A Hybrid Learning Agent Approach for Solving the Flight Trajectory Optimization

Kurzfassung

It is believed that combinatorial optimization is to be among the first applications where quantum computers can demonstrate a practical advantage over classical systems. One such problem is the flight trajectory optimization, which aims to find the optimal path between two airports for an airplane. Due to non-local constraints of the flight path, the problem is part of complexity class NP-hard and becomes difficult to solve for classical algorithms. To address this, we employ a Reinforcement Learning algorithm in which the learning process of an agent is sped up by using a quantum communication channel.