Today, most geodetic networks are being measured by differential Global Positioning System (GPS). Nevertheless, producing optimal baseline configuration and optimal baseline weights "Mixture-Order Design" in designing stage are necessary pre-requisites for an accurate geodetic GPS network. In this paper, a very efficient method based on the concept of the genetic algorithms metaheuristic technique is developed for providing an efficient solution for Mixture-Order Design problem. In the present study, the proposed optimization modeling is applied only to determine which of the baselines could be deleted while adhering to the prescribed precision of the network, which may result in significant savings on the cost of the project.