| dc.description.abstract | Supply chains start with a customer need/request and encompass all the components involved in fulfilling this request, including the suppliers, manufacturers, transporters, warehouses, retailers, and lastly, the end customers. Their main purpose is to satisfy that initial need while making the process as efficient as possible. The objective of the supply chain is to maximize the overall value produced, which is often correlated with the supply chain profitability. In the profitability (and success) of an organization, operational decisions are key. The environment, technology, and customer needs/expectations are continually changing. Consequently, it is imperative that companies adapt and restructure their supply chain accordingly to remain competitive. This work challenges traditional truck-based last mile operations by proposing a more sustainable approach to home delivery. Two different problems are addressed using a “three paper” dissertation structure.
The first paper considers the two-echelon, multi-trip, capacitated vehicle routing problem with home-delivery and self-pickup services using drones and electric-assisted bikes. In this proposed network, parcels are transported from a depot to parcel lockers via drones and are then delivered to customer locations via e-bikes. A mathematical programming formulation that seeks to determine the vehicle routes that minimize the total cost (drone/e-bike operational cost and operator wage) is proposed. The mathematical model is coded in Python to obtain the optimal solution of each instance. Performance is assessed based on the metrics of total cost (measured in dollars) and emissions (measured in grams of CO2). An analysis of variance (ANOVA) is conducted to assess the effects of three factors on total cost; these factors are transportation approach, self-pickup customer percentage (of total customers), and depot location. Experimental results suggest that with a modest increase in total cost (as little as 13%), emission reductions of up to 92%, on average, can be achieved when using the greenest delivery strategy of drones and e-bikes.
The second paper considers the technician routing and scheduling problem with drone resupply and stochastic demands. In general, repair services usually involve service technicians carrying a repair kit in their truck with spare parts and performing repair jobs; however, technicians often need to return to the depot to collect spare parts that they don’t have in the truck. In this work, drones can resupply the trucks at the customer locations. Demands for spare parts are revealed over the course of the day; that is, after each technician’s first visit to each location. At this time, the demand of that customer is immediately revealed/known. Spare parts can be transported by either truck or drone, subject to capacity constraints. The problem seeks to balance the tool and part supply inventory on the vehicles (trucks and drones) and at the depot. The problem is optimized based on the information revealed at the moment of need and there is no anticipation of future requests. Moreover, all customer locations (but not their demands) are known at the start of the day. A mathematical model is proposed. In the model, the sequence of customers that a repair truck will follow is already determined and given. The model seeks to balance truck, drone, and depot inventories. Moreover, we are interested in determining an efficient and environmental friendly network with the incorporation of faster and low-emission vehicles (drones). Results suggest that incorporating a drone to the network can improve the makespan by over 30%.
The third paper proposes a network where a mobile parcel locker (MPL) transports parcels for four different customer types (attended home delivery, unattended home delivery, self-pickup, and roaming) from the depot to stopovers. Self-pickup customers collect their parcel at the parked MPL, while the rest of the customers receive their parcels through a heterogeneous fleet of vehicles (drone and electric-assisted bike) integrated into the MPL. We develop a mixed integer linear programming formulation that seeks to find the route/schedule of MPL and delivery vehicles, MPL location, and the customer-to-stopover assignments that minimize the total time. The customer type distribution more suitable for the proposed network is investigated. Furthermore, the effects of customer type, adding a second integrated delivery vehicle to the MPL, and delivery vehicle capacity are studied. Experiments reveal that the bike is the bottleneck of the system and that having more roaming customers than attended home delivery customers may reduce the average bike total travel time by 9.5%. | en_US |
