TU Delft Repository

In the field of public transportation, environmentally friendly and convenient transportation modes are the future trends. The ride-hailing services is an important component of them. However, current ride-hailing systems, particularly the matching systems, still have issues related to low system efficiency and bad user experience. Although existing ride-hailing rider-driver matching system can allocate travel demands and drivers to a certain extent, they still have deficiencies in certain scenarios. For example, they cannot ensure effective rider-driver matching during peak hours, or they cannot find a good balance between pick-up distance and matching rate. As Reinforcement learning (RL) has been proven in many studies to be applicable and effective in solving complex and dynamic optimization problems. This study aims to explore how Reinforcement Learning (RL) can be adapted to the ride-hailing matching system to optimize system efficiency and user experience through a dynamic matching radius policy. The research objective of this study is to simulate an actual ride-hailing system and use RL to train a policy. This policy can output an optimized dynamic matching radius in real-time based on real-time rider-driver demand-supply relationship, hence achieving a higher matching rate, a shorter average pick-up distance, and a higher driver utilization rate of the ride-hailing system.
Adapting Reinforcement Learning (RL) to optimize the ride-hailing system's matching radius has several difficulties and challenges due to the uncertainties in the real-world rider-hailing market. Traditional approaches are normally static, solving the matching problem at specific times through mathematical models. However, these methods often perform inconsistently when dealing with fluctuating ride-hailing supply-demand relationships, particularly during peak hours. On the other hand, the dynamics and complexity of the ride-hailing market and the ride-hailing environment also make it difficult to model the ride-hailing system. The ride-hailing market is easily affected by many variables, such as weather conditions and local traffic conditions. When quantitatively optimizing the matching radius of the ride-hailing matching system, it is critical to reasonably control irrelevant variables. To address these challenges, this study models the ride-hailing matching problem as a Markov Decision Process (MDP). Based on the defined MDP, a ride-hailing matching simulator is developed. Some assumptions and simplifications are also made to ensure high realism while reasonably controlling irrelevant variables and uncertainties. Multi-replay-buffer Deep Deterministic Policy Gradient (MDDPG) algorithm is then applied to handle the optimization problem of the ride-hailing matching radius. Through the interactions between the MDDPG agent and the developed simulator, feedback rewards are received for the agent to improve the policy. The proposed method is then validated in a case study showcasing the application of the developed simulator and the RL algorithm in a real-world scenario in Austin, Texas. The case study includes an analysis of the current ride-hailing market in Austin, how to apply the simulator based on it, the implementation details of the RL algorithm, and the resulting performance improvements. The results of the case study show that the actions obtained from the proposed method outperform all the baselines in multiple scenarios, highlighting the benefits of using Reinforcement Learning to improve ride-hailing efficiency and user experience.
To conclude, the optimization method proposed in this study applies an advanced Reinforcement Learning approach to the ride-hailing system, successfully improving overall efficiency and user experience. The results of this research demonstrate the potential of Reinforcement Learning in optimizing ride-hailing matching systems, offering a promising direction for further exploration. This study lays a solid foundation for future research to build upon, encouraging the development of more optimization methods with RL technologies that can enhance the effectiveness and adaptability of ride-hailing system in increasingly complex and dynamic environments.

User feedback plays a significant role in helping recommendation systems to make personalized and accurate predictions. Despite the fact that many methods of collecting user feedback have been proposed, little research exists that addresses both the breadth and depth of data collected. In this study, we incorporate personal reflection into traditional crowdsourcing tasks and investigate how it facilitates people's reflection on their usage of YouTube. We present a novel crowdsourcing approach with personal reflection integration based on several design principles, which allows participants to reflect on contextual factors and personal values of using YouTube through guided context recall and evaluations, therefore gathering deeper insights into people's preferences and behaviors on a large scale. We conducted a user study involving 20 participants and explored the insights generated from their textual answers and the role of design principles in the reflection process. This approach successfully enables multiple participants to conduct the study simultaneously, thereby reflecting on their watching behaviors and preferences. Based on the quantitative and qualitative analysis of the findings, we sum up the implications of this approach to provide guidance for the YouTube recommendation system and point to the directions for the design of similar studies in the future.

This research addresses the challenges of installing floating wind turbines on semi-submersible platforms, with a focus on managing relative motions during offshore integration. Traditional methods, which rely on large onshore marshaling yards, are constrained by the limited space available at ports to meet the increasing demands of the offshore wind industry.

In response, the study proposes a shift to floating-to-floating installation, where turbine integration takes place on a moored floater directly at the wind farm. A major challenge in this approach is controlling the relative motions between the two floating bodies. To address this, a multibody model is developed in OrcaFlex, simulating the interaction between a heavy lift vessel and the VolturnUS-S platform. This model highlights the complexity of relative displacements, providing the foundation for introducing concepts aimed at minimizing these horizontal motions.

The chosen concept features a donut-shaped external platform equipped with winches for horizontal motion compensation. In the proposed 10-step installation process, the turbine tower is suspended from a crane while a gripper restricts pendulum motions, and an Active Heave Compensation system manages the relative heave motions. The external platform, suspended below the tower as it is attached to the gripper, is first lowered and secured onto the floater. Once in place, the tower is lowered, with the winches providing the necessary stiffness to control relative horizontal displacements during this phase.

The winch wires are modeled as springs with equal stiffness applied in both horizontal directions. Time-domain simulations in OrcaFlex, validated by an analytical model in MATLAB, reveal that a stiffness of 8887 kN/m is required to effectively limit horizontal displacements to 0.2 m for significant wave heights up to 2.5 m. This design choice ensures 92% workability for this installation step, considering a maximum crane angle of 2 deg and a maximum horizontal relative displacement of 0.2 m. Ensuring functionality under the simulated sea states reveals a design requirement of 1823 kN for the compensation system in the horizontal direction for Hs = 2.5 m and a Tp ranging from 6 to 11 s. While this force exceeds the capacity of a simple winch, it remains within a feasible range for a dedicated compensation system.

This study analyzed the profitability and cost constraints of thin film solar PV modules for commercial sized lightweight roofs by using the reversed LCOE method. The focus will be on a particular technology of solar PV, namely, thin film solar PV module technology. Thin film solar PV modules are able to serve the niche market of commercial scale lightweight roofs, without competition with the highly matured traditional solar PV market. As such, the objective of this study is to derive actionable insights into the cost constraints of thin-film solar PV modules for these applications. Using the reversed LCOE method, figures show that there is a high variability in conditions that influences the performance and the profitability of thin film solar PV modules in a commercial scale lightweight roof project. This study shows that both profitable and unprofitable scenarios are possible and provides an analysis on how various conditions influence the project’s profitability.

Hazards that arise from volcanic settings are abundant, and the continuous monitoring of volcanoes remains an important task to protect civilians. Early warning systems have an enormous impact in effectively informing the general public of potential hazards and reducing the associated risks involved with volcanic eruptions. Volcano monitoring infrastructures rely on various complementary geophysical techniques, and the maintenance of such a diverse system of instruments is often challenging. This dissertation concerns the automated assessment of geophysical instruments, and the application of microgravity observations to long-term volcano monitoring.