A Quick Employment of Markov Decision Process (MDP) in Partially Unknown Three-dimensional Discrete Space

This project aims to develop a planning algorithm under partially unknown 3D discrete space. Value iteration is employed since MDP complies with the state transition and decision patterns here, with an adaption for changing transitional probabilities and rewards. Eventually, the robot conducts an optimal policy on a fully known map and reaches the destination in a partially unknown environment with less efficiency.