An algorithm for solving inverse problems characterized by multi-extremal functions

This article addresses the solution of a single-point inverse problem for multi-extremal functions in the context of target indicator formation tasks. A description of the proposed algorithm is provided, which was developed based on an inverse approach and coordinate descent with stochastic argument selection. The algorithm was tested using well-known benchmark functions, including comparisons with methods implemented in existing Python libraries. The application of the algorithm for solving the practical problem of pricing under oscillatory demand is examined. The experimental results demonstrate the feasibility of applying the algorithm to solve problems involving multi-extremal functions.