Abstract
We introduce a novel method called subdividing labeling genetic algorithm (SLGA) to solve optimization problems involving n - dimensional continuous nonlinear functions. SLGA is based on the mutation and crossover operators of genetic algorithms, which are applied on a subdivided search space where an integer label is defined on a polytope built on the n - dimensional space. The SLGA method approaches a global optimal solution by reducing the feasible search region in each iteration. One of its main advantages is that it does not require computing the derivatives of the objective function to guarantee convergence. We apply the SLGA method to solve optimization problems involving complex combinatorial and large-scale systems and illustrate numerically how it outperforms several other competing algorithms such as Differential Evolution even when considering problems with a large number of elements.