Optimal Path Planning and Cost Minimization for a Drone Delivery System Via Model Predictive Control

2025-03-25

AI Summary

This work addresses the joint optimization of path planning and operational cost in collaborative drone delivery. We propose a dynamic multi-scale optimization framework based on Model Predictive Control (MPC), marking the first systematic integration of MPC into multi-drone cooperative delivery tasks—offering superior dynamic environment modeling and real-time re-planning capabilities compared to existing reinforcement learning approaches. Evaluated on both high- and low-resolution grid simulations, our method consistently outperforms state-of-the-art multi-agent RL algorithms: reducing delivery cost by 12–18%, accelerating convergence by over 40%, and decreasing required drone count by more than 30%. The core contribution lies in establishing a scalable, interpretable, and real-time-responsive MPC-based cooperative optimization paradigm, providing an efficient and practically viable decision-making foundation for real-world low-altitude logistics systems.

Technology Category

• Search and Optimization: Sampling/Simulation-based Search
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Machine Learning: Optimization

Application Category

• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

In this study, we formulate the drone delivery problem as a control problem and solve it using Model Predictive Control. Two experiments are performed: The first is on a less challenging grid world environment with lower dimensionality, and the second is with a higher dimensionality and added complexity. The MPC method was benchmarked against three popular Multi-Agent Reinforcement Learning (MARL): Independent $Q$-Learning (IQL), Joint Action Learners (JAL), and Value-Decomposition Networks (VDN). It was shown that the MPC method solved the problem quicker and required fewer optimal numbers of drones to achieve a minimized cost and navigate the optimal path.

Problem

• Optimize drone delivery path planning using MPC
• Compare MPC with MARL methods for efficiency
• Minimize cost and drones needed for delivery

Innovation

• Model Predictive Control for drone path planning
• Benchmarked against Multi-Agent Reinforcement Learning
• Minimized cost and optimal drone deployment