Modern logistic solutions for pickup and delivery problems with alternative and roaming locations, intermodal transportation, and inventory constraints and transfers

The thesis is comprised of topics dealing with variations of the pickup and delivery problem (PDP), a subfield of commercial transportation problems. The research on PDPs has been advanced thematically and methodologically by studying new topics and developing extensions of existing algorithms and new solution methods. In particular, this thesis considers transportation networks operating in multiple regions that are linked by a long-haul connection. An extensive literature review for the multi-region multi-depot pickup and delivery problem (MR-MDPDP) and its preceding problems is provided as well as more detailed work on a three-part network structure to solve a PDP with long-hauls without direct shipments between regions while fulfilling capacity and time window constraints. The insights gained from this problem show that long-haul flexibility influences short-haul routing cost, such that large improvements are possible merely by increasing long-haul flexibility but not long-haul cost. This work was also extended as a basis to train a regression model to enable the fast estimation of cost for the MR-MDPDP. The thesis also examines the pickup and delivery problem with alternative locations and overlapping time windows in one region. Each request may have multiple roaming pickup locations throughout the day with non-overlapping time windows. Requests may also have multiple roaming delivery locations and additionally an alternative recipient with its own set of roaming locations. The assumption is that it is no longer necessary to deliver to a single specific person. Multiple persons in different locations can be available simultaneously to accept a delivery. Additionally, recipients of parcels can use 24-hour locker boxes supplied by the logistics provider if they are located near their home. In particular, we explore the benefits of locker boxes, roaming locations, alternative recipients, and mixed customer profiles with different preferences concerning data sharing and convenience. We find that an increase in convenience for the customers and flexibility translates into large cost savings for the carriers. A rich variation of the pickup and delivery problem was solved as part of the VeRoLog Solver Challenge 2016–2017. This problem required the redistribution of expensive and, therefore, scarce equipment over customer locations within time windows. The difficulty was in creating combinations of pickups and deliveries that reduce the amount of equipment needed to execute the schedule, as well as the lengths of the routes and the number of vehicles used. The developed algorithm was awarded third place in the challenge. Methodologically, this thesis applied heuristics, metaheuristics, matheuristics, and decomposition approaches. Some established methods (adaptive large neighborhood search, variable neighborhood decent, genetic algorithms) are applied successfully, while others (extension of the pilot method, variations of the savings algorithm and the insertion algorithm) are adapted and extended for a better applicability on the problems studied. The quality of the solution methods is ensured by comparing with suitable previous results from the literature. Managerial insights are gained by solving realistic and real-life scenarios.