This study focuses on two-echelon synchronized logistics in a multimodal distribution network for integrated road-based and waterborne transportation motivated by practical city logistics challenges. To evaluate the feasibility and requirements of a multimodal transport system as a potential future logistics solution for supplying hotels, restaurants, and cafés (HoReCa) in the city center of Amsterdam, we propose a methodology to evaluate the effect of various satellite service network designs, fleet sizes of vessel and street vehicles, and city access time window constraints (permitted working hours that can be set by the municipality through policies). To tackle the complexity of the problem, a decomposition-based heuristic is developed to solve large-size instances for a rich 2E-VRP variant. The decomposition-based heuristic consists of three main phases: a location phase for determining satellite locations, an allocation phase for establishing assignments of satellites and customers, and a routing phase for solving the routing problem given operational settings. In addition, we propose various modeling approaches to represent system designs with different operational limitations in practice. First, we consider two transshipment types: one-to-many and one-to-one transshipments. The transshipment type is related to satellite resources in terms of space availability and lifting capabilities. Secondly, we propose two simulation-optimization frameworks for iteratively configuring the workload of and workforce at satellites. We demonstrate the methodology by conducting a case study for supplying over 1600 HoReCa in the city center of Amsterdam. The results indicate that balancing the workload of satellites is critical for scenarios with few permitted working hours. By configuring the workforce at satellites, a good balance between the number of vessels, street vehicles, and street-level distance is found for scenarios with more permitted working hours. We show that the transshipment type significantly affects the service level and system requirements to meet customer demand, with a sensitivity analysis on transshipment times at satellites indicating significant benefits for reduced transshipment time in the case of one-to-one transshipments. Moreover, the computational experiments show several trade-offs between the required number of satellites, vessels, street vehicles, and permitted working hours to meet customer demand.

Vehicle routing problems (VRPs), a generalization of the traveling salesman problem, are extensively studied combinatorial optimization problems for their practical application. Many solution methods, e.g., exact and heuristic algorithms, have been proposed in the last few decades but require relatively much computation time due to the NP-hard nature of the VRP. Additionally, to build such algorithms, much expert knowledge is required. The recent developments in a subfield of machine learning, deep learning, make it possible to solve routing problems in a purely data-driven manner or assist heuristic methods. This requires less problem-specific knowledge and can outperform the traditional solution methods in terms of objective value and computation time. In this literature review, introductions for the vehicle routing problem and machine learning are given first. Then, an existing categorization for algorithmic machine learning structures is described. The main body of this report reviews recent machine and deep learning methods to solve static and dynamic routing problems. The reviewed literature is summarized in tables that indicate the main characteristics of each work. Moreover, the results of several machine and deep learning based solution methods for the static VRP are compared to each other. Lastly, the challenges and opportunities for future research of deep learning based solution methods for the vehicle routing problem are discussed. The main challenges of these methods are scalability, generalization, and adaptability. Therefore, future research could be focused on addressing these challenges to improve deep learning methods for practical routing applications.