Research on refined classification method of surrounding rock based on game theory combination weighting - fuzzy cloud model

The surrounding rock classification system suffers from wide variations in surrounding rock quality across grades and significant disparities within the same grade, making it difficult to accurately assess surrounding rock quality and implement differentiated management in engineering practice. Accordingly, this paper summarises commonly used domestic surrounding rock classification methods and integrates five evaluation indicators to establish a refined surrounding rock classification index system. Subsequently, the membership degrees of the evaluation indicators were calculated using a cloud model, and a game-theoretic combination weighting method was applied to assign both objective and subjective weights to the evaluation indicators. Ultimately, surrounding rock grades were determined using a fuzzy comprehensive evaluation method, thereby establishing a surrounding rock grading assessment model based on a game theory–based combined weighting–fuzzy cloud model. In a preliminary proof-of-concept on a small-scale dataset of 15 samples, this model achieved a classification accuracy of 93%; subsequent studies on larger datasets wil be conducted to verify its effectiveness and generalizability.