ArticlePDF Available

Abstract

In pickup and delivery problems vehicles have to transport loads from origins to destinations without transshipment at intermediate locations. In this paper, we discuss several characteristics that distinguish them from standard vehicle routing problems and present a survey of the problem types and solution methods found in the literature.
... The VRP with time windows (VRPTW) studies problems where each customer should be served only within a specified time interval or time window (Bräysy and Gendreau, 2005a). In a more specific context, the pick-up and delivery problem with time windows (PDPTW) (see, e.g., Savelsbergh and Sol, 1995;Parragh et al., 2008) assumes that each request specifies the size of the load needed to be transported, the locations where it is to be picked up (the origins), and the locations where it is to be delivered (the destinations). Note that in all of these problems, the time windows are input parameters selected by customers and are not allowed to be optimized in their model. ...
... Our problem is closely related to VRPTW (see, e.g., Desrochers et al., 1992;Bräysy and Gendreau, 2005a,b), Dial-a-Ride problem (DARP) (see, e.g., Cordeau and Laporte, 2007;Berbeglia et al., 2012), pick-up and delivery problem (PDP) (see, e.g., Savelsbergh and Sol, 1995;Berbeglia et al., 2010) and appointment scheduling problems (see, e.g., Gupta and Denton, 2008;Erdogan and Denton, 2013;Berg et al., 2014;Deng and Shen, 2016;Jiang et al., 2017). We refer to Laporte (1992) and Laporte (2007) for classical models for VRP and the related exact algorithms, classical heuristics, and metaheuristics. ...
... Cordeau and Laporte (2007) review the literature of DARP, demonstrate the main features of the problem and provide a summary of the most important models and algorithms. Savelsbergh and Sol (1995) distinguish PDP from standard VRP and present a survey of its models and solution approaches, with a primary focus on deterministic problems. Berbeglia et al. (2010) and Pillac et al. (2013) provide thorough reviews of dynamic PDP and VRP, respectively, where objects or people have to be served in real-time. ...
Thesis
The primary focus of this dissertation is to develop mathematical models and solution approaches for sequential decision-making and optimization under uncertainty, with applications in transportation, logistics, and healthcare-related operations management. In real-world applications, system operators often need to make sequential decisions, that may involve both discrete and continuous variables under data uncertainties. These problems can be modeled by multistage stochastic integer programs (MS-SIP) that are, however, computationally intractable due to the well-known “curse of dimensionality” issue. MS-SIP assume that the distributions of uncertainty parameters are known and one has access to a finite number of samples of the distributions. In contrary to MS-SIP, multistage distributionally robust integer programs (MS-DRIP) make no assumption on distributions of uncertain parameters. Instead, the optimal solutions are sought for the worst-case probability distributions within a family of candidate distributions, namely, the ambiguity set. Compared to multistage sequential decision models, the two-stage counterparts, namely TS-SIP and TS-DRIP, are easier to solve, where planning decisions are made before uncertainty realizes. In this dissertation, we investigate the four models by developing highly efficient and scalable algorithms and recommend the most practical one in the context of designing and operating complex service systems. Specifically, in Chapter 2, we first study MS-DRIP under endogenous uncertainty, where the probability distribution of stage-wise uncertainty depends on the decisions made in previous stages. We derive mixed-integer linear programming or mixed-integer semidefinite programming reformulations for the min-max Bellman equations, and for the latter we show how to obtain upper and lower bounds of the optimal objective value. We employ the Stochastic Dual Dynamic integer Programming method for solving the resultant MS-SIP. Our numerical results based on facility-location instances show the computational efficacy of our approaches and demonstrate the cost effectiveness of considering decision-dependent uncertainty in the dynamic risk-aware optimization framework. In Chapter 3, we examine the gaps between MS-SIP and TS-SIP with facility-location instances. It remains an open question to bound the gap between these two models using risk-averse objective functions, which indicates at least how much benefits we can gain from solving a more complex multistage model. We provide tight lower bounds for the gaps between optimal objective values of risk-averse multistage stochastic facility location models and their two-stage counterparts using expected conditional risk measures. To speed up computation, two approximation algorithms are proposed to efficiently solve risk-averse TS-SIP and MS-SIP. The aforementioned models and approaches can be applied to a wide range of applications, including smart transportation and mobility-as-a-service. In Chapter 4, we first consider integrated vehicle routing and service scheduling problems with either customer-imposed or self-imposed time windows. We propose TS-SIP to optimize vehicle routes and estimated arrival time or time windows to reduce customers’ waiting, vehicle idleness, and overtime. To fulfill real-time arrived service requests, we develop K-means clustering-based algorithms to dynamically update planned routes and schedules, which can quickly compute high-quality solutions for large-scale instances. Finally, in Chapter 5, we extend the TS-SIP for vehicle routing and service scheduling to cover on-demand ride pooling requests, where we dynamically match available drivers to randomly arriving passengers and also decide pick-up and drop-off routes. We design a spatial-and-temporal decomposition scheme and apply Approximate Dynamic Programming (ADP) to improve computation. Our ADP approach reduces the unsatisfied demand rate dramatically compared to other benchmarks that do not incorporate future information or pooling options.
... The related pickup and delivery problem (PDP) has been extensively studied in recent years, which can be classified into many-to-many, one-to-many-to-one, and one-to-one problems [14]. Savelsbergh and Sol [15] introduce the general PDP, and the loads are transported from origins to destinations without transshipment. Ropke and Pisinger [16] develop a heuristic solution framework based on ALNS to solve the PDP with time [18] study the PDPTW with the mixed-load strategy for catering distribution services, and construct the multi-commodity flow optimization model in the framework space-time-state network. ...
... M c is the maximum number of changed service orders, which represents the number of orders that use the transfer station-based delivery instead of special delivery after adopting the delivery mode based on transfer stations. S 1 is the total profit of the traditional delivery mode, as shown in Equation (15). S 2 is the total profit of the delivery mode based on transfer stations, as shown in Equation (16). ...
Article
Full-text available
The rapid development of instant logistics services has brought not only convenience to people's life but also a great challenge to traffic management. Due to the limited delivery range of instant delivery systems, customers are usually recommended the meals nearby or pay much higher delivery fees for long‐distance delivery. This study proposes a novel order splitting strategy based on transfer stations for long‐distance meal orders to satisfy the diverse customer demands. The meal delivery routing problem is addressed with an order allocation strategy based on transfer stations through developing a three‐stage modelling framework consisting of order combination, splitting and delivery routing for the online‐to‐offline instant logistics services. Normal meal orders are combined by the DBSCAN algorithm, and the cross‐regional long‐distance orders are split by transfer stations. Based on order combination and splitting, a mixed integer programming model is constructed for the meal delivery routing problem and solved by the adaptive large neighbourhood search algorithm. The proposed algorithm converges quickly for the tested instances constructed based on real platform data. The proposed order allocation strategy can expand the delivery scopes of couriers effectively, stimulate more potential orders and guarantee the timeliness of meal delivery.
... Microtransit services involve the movement of commuters between predetermined locations and transit hubs. e services operate either on a fixed schedule and fixed-route manner [22,23] or by following flexible scheduling and flexible routes [24,25]. e authors showed that the more flexible system offers cost advantages over regional systems, especially when transit services are frequent, or transit hubs are close together, with little impact on passenger convenience. ...
Article
Full-text available
With the aid of recent technological advancements, seamless integration of shared mobility services and public transit may offer efficient and affordable connectivity to the transit stations in urban settings, thereby enhancing residents’ mobility. A previous research mainly focused on car-sharing services as a self-standing mode of transportation. However, due to rapid urbanization acceleration and regions’ extension, commuters often combine the fixed-route/fixed schedules public transportation and car-sharing service in one journey. To this end, we study a one-way, station-based electric car-sharing service interaction with public transportation. We propose an integrated route choice and EV assignment model to address the potential of car-sharing services as a feeder to the public transit network. The integrated model consists of two components, operations of the car-sharing service and the commuter’s route choice and the associated mode choice. The service provider decides on the resource levels, allocations, and relocation strategy in the first component. In the second component, the travel options for the commuters are modeled. The two-component model was simulated in an agent-based simulation based on a case study from the state of Qatar. We further extend the integrated model to include the carpooling option, in which multiple passengers sharing the same route can share the same vehicle. Extensive simulation analyses show that the integration can considerably enhance urban mobility and increase public transportation accessibility through enhanced first and last miles linkages. Moreover, the influence of transportation supply and spatial characteristics on the individual mode choice was estimated. Results indicate that public transit ridership can increase up to 17%. Moreover, adding the carpooling option can significantly decrease the number of relocations operations at a minimal impact on the commuters’ trip performance.
... Cette section s'intéresse à une catégorie de problèmes en particulier : "the General Pickup and Delivery Problem" (GPDP). Cette classe a été introduite en 1995 par Savelsbergh & Sol (1995) (2011), la classification permettant de différencier ces problèmes se divise tout d'abord en deux catégories : statique et dynamique. Attention, il faut cependant noter que même si la terminologie est similaire à celle vue dans les systèmes de covoiturage (statiques et dynamiques), les notions désignées ici sont différentes. ...
Thesis
L’utilisation massive des véhicules personnels pour les trajets urbains, a engendré ces dernières années des difficultés qui paraissent aujourd’hui difficiles à surmonter. L’augmentation de la congestion et, de surcroît, des émissions de gaz à effet de serre dans les grandes agglomérations, sont des exemples concrets de ces dérives. De nombreuses solutions permettant de limiter ces aspects ont cependant été évoquées dans la littérature, comme par exemple, le développement de nouvelles lignes de transports en commun, ou encore la création de nouveaux services de mobilité partagée. Bien que ces avancées aient suscité une vive attention, la plupart d'entre elles se sont orientées vers un fonctionnement de plus en plus réactif à la demande des utilisateurs. En d’autres termes, de nombreux travaux visaient à implémenter des méthodes d’appariement dynamique entre les usagers. Ces récentes évolutions se démarquent des approches dites « conventionnelles », classiquement utilisées pour la mise en place de nouvelles lignes de transports en commun.Il semble néanmoins évident que les objectifs de ces différentes approches soient assez éloignés les uns des autres. En effet, outre les objectifs en termes d'aménagement du territoire, les lignes de transports en commun visent à répondre à une demande de mobilité massive et régulière. Alors que les services de mobilité partagée tendent plutôt à satisfaire une demande individuelle et très ciblée. C'est donc pour tenter de s'affranchir des problématiques inhérentes à chacun de ces systèmes, que des méthodes hybrides ont fait leur apparition. Parmi elles, nous pouvons par exemple citer le microtransit, ou les bus à la demande, qui constituent de réelles innovations dans le cadre de la mobilité urbaine. Au-delà de ces nouveaux services, l'émergence de récentes technologies telles que les véhicules autonomes, est venue renforcer l'idée que les transports collectifs pourraient constituer, à terme, une solution réaliste à la décongestion des agglomérations. Cette thèse vise donc à identifier les limites des systèmes de transports collectifs actuels en zones urbaines. Puis à constituer un socle théorique, permettant la conception de lignes de transports optimisées en fonction de la demande de mobilité. Les méthodes déployées devront permettre de quantifier précisément la demande, ainsi que sa répartition spatiale et temporelle. Elles devront également assurer le traitement de données massives, afin de répondre aux problématiques soulevées dans la littérature. Enfin, la généricité des méthodes appliquées devra permettre leur réutilisation dans différents contextes, et avec différents types de services ; permettant ainsi de préfigurer l’arrivée potentielle des véhicules autonomes dans le cadre d'une mobilité urbaine collective et durable.
Article
Deploying drones for rapid pickup and delivery for on-demand customers in the hyperlocal market is unexplored in literature and demands attention. Maximizing customer pickup and deliveries with a limited drone flight endurance is essential but is hard to achieve in practice because of many service requests by on-demand customers. It mandates that drones cover extra mileage to visit docking stations in between services for recharge. Optimally allocating the services on available flight endurance and minimizing the docking station visits to complete all scheduled services is a predominant requirement for effective drone operations in a hyperlocal market. This problem is formulated as a mixed-integer linear programming model, and a heuristic algorithm is proposed to attempt various practical size problems, with near-optimal solutions reported. The paper offers valuable insights for practitioners and future researchers wishing to analyze the performance of drone operations and determine the appropriate number of drones required for the hyperlocal market based on service demand.
Chapter
This paper presents the development of a hybrid approach as a solution to the multiple Traveling Salesman Problem (mTSP) applied to the route scheduling for self-drive cars. First, we use k-means to generate routes that equality distribute delivery locations among the cars. Then, these routes are set as the initial population for bio-inspired algorithms, such as Genetic Algorithm (GA) and Ant Colony System (ACS), that perform an evolutionary process in order to find a route which minimizes the overall distance while keeping the balance of individual tours of each car. The experiments were conducted with our route scheduling system in real and virtual environments. We compared our hybrid approaches using k-means in conjunction with GA and ACS against GA, ACS and Particle Swarm Optimization (PSO) initialized with random population. The results showed that, as the number of cars and target locations increase, the hybrid approaches outperform GA, ACS and PSO without any pre-processing.
Article
Compared with public transport operations, urban freight traffic and its associated delivery operations seem to be frequently overlooked in urban traffic management and traffic flow theory. One explanation for this is certainly the lack of available data, as the competitive freight transport market is fragmented and several actors are unwilling to collect or share tactical and operational data. In this study, we use the unique pNEUMA drone data set from Athens, Greece, to shed light on urban freight operations. We discuss macroscopic traffic indicators in a multimodal context. As the vehicle stopping behavior can adversely influence traffic flow, we reveal the stopping behavior of the different modes represented in the data set using clustering techniques. We find that urban freight vehicles’ stopping frequency lies between the stopping frequencies of cars and buses. We reveal the distribution of stopping times for loading and unloading stops in Athens to have a mean of around 380 seconds. Clustering all loading and unloading stops further reveals three groups of loading and unloading stops that could be labeled by incorporating knowledge and expertise about local particularities. The limited flight time of drones, owing to their battery capacities, did not allow reconstruction of longer vehicle routes, such as an entire vehicle tour within the network. However, this could be addressed in future research by realizing continuous large-scale monitoring routines. The revealed vehicle behavior parameters can be used in traffic models to generate further insights into the impacts of urban freight transport to inform public sector decision makers.
Article
The pickup and delivery problem with time windows (PDPTW) is an important problem to be solved. This is due to its resemblance to real-life services such as food delivery, courier service, curbside pickup service, and dial-a-ride, among others. However, the research on multi-objective pickup and delivery problems with time windows (MOPDPTW) remains scarce even after many years. The MOPDPTW objective is to achieve the optimal solution on both the number of vehicle and total travelled distance. It is challenging to obtain optimal results, the recurring optimal results that has least difference in magnitude and broader Pareto sets at the same time. Hence, an algorithm that can fulfil these requirements are highly sought after. This paper presents a mediocre evolutionary distributed microservices re-optimization algorithm (MEDMRA) to solve the MOPDPTW. The combination of the re-optimize algorithm and mediocre evolutionary algorithm escape the local optimal, optimized results and produce broader Pareto value. We compare our results using Li & Lim instances. Our results outperform the latest published hybrid algorithms and competitive with the best-known solutions.
Article
In this paper, we address the pickup and delivery problem with time windows (PDP-TW) and heterogenous vehicles for minimisation of total tardiness by learning heuristics from a given set of solutions. In order to extract scalable heuristics from optimal or best feasible solutions, we propose a machine-learning (ML)-based approach called ENSIGHT (Evolutionary Neural network with Scalable Information for Generation of Heuristics for Transportation). ENSIGHT consists of three phases: solution generation, interpretation of solutions, and improvement of heuristics by an evolutionary neural network (ENN). First, a set of optimal or best feasible solutions for the training set of problem instances is acquired by using the proposed mathematical model. Second, as for the process interpreting those solutions, an approach for transforming them into training data by way of scalable input attributes as well as output discretisation is followed. Third, the ENN improves the learned heuristics by an evolutionary parameter optimisation process for minimization of total tardiness. To verify the performance of the proposed ENSIGHT, we conducted experiments and the results of which showed that it outperforms other ML techniques and the current dispatching rules (DRs). Moreover, the approach was demonstrated to be effective in learning scalable heuristics based on combined scalable inputs and discretisation as well as an evolutionary improvement process.
Article
In this paper, several variants of multi-objective Selective Pickup and Delivery Problems with Time Windows are investigated. These problems have been widely addressed from a single-objective point of view to look for a solution with the most profitable set of requests while respecting a set of constraints. Handling simultaneously profit maximization and travel cost minimization poses a challenging optimization task in this class of routing problems. We propose a two-phase framework based on the decomposition of the search space in several linearly aggregated sub-problems. The aggregated problems are first optimized by an efficient local search with dedicated removal and insertion operators. An update is then applied on the weights of the least efficient sub-problems. We show that the perturbation of these weighted sum problems enables the exploration of more regions of the search space, and thus ensures the diversification of the Pareto front approximation. The obtained results on the selective variants of the Pickup and Delivery Problem show the effectiveness of our algorithm based on solutions quality and computational time. The proposed algorithm strictly improves 36 best known solutions of the single-objective problem and achieves the best results on all the instances of the lexicographic variant. Its performance is also confirmed on the bi-objective variant since we obtain better Pareto front approximation in terms of hyper volume, set cover, and computational time indicators. We discuss the results and explain the positive contribution of each component based on statistical tests.
Article
Since the late 70s, much research activity has taken place on the class of dynamic vehicle routing problems (DVRP), with the time period after year 2000 witnessing a real explosion in related papers. Our paper sheds more light into work in this area over more than 3 decades by developing a taxonomy of DVRP papers according to 11 criteria. These are (1) type of problem, (2) logistical context, (3) transportation mode, (4) objective function, (5) fleet size, (6) time constraints, (7) vehicle capacity constraints, (8) the ability to reject customers, (9) the nature of the dynamic element, (10) the nature of the stochasticity (if any), and (11) the solution method. We comment on technological vis-à-vis methodological advances for this class of problems and suggest directions for further research. The latter include alternative objective functions, vehicle speed as decision variable, more explicit linkages of methodology to technological advances and analysis of worst case or average case performance of heuristics. © 2015 Wiley Periodicals, Inc. NETWORKS, 2015
Article
This paper considers the vehicle routing problem with stochastic demands. The objective is to provide an overview of this problem, and to examine a variety of solution methodologies. The concepts and the main issues are reviewed along with some properties of optimal solutions. The existing stochastic mathematical programming formulations are presented and compared and a new formulation is proposed. A new solution framework for the problem using Markovian decision processes is then presented.
Article
The aim of this paper is to develop an exact algorithm for the asymmetrical distance-constrained vehicle routing problem. The problem is solved by means of a branch-and-bound tree in which subproblems are modified assignment problems subject to some restrictions. Computational results for problems involving up to 100 nodes are reported.Cet article décrit un algorithme exact pour le probléme de tournées avec contraintes de temps et une matrice de distance asymétrique. On résout le probléme au moyen d'un arbre de “branch and bound” dans lequel les sous-problémes sont des problémes d'affectation généralisée. On présente des résultats numériques pour des problèmes contenant justu'à 100 points de livraison.