A new method for global optimization

This paper presents a new method for global optimization. We use exact quadratic regularization the transformation of multimodal problems to problem maximum norm vector on convex set. Quadratic often allows you convert into unimodal problem. For this, we shift feasible region along bisector positive orthant. only local search (primal-dual interior point method) and dichotomy extremum in problems. The comparative numerical experiments have shown that this is very efficient promising.