The PhD work is motivated by fact that mixed-integer nonlinear optimization is a highly versatile tool for modeling application problems in many areas. However, in some applications it is not possible to compute or approximate the derivatives. A promising approach for solving nonsmooth MINLP problems combines a heuristic with stochastic techniques. In this work, we intend to explore and extend the Firefly stochastic algorithm to solve nonsmooth nonconvex MINLP problems with equality and inequality constraints. |