Traditional mobile robots commonly utilize the Q-learning algorithm in reinforcement learning for path planning. This paper proposes an reinforcement learning particle swarm algorithm(RLPSO) that combines the particle swarm algorithm and the Q-learning algorithm to address the convergence problems of the traditional Q-learning algorithm in path planning. These problems include slow speed, long planning path, and low search efficiency. The proposed approach first explores the unknown environment through Q-learning to create a Q matrix, which is then used to calculate the fitness value of each particle. Subsequently, the environment is further explored using the particle swarm algorithm, and the optimal hyperparameters in the Q-learning algorithm are determined based on a predefined multi-objective planning function. Finally, the RLPSO algorithm is simulated and calculated using MATLAB software. Experimental results demonstrate that the RLPSO algorithm accelerates convergence, reduces the planning path length, and enhances search efficiency, enabling the mobile robot to find the optimal path efficiently.
