Efficient algorithm for finding a fuzzy Sugeno measure for solving applied problems

The Sugeno parametric λ-measure is widely used in various management decision support problems, in particular, in multi-criteria decision making to eliminate criterion uncertainty. An effective algorithm for searching the λ parameter is presented, based on rigorously proven monotonicity properties of the Sugeno λ-measure. The algorithm includes: a logarithmic transformation of the objective function to ensure computational stability at extreme values of the arguments; a procedure for automatically expanding the search interval; handling of symmetric search bounds and a punctured point; a combined stopping criterion based on the argument and the function value. Numerical experiments demonstrate the convergence of the proposed algorithm in all considered test cases, including highdimensional problems ( n = 1000), where traditional methods, such as Newton’s method, exhibit divergence. The proposed algorithm can be used in decision support systems and applied tasks that require multiple computations of the λ-measure.