ABSTRACT Medical on‐demand couriers play an important role in urban medical services. They transport medical supplies, tests, blood, or organs between different locations in the city within a short amount of time. Transportation orders are issued spontaneously and differ in their priority and time restrictions. Since all orders must be served, the goal is to minimize the priority‐dependent delay penalty of all orders. To this end, we propose an anticipatory re‐optimization approach by integrating a parametric Cost Function Approximation (CFA) into a Large Neighborhood Search (LNS). Whenever a new order is placed, the LNS iteratively optimizes the current route plans. The CFA addresses the tradeoff between minimizing delays for pending orders and the fleet's flexibility to react to future orders. To this end, it is used both to evaluate decision candidates (i.e., new route plans) and to assess insertion options within the LNS repair operator. We demonstrate the advantages of our approach in comprehensive computational experiments and provide insights into the method and problem.

ABSTRACT With rising customer expectations and increasing computational potential, many transport, manufacturing, and production operations face real‐time decision making in stochastic dynamic environments. Decision makers must find and adapt complex plans that are effective now but also flexible with respect to future developments. The challenges of searching a high‐dimensional constrained decision space for effective and flexible decisions are reflected in the three parts of the Bellman equation: the reward function, the value function, and the decision space. In the literature, reinforcement learning (RL) has shown potential to quickly evaluate the reward‐ and value function for a limited number of decisions but struggles to search a constrained decision space present in most planning problems. The question of how to combine the thorough search of the complex decision space with RL‐evaluation techniques is still open. We propose two RL‐based solution methods and detail a third one to search for and evaluate decisions in an integrated manner. Each method is inspired by one component of the Bellman equation. The first two methods dynamically shape the reward function or decision space to encourage effective and flexible decisions or prohibit inflexible decisions. The third method models the Bellman equation as a mixed‐integer linear programming formulation in which the value function is approximated by a neural network. We compare our proposed solution methods in a structured analysis for carefully designed problem classes. We demonstrate the effectiveness of our methods compared to prominent benchmark methods and highlight how the methods' performances depend not only on the problem classes but also on the instances' parameterizations.

This special issue of Networks focuses on recent and forthcoming trends in city logistics and urban mobility. Specifically, new business models and services are enabled due to technological advancements in information technology, vehicle autonomy and connectivity, vehicle electrification, payment methodologies, and clearing solutions. Notable trending topics include: drones and autonomous ground vehicles, physical internet for parcel and cargo delivery, electric vehicles and micromobility, ride-sharing and ride-hailing, crowd shipping and the gig economy, same-day delivery of goods and meals, collaborated transportation, and brokering. We have invited and received a broad range of operational research papers that study these topics for this special issue. In what follows, we briefly describe the seven research papers included in this issue, covering a wide range of trending topics. To these papers, we have added a survey paper that provides an overview of future directions for research in urban mobility and city logistics 4. In particular, the survey focuses on three main developments: vehicle autonomy, crowdsourced logistics, and urban micro-consolidation centers. Abbaas and Ventura 1 study the continuous deviation-flow refueling station location problem on a general network that arises in the context of alternative fuels. They propose an exact algorithm that determines the endpoints of all refueling segments on each edge of the network that cover the corresponding origin-destination flows. This set of endpoints is used in a set covering model to locate multiple refueling stations under the assumption that every vehicle only refuels once on each way of its round trip. A numerical experiment illustrates the performance of the proposed methodology. Arrieta–Prieto et al. 2 formulate the last-mile delivery problem assisted by urban micro-consolidation centers as a mixed-integer quadratically-constrained program (MIQCP) and develop a greedy heuristic solution, inspired by decomposition algorithms for large-scale optimization. Simulation results and a case study in Manhattan (NY, USA) demonstrate that the proposed heuristic can provide good results with respect to the exact solution and can be used for freight-efficient urban design and policy planning. Haferkamp and Ehmke 3 study demand and fulfillment control policies in ride-sharing systems. They classify the existing policies in the literature and explore the effectiveness of such policies under varying conditions in order to identify benefits and risks for ride-sharing systems. For this purpose, the authors define policies that differ in the optimization of demand and/or fulfillment control through the exploitation of either confirmed or complete information. Experimental results demonstrate that demand and fulfillment control have different effects on ride-sharing systems' performance and service quality. Le et al. 5 analyze the impact of optimized coordination between autonomous vehicles and traffic-light controlled intersections. To this end, they develop a mixed-integer linear programming (MILP) model based on a microscopic traffic model with centrally controlled autonomous vehicles and they incorporate traffic-light switching regulations. Their model allows an estimation of the maximum performance gains due to improved communication and serves as a benchmark for decentralized approaches. An evaluation of the numerical results with a traffic simulation tool shows that performance indicators such as time, energy, and emissions can be reduced significantly compared to simulated real-world traffic. In addition, the authors evaluate the performance of a moving horizon approach which exposes the trade-off between performance and real-time feasibility. Ostermeier et al. 6 present an approach to cost-optimal routing of a truck-and-robot system for last-mile deliveries with time windows. Their solution algorithm is based on a combination of a neighborhood search with cost-specific priority rules and search operators for the truck routing. An exact solution approach and a heuristic approach are proposed for the robot scheduling sub-problem. The numerical experiments show that the approach can reduce last-mile delivery costs significantly. In particular, a case study indicates that the truck-and-robot concept reduces last-mile costs by up to 68%, compared to truck-only delivery. Finally, a sensitivity analysis provides managerial insights regarding the setting in which truck-and-robot deliveries can efficiently be used in the delivery industry. Rosenfeld 7 studies a retrieval problem arising in puzzle-based storage systems. The problem consists of a general number of loads to be retrieved and arbitrary numbers of I/O points and empty locations on a two-dimensional lattice graph. The author analyzes the theoretical characteristics of the problem and proposes a set of graph search algorithms to tackle these characteristics. Numerical results demonstrate that the proposed algorithms can optimally solve moderate-sized puzzle-based storage problems using a limited amount of memory and in a reasonable amount of time. Voigt and Kuhn 8 consider a parcel delivery service that ships parcels from pickup to delivery points using regular drivers and occasional drivers where drivers can hand over or collect parcels at predefined transshipment points. The company seeks to minimize the overall costs that arise from the total distance traveled by regular drivers and the compensation paid to occasional drivers. The authors model the pickup and delivery problem with transshipments and occasional drivers (PDPTOD) as a mixed-integer programming (MIP) problem. They develop a specialized heuristic approach based on an adaptive large neighborhood search (ALNS). The numerical study reveals the impact of the number and location of transshipment points on the cost advantages achieved by integrating occasional drivers into the delivery process. It also shows that the cost savings are highly sensitive to the assumed flexibility and compensation scheme of occasional drivers. The authors thank the authors for their excellent submissions that have made this special issue come true. In addition, the authors extend our thanks and appreciation to the reviewers who contributed their time and hard work, providing valuable comments and suggestions. Lastly, the authors thank Professor Bruce Golden for delegating this special issue to us and for his support throughout the editorial process. Data sharing is not applicable to this article as no new data were created or analyzed in this study.

Abstract This survey article provides an overview on future directions for research in urban mobility and city logistics. It sets a focus on three particularly serious changes in the business models: vehicle autonomy, crowdsourced logistics, and urban micro‐consolidation centers. In the future, service fleets might fully or partially be autonomous which brings new operational opportunities and challenges. In many business models, crowdsourcing jobs are already common. While this might save costs, it also leads to uncertainty in the available workforce and their behavior. Finally, micro‐consolidation centers enable the use of smaller, cheaper, and emission‐friendlier vehicles, but lead to more complex planning and operations. For each topic, the article presents an overview on the relevant literature as well as important and open research challenges.

Due to new business models and technological advances, dynamic vehicle routing is gaining increasing interest. Especially solving dynamic vehicle routing problems with stochastic customer requests becomes increasingly important, for example, in e‐commerce and same‐day delivery. Solving these problems is challenging, because it requires optimization along two dimensions. First, as a reaction to new customer requests, current routing plans need to be reoptimized. Second, potential future requests need to be anticipated in current decision making. Decisions need to be derived in real‐time. The limited time often prohibits extensive optimization in both dimensions and the question arises how to utilize the limited calculation time effectively. In this paper, we analyze the merits of reactive route reoptimization and anticipation for a dynamic vehicle routing problem with stochastic requests. To this end, we compare an existing method from each dimension as well a policy allowing for a tunable combination of the two approaches. We show how the appropriate optimization combination is strongly connected to the degree of dynamism, the percentage of unknown requests. We also show that our combination does not provide significant benefit compared to the respectively best optimization dimension.

In this paper, we analyze how drones can be combined with regular delivery vehicles to improve same‐day delivery performance. To this end, we present a dynamic vehicle routing problem with heterogeneous fleets. Customers order goods over the course of the day. These goods are delivered either by a drone or by a regular transportation vehicle within a delivery deadline. Drones are faster, but have a limited capacity as well as require charging after use. In the same‐day context, vehicle capacity is not a constraint, but vehicles are slow due to urban traffic. To decide whether an order is delivered by a drone or by a vehicle, we present a policy function approximation based on geographical districting. Our computational study reveals two major implications. First, geographical districting is highly effective increasing the expected number of same‐day deliveries. Second, a combination of drone and vehicle fleets may significantly reduce the required delivery resources.