However, the traditional Q-Learning algorithm lacks the iteration stability and computational efficiency required in high-dynamic scenarios, and the shortest paths found often fail to meet the ...