Optimize Path Planning for UAV COVID-19 Test Kits Delivery System by Hybrid Reinforcement Learning

Abstract

This paper aims to solve an optimization problem in the UAV-enabled COVID-19 test kits delivery system. The UAV intends to find the optimal path to deliver the COVID-19 test kits to people with a high probability of COVID-19 infection in the shortest time. The traditional Deep Reinforcement Learning doesn't perform well in solving the optimization problem because of the slow converging speed and difficult parameters-tuning. In order to solve this problem efficiently, a low-complexity Hybrid Reinforcement Learning is proposed. The algorithm consists of a heuristic algorithm and a Q Learning algorithm. At first, a heuristic algorithm is utilized to calculate the optimal path between any two users. Next, Q learning is applied to determine the sequence of the users to deliver the COVID-19 test kits. As a result, both the delivery sequence and the specific path from one user to another can be generated. The simulation results prove the superiority of the proposed Hybrid Reinforcement Learning in solving the proposed optimization problem compared with the state-of-arts.