A similarity degree-based three-way decision method with application to mine potential evaluation

Many existing interval-valued three-way decision methods primarily rely on crisp similarity assessments. Besides, they usually focus on single-objective modeling based on utility maximization or loss minimization, which may lead to conservative and biased outcomes. To address these issues, this paper develops a similarity degree-based interval-valued three-way decision method that integrates utilities and losses within a unified framework. Specifically, we define similarity degrees between interval values by a distance function, and further derive similarity-driven conditional probabilities, relative utility and loss functions. On this basis, we compute expected utilities and expected losses. Furthermore, we combine them via a risk-preference coefficient to obtain the expected values of alternatives, which can be used to reasonably categorize and rank alternatives. Moreover, the proposed method is applied in mine potential evaluation, and the evaluation results exhibit a high level of agreement with practical exploitation outcomes. What’s more, comparative analyses and sensitivity experiments further confirm that the method achieves more accurate and robust decision performance than some representative methods.