A Variable Neighborhood Search for Flying Sidekick Traveling Salesman Problem (1804.03954v2)

Abstract: The efficiency and dynamism of Unmanned Aerial Vehicles (UAVs), or drones, present substantial application opportunities in several industries in the last years. Notably, the logistic companies gave close attention to these vehicles envisioning reduce delivery time and operational cost. A variant of the Traveling Salesman Problem (TSP) called Flying Sidekick Traveling Salesman Problem (FSTSP) was introduced involving drone-assisted parcel delivery. The drone is launched from the truck, proceeds to deliver parcels to a customer and then is recovered by the truck in a third location. While the drone travels through a trip, the truck delivers parcels to other customers as long as the drone has enough battery to hover waiting for the truck. This work proposes a hybrid heuristic that the initial solution is created from the optimal TSP solution reached by a TSP solver. Next, an implementation of the General Variable Neighborhood Search is used to obtain the delivery routes of truck and drone. Computational experiments show the potential of the algorithm to improve the delivery time significantly. Furthermore, we provide a new set of instances based on well-known TSPLIB instances.

• The paper presents a hybrid algorithm (HGVNS) that combines an exact MIP method with variable neighborhood search to solve the FSTSP.
• It achieves significant efficiency improvements, reducing delivery time by up to 67.79% and outperforming previous methods by up to 24.84%.
• The results underscore the practical benefits of integrating drones with trucks to enhance modern logistics systems.

Analyzing a Variable Neighborhood Search for the Flying Sidekick Traveling Salesman Problem

The development of drone technology has prompted significant interest in optimizing logistics networks, particularly through the integration of drones with traditional delivery vehicles like trucks. The paper "A Variable Neighborhood Search for Flying Sidekick Traveling Salesman Problem" presents an advanced approach to address this challenge by focusing on the Flying Sidekick Traveling Salesman Problem (FSTSP), a complex variant of the classical Traveling Salesman Problem (TSP).

Overview of the Flying Sidekick Traveling Salesman Problem

FSTSP involves assigning delivery tasks between a truck and a drone. The truck acts as a mobile base from which the drone can be dispatched to serve customers, allowing the truck to continue on its route. This hybrid delivery mode takes advantage of the high speed of drones and the large carrying capacity of trucks, presenting a promising solution to the burgeoning demand for rapid parcel delivery. The primary objectives in solving FSTSP include minimizing total delivery time and ensuring adherence to operational constraints such as drone endurance and payload capacities.

Methodology: Hybrid Variable Neighborhood Search

The authors introduce a hybrid heuristic algorithm named HGVNS (Hybrid General Variable Neighborhood Search) to solve FSTSP. The strength of this approach lies in its combination of an exact method to find initial solutions and the application of the General Variable Neighborhood Search (GVNS) for iterative improvement.

• Initial Solution Creation: The process begins with an exact method using a Mixed-Integer Programming (MIP) solver to generate an optimal TSP solution for a truck-only scenario.
• General Variable Neighborhood Search: The GVNS iteratively explores various neighborhood structures to diversify the search and escape local optima effectively. Various neighborhood operations, such as reinsertion and customer relocation, are explored to refine both truck and drone routes collaboratively.

Computational Experiments and Results

The proposed HGVNS was evaluated on several instances, including those provided by the literature and newly derived from the TSPLIB library, across different geographic configurations and drone speed scenarios.

• Performance Evaluation: The algorithm demonstrated significant improvements in total delivery time compared to traditional truck-only routes, with reported reductions up to 67.79%. When compared to existing methods, HGVNS consistently found superior solutions, thus establishing new best-known solutions (BKS) for all evaluated FSTSP instances.
• Improvements Against Baselines: Notably, the improvement was up to 24.84% over previously reported solutions in instances from the literature. The results underscore the effectiveness of integrating drones into logistical routes and the efficacy of the HGVNS approach.

Implications and Future Directions

The practical implications of this research are extensive, advocating for a transformative shift in parcel delivery systems by using drones for improved efficiency. The theoretical advancement in solving NP-hard problems like FSTSP through advanced metaheuristic approaches like HGVNS illustrates a robust methodology adaptable to diverse operational constraints.

Future research could expand on the framework to incorporate multiple drones and trucks, capacitated vehicles, and real-time adaptability, further aligning the model with complex real-world logistics scenarios. Additionally, exploring the integration of Machine Learning for predictive logistics and dynamic routing could augment the performance of such systems under variable customer demands and stochastic travel conditions.

In conclusion, the paper provides a compelling solution to the FSTSP while highlighting the synergy between trucks and drones in logistics. The HGVNS algorithm presents promising avenues both for theoretical exploration and practical application in the modern logistics landscape.