Environmental protection of rural ecotourism using PSR and MDP models

The relationship between the tourism economy and the ecological environment has always been a hot issue in tourism research today. Because increasing tourism in rural regions frequently endangers the environment and rural culture, this study uses the Markov Decision Process (MDP) to determine the best policy balance between environmental preservation and tourism infrastructure. First, this research describes the methods used to incorporate a decision-making framework based on MDP into rural ecotourism, with a particular emphasis on the Zibo region of China. To navigate this complex game interplay, we proposed a Markov Decision Process MDP model bearing three pivotal states for this problem. These states are the Economic State, Ecological State, and Environment State. In this model, we also identified three distinct actions. Each action carries its significance. These actions include Promoting Tourism, Environment Protection, and Infrastructural Investment. We relied on data spanning a decade in Zibo of China to calculate the likelihood and real-world dynamic probability. The optimum policy is driven after the MDP simulation. Second, to validate the optimum policy, the policies were also iterated. In the last phase, the Sequential Markov Decision Process is simulated. This simulation not only shows the current state and policy for the area but also predicts the future strategies and policies for Tourism Economy and Rural Environment protection. The experimental results show that the model achieved the best policy and strategy. The experimental results show that the model successfully achieved the best policy and strategy. After 10,000 iterations, the score of γ = 0.9 is achieved, and the iteration of policies provides a better insight into the chosen strategy. The results are optimistic, with our method achieving an accuracy of 85%, recall of 88%, precision of 82%, F1-Score of 85%, sensitivity of 88%, and consistency of 90%. Our approach outshines the competition when assessing performance in specific geographic areas, achieving the highest accuracy in Area A (92%) and Area C (94%). These findings demonstrate the adaptability and the effectiveness of our approach in handling various ecological contexts.