for multi-modal transportation systems that leverages waterways alongside traditional road networks. The problem is defined as a two-echelon multi-trip location routing problem with satellite synchronisation (2E-MTLRP-SS), incorporating capacitated vehicles, multiple depots and time constraints. A decomposition-based decision model is introduced, breaking down the problem into manageable sub-problems interconnected through synchronisation in time, space, and load. The decision model uses metaheuristics to be able to handle large-scale, realistic problems and provide feasible solutions for real-life applications. The model’s effectiveness is demonstrated through a case study in Amsterdam, showing the potential of IWLT systems to reduce congestion-related issues and improve the livability of cities. Different scenarios for the IWLT system are investigated, to assist Amsterdam's system developers in making design choices for implementation. The proposed decision model is widely applicable to multi-modal transportation systems all over the world.

There is a growing demand of various metals originating from the sustainable energy transition. The current supply and investments plans for these critical minerals are insufficient to enable this green energy transition. Other problems are the dependency of third countries and multiple drawbacks of land-mining. Deep-sea mining could be a great outcome, by collection of nodules on the seabed in the CCZ area. Different mining trials have been executed, but large-scale mining has not started yet. So far, there is no design guideline for a polymetallic nodule dry bulk (transshipment) terminal. The design and construction phase for a dry bulk terminal takes several years, therefore, the design of such a terminal should be investigated. This thesis will focus on the design of a transshipment terminal to accommodate these large-scale mining activities in the future. The logistics and location determination of this terminal are also taken into account. The thesis contains a case-study which is formed in collaboration with Allseas. For this case-study, the simulation-based approach is chosen to design the transshipment terminal. This approach integrates many stochastic variations, identifies bottlenecks, and allows for sensitivity analysis or ’what if’ scenarios. This method was compared to traditional design methods that rely on rules of thumb and practical experiences.

Multi-agent path finding (MAPF) is the task of finding non-conflicting paths for multiple agents that operate in a environment with shared resources.
Finding an optimal solution quickly becomes intractable for many applications and consequently suboptimal methods are also explored extensively in literature.
This work presents the Decentralized Optimization (DECOP) algorithm: a novel receding horizon control algorithm that exploits insights from MAPF research as well as decentralized control. In the proposed framework, each travelling agent communicates with agents in its proximity to solve a local MAPF problem that considers only a selected tractable number of agents. Inter-agent cooperation and conflict free operation are induced through applying a common local optimization policy during parallel local optimization and through a subsequent path reservation scheme based on random priorities. Inter-agent communication consists of sharing respective route alternatives from which additional information with regard to an agents' entanglement can be inferred which can also be included in the local optimization cost function.
Comparative results with other decentralized algorithms show that the DECOP algorithm yields competitive results while guaranteeing conflict free operations, with limited required communication and without the need of any training time. Among many degrees of freedom to be explored further, including information about the entanglements of an agent's route alternatives in the common policy for local optimization yields an increase in performance and suggests an increased extent of induced cooperation.

Due to the challenges of traffic congestion, pollution, and low transportation efficiency in urban areas, Shared Automated Vehicles (SAVs) are considered a solution that can integrate into existing transportation frameworks to improve traffic performance and environmental sustainability. Currently, there is a research gap in managing the fleet size of SAVs under different matching strategies. Although the literature on the operational and environmental benefits of SAVs is growing, few studies systematically analyze how different fleet management strategies affect service levels and road congestion. This thesis fills this gap by exploring the impacts of Immediate Decision and Batch Offer strategies on fleet size and their corresponding effects on urban traffic performance.

The thesis focuses on the integration of SAV with mixed mobility systems, using SAVs to entirely replace current Mobility on Demand (MoD) services, employing an agent-based simulation method to study the impacts of fleet size and matching strategies on urban traffic performance. This research employs a case study approach, selecting the Haidian District in Beijing as the research area. It uses an innovative demand generation model based on travel distances as simulation request inputs. The model utilizes travel distance distribution functions based on historical data, road network information, and taxi GPS data to generate traffic demand based on network nodes through traffic zone division, probability density calculation, and travel distribution. The model employs the K-means clustering algorithm to divide traffic areas based on the density of traffic nodes rather than specific road network conditions, a macroscopic, low-resolution method with lower computational complexity suitable for large-scale road networks.

The mode choice model in this study analyzes how passengers make decisions between different modes of travel. The model, using a discrete choice framework, studies how passengers choose between SAVs(Shared Automated Vehicles) and POVs(privately operated vehicles). Employing the Logit model, the study considers several key factors influencing travel choice, including travel time, comfort, reliability, and environmental impact, to provide mode selection in simulations. This part emphasizes the importance of travelers' subjective perceptions.

The application of simulation and the Macroscopic Fundamental Diagram (MFD) are central methodologies of this study. In the simulation analysis part, this research uses the agent-based simulation tool FleetPy, focusing on evaluating two key matching strategies: Immediate Decision Simulation (IDS) and Batch Offer Simulation (BOS). By simulating different fleet sizes and matching strategies, it explores the dispatching and operational performance of Shared Autonomous Vehicles in urban road networks and compares them with traditional taxi systems.

The findings indicate that replacing taxi systems with an SAV system can significantly improve fleet efficiency, achieving the same service level with a smaller fleet size. In both small and large fleet scenarios, the SAV system demonstrates higher operational efficiency. In scenarios with smaller fleet sizes, the Batch Offer (BOS) strategy provides higher service levels. By batch processing and optimizing multiple requests, it can effectively improve the success rate of matching and reduce waiting times, while Immediate Decision (IDS) performs better in terms of service timeliness. Simulation results also show that as the fleet size increases, the advantages of the IDS strategy become apparent but may lead to wastage of vehicle resources.

Container transport requires constant improvement to be more efficient and sustainable. Improvements can take the form of innovations, investments or policies. This thesis develops choice-driven methods to enhance decision-making in intermodal freight transport by focusing on demand, supply, and competition aspects considering heterogeneous behaviors and interactions of actors. The proposed methods support policy-makers, transport companies, and planners in foreseeing the impacts of their decisions on the involved actors.

This thesis concerns the fundamental problem of learning the behavior of decisionmaking agents using only observations of how they act in different situations. As humans, we do it all the time, and have been doing it since birth: think about how a child learns to speak and walk. However, this thesis does not only focus on imitating. We go a step further and try to learn why an agent does what it does. In other words, what is the agent’s objective, which led them to act in a certain way? This is a much harder question, but also much more rewarding when answered correctly: knowing the agent’s motivation allows us not only to imitate but also to understand or even influence the agent’s behavior. For this purpose, we ask the question: what is the agent optimizing for when making decisions? For instance, imagine a consumer agent that buys a certain set of products given their budget. To model the consumer’s behavior, we interpret their action as buying the products with maximum utility, given a limited budget. Thus, learning how the consumer evaluates each product (that is, the utility of each product for the consumer) would allow us to understand, replicate, and possibly influence their behavior. Mathematically, we model the decision process of the agent as an optimization program, and we use Inverse Optimization (IO) as a tool to “reverse engineer” the agent’s optimization program from observed behavior. This thesis can be divided into two parts. First, we dive into the mathematical formalization of the IO problem. We look at the geometry of thema thematical objects emerging from IO problems, and we discuss what it means to solve the IO problem and different ways to do it, proposing tractable reformulations and efficient algorithms. In the second part of this thesis, we develop a tailored IO methodology to solve IO problems emerging from routing problems. We test the potential of our methodology for modeling human driving behavior on real-world problems using data from the Amazon Last Mile Routing Research Challenge. We achieve excellent results, showcasing the potential of IO to solve real-world problems. Additionally, we also developed Inv Opt, an open-source Python package to solve general IO problems.

A comprehensive understanding of shippers’ preferences can help transport freight forwarders create targeted transport services and enhance long-term business relationships. Nevertheless, limited research examined the benefit of considering shippers’ preferences in the decision-making of synchromodal transport planning and the collection of relevant data is still not straightforward.
This research proposes an innovative framework to learn shippers’ preferences in synchromodal transport operations and optimize transport services accordingly. A preference learning method is developed to capture shippers' preferences through pairwise comparisons of transport plans. In order to model the underlying complex nonlinear relationships and detect heterogeneity in preferences, artificial neural networks are employed to approximate shippers' utility for a specific plan. Based on the learned preference information, a synchromodal transport planning model with shippers’ preferences (STPM-SP) is proposed, with the objectives of minimizing the total transportation cost and maximizing shippers’ satisfaction. An Adaptive Large Neighborhood Search algorithm is developed for solving this optimization problem. This algorithm takes into account the two different objective functions and searches for Pareto solutions to the planning problem.
A case study is conducted based on the European Rhine-Alpine corridor to demonstrate the feasibility and effectiveness of the proposed methodological framework. Basic discrete choice models, binary logit models, are used as benchmarks for preference learning and the synchromodal transport planning model without preferences (STPM) is used as the benchmark for planning. The results show that the proposed preference learning method has better predictive power than the baseline model, achieving higher accuracy and lower variation. With the consideration of shippers’ preferences, STPM-SP can significantly increase shippers' satisfaction with transport services. Scenarios with different types of preferences are tested and results show that the average of maximum improvements in satisfaction reached 37.76%. This research contributes to learning shippers' preferences in the transport operation process and highlights the importance of incorporating these preferences into the decision-making process of synchromodal transport planning.

In the upcoming years the Municipality of Amsterdam has to cope with renovation works to approximately 200 kilometers of quay walls and 800 bridges. These projects need to be executed in the least burdensome manner, which is why executing parties are obliged to perform their construction logistics via waterways instead of the road. Stakeholders want to have more insight into the consequences of using waterborne construction logistics and what the possible choices and decisions are within the system. In this research a waterborne construction logistics system is presented from which waterborne scenarios can both be compared with each other and with the conventional road modality. A decision support tool is developed and used to quantify these consequences and it can be concluded that water is always the preferred alternative when looking at movements and vessel/vehicle kilometers through the city, which is for both indicators a decrease of approximately 80%. From this analysis can be concluded that in the most optimistic scenario it is possible for waterborne logistics to become less expensive than road transport, but under the conditions that the average transported load and vessel speed are rather high. It is therefore recommended to further investigate the optimum between vessel capacity, loading rate and vessel speed and for the Municipality of Amsterdam to involve executing stakeholders to contribute to reaching this optimum.