We propose a genetic algorithm optimization technique (GAOT) designed to obtain the lattice geometry parameters and the electronic structure of pi-conjugated polymer systems. Particularly, we obtain the ground state structure of a completely dimerized trans-polyacetylene chain (CH)(x) using this approach. We also use this approach to determine the polaron and soliton solutions. These solutions are then compared to the traditional self-consistent-field (SCF) solutions. We conclude that the optimization technique proposed here generates solutions as those obtained using traditional (SCF) techniques, with the important advantage of reducing significantly the number of iterations needed to reach convergence. Besides, the convergence of the GAOT technique is more robust, reaching solutions independently of initial guesses. (C) 2012 Elsevier B. V. All rights reserved.