A Survey On Enhancing Reinforcement Learning in Complex Environments: Insights from Human and LLM Feedback