209 Copyright © 2016, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited. Chapter 8 DOI: 10.4018/978-1-4666-9644-0.ch008 ABSTRACT This chapter addresses the family of problems known in the literature as Capacitated Vehicle Routing Problems (CVRP). A procedure is introduced for the optimization of a version of the generic CVRP. It generates feasible clusters and, in a first step, yields a coding of their ordering. The next stage provides this information to a genetic algorithm for its optimization. A selective pressure process is added in order to improve the selection and subsistence of the best candidates. This arrangement allows improving the performance of the algorithm. We test it using Van Breedam and Taillard’s problems, yielding similar results as other algorithms in the literature. Besides, we test the algorithm on real-world problems, cor- responding to an Argentinean company distributing fresh fruit. Four instances, with 50, 100, 150 and 200 clients were examined, giving better results than the current plans of the company. INTRODUCTION The competitive pressures faced by companies pushes for the development of new techniques for opti- mal decision-making. These methods are also intended to be fast enough to achieve highly competitive services. One of the main issues that such techniques should address is the design of plans for the dis- tribution of products. The high complexity of these problems constitutes a roadblock for the creation of A Genetic Algorithm’s Approach to the Optimization of Capacitated Vehicle Routing Problems Mariano Frutos Universidad Nacional del Sur, Argentina & CONICET, Argentina Fernando Tohmé Universidad Nacional del Sur, Argentina & CONICET, Argentina Fabio Miguel Universidad Nacional de Río Negro, Argentina