Path planning of a mobile robot as a discrete optimization problem and adjustment of weight parameters in the objective function by reinforcement learning

In a previous paper, we proposed a solution to path planning of a mobile robot. In our approach, we formulated the problem as a discrete optimization problem at each time step. To solve the optimization problem, we used an objective function consisting of a goal term, a smoothness term and a collision term. This paper presents a theoretical method using reinforcement learning for adjusting weight parameters in the objective functions. However, the conventional Q-learning method cannot be applied to a non-Markov decision process. Thus, we applied Williams's learning algorithm, REINFORCE, to derive an updating rule for the weight parameters. This is a stochastic hill-climbing method to maximize a value function. We verified the updating rule by experiment.