A randomized variable neighborhood descent heuristic to solve the flying sidekick traveling salesman problem.

• Autor(a) principal: Freitas, Júlia Cária de
• Data de Publicação: 2018
• Outros Autores: Penna, Puca Huachi Vaz
• Tipo de documento: Artigo
• Idioma: por
• Título da fonte: Repositório Institucional da UFOP
• Texto Completo: http://www.repositorio.ufop.br/handle/123456789/10382; https://www.sciencedirect.com/science/article/pii/S1571065318300593
• topic: Unmanned aerial vehicle; Drone delivery; Last mile delivery

Unmanned aerial vehicles (UAV), or drones, have the potential to reduce cost and time in last mile deliveries. This paper presents the scenario which a drone works in collaboration with a delivery truck to distribute parcels. This Traveling Salesman Problem (TSP) variant has some particularities that make the originals constraints insufficient. In more detail must be considered the flying time-limit of the drone that inhibits them from visiting all customers and the parcel must not exceed the payload of the drone. To solve the problem, the initial solution is created from the optimal TSP solution obtained by the Concorde solver. Next, an implementation of the Randomized Variable Neighborhood Descent (RVND) heuristic is used as a local search to obtain the problem solution. To test the proposed heuristic, 11 instances based on the well-known TSP benchmark set were created. Computational experiments show the use of drones for last mile delivery can reduce the total delivery time up to almost 20%. Moreover providing a faster delivery this system has a positive environmental impact as it reduces the truck travel distance.