Development of a Reinforcement Learning Inspired Monte Carlo Tree Search Design Optimization Algorithm for Fixed-Wing VTOL UAV Propellers

A novel Monte Carlo Tree Search Optimization Algorithm that is trained using a Reinforcement Learning approach is developed for the application to geometric design tasks. It is capable of evaluating design parameters and demonstrates the successful application of reinforcement learning strategies on a physics informed design optimization task. The algorithm is intended to be used for the parametric design of the optimal geometry of a propeller for Fixed-Wing VTOL UAV but is also applied to an aircraft design problem with ease.     «

A novel Monte Carlo Tree Search Optimization Algorithm that is trained using a Reinforcement Learning approach is developed for the application to geometric design tasks. It is capable of evaluating design parameters and demonstrates the successful application of reinforcement learning strategies on a physics informed design optimization task. The algorithm is intended to be used for the parametric design of the optimal geometry of a propeller for Fixed-Wing VTOL UAV but is also applied to an ai...     »