Yagmur, Ece2025-03-222025-03-2220250957-41741873-6793https://doi.org/10.1016/j.eswa.2025.126617https://hdl.handle.net/20.500.13091/9922This paper presents the joint vehicle and courier routing problem in last mile delivery using parcel lockers, which serve as both transshipment stations for crowd shipping and pick up locations for customers to self-pick up their orders. In the problem under study, any customer prefers to be a courier in return for a certain amount of compensation and to serve other customers along with himself. The objective of the problem is to minimize the sum of total travel costs arising from vehicles, total tardiness costs, and total compensation costs paid for couriers. A novel Mixed Integer Programming (MIP) model is developed, and comprehensive computational experiments are conducted on newly generated test instances. According to the results, CPLEX provides acceptable solutions only for small-sized instances, as it is rather sensitive to the number of customers and lockers in terms of solution time and quality. It is also observed that the performance of CPLEX deteriorates dramatically as the problem size increases, and it suffers to find any upper bound even for medium-sized instances within 3 h. So, we offer two metaheuristic algorithms for practical-sized cases: the Memetic Algorithm (MA) and Simulated Annealing (SA). Computational experiments demonstrate that both metaheuristics reach optimal solutions in a reasonable amount of time in all instances whose optimality is known. For medium-sized instances, although objective function values are close to each other, in 643 out of 864 medium-sized instances, MA provides superior results as compared to SA. Finally, for large-sized instances, both algorithms are run for both 180 and 360 s. According to computational experiments, the performance of SA in 180 and 360 s is found to be almost the same, as it shows fast convergence. While SA is more successful than MA within 180 s, both algorithms show similar performances in terms of the average objective function value for large-sized instances in 360 s.eninfo:eu-repo/semantics/closedAccessCourier RoutingCrowd ShippingLast Mile LogisticsParcel LockerVehicle RoutingArticle10.1016/j.eswa.2025.1266172-s2.0-85216840972