Improved Differential Evolution Algorithms for Solving Mixed-integer Nonlinear Programming

摘要

A new method called hybrid line-up differential evolution algorithm is provided for the difficult problem of mixed-integer nonlinear programming (MINLP) with multi-modal and non-linear object functions in the paper. At the same time, two novelty strategies are adopted in order to solve the MINLP conveniently. One is to introduce the adaptive penalty parameter strategy; the other is to introduce a mixed coding ways and a rounding operation so that the hybrid line-up differential evolution algorithm can solve the mixed-integer nonlinear optimization problems. The superiority of the algorithm is tested through cases and the simulated result shows high efficiency, rapid speed of convergence and strong capability of global search in searching solution.