goal-oriented reinforcement learning, where the agent needs to reach the goal
state while simultaneously minimizing the cost, has received significant
attention in real-world applications. Its theoretical formulation, stochastic
shortest path (SSP), has been intensively researched in t