Solving a nonsmooth and nonconvex minimization problem can be approached as finding a zero of a set-valued operator. With this perspective, we propose a novel Majorizer–Minimizer technique to find a local minimizer of a nonsmooth and nonconvex function and establish its convergence. Our approach leverages Bregman distances to generalize the classical quadratic regularization. By doing so, we generate a family of regularized problems that encompasses quadratic regularization as a special case. To further demonstrate the effectiveness of our method, we apply it on a lasso regression model, showcasing its performance.

In this paper we establish necessary as well assufficient conditions for a given feasible point to be a globalminimizer of smooth minimization problems with mixed variables.These problems, for instance, cover box constrained smooth minimizationproblems and bivalent optimization problems. In particular, ourresults provide necessary global optimality conditions for differenceconvex minimization problems, whereas our sufficient conditionsgive easily verifiable conditions for global optimality of variousclasses of nonconvex minimization problems, including the class ofdifference of convex and quadratic minimization problems. Wediscuss numerical examples to illustrate the optimalityconditions