Network selection is one of the most significant challenges for the next-generation wireless heterogenous networks. Many network selection schemes have been proposed, which can mainly be divided into two types: distributed schemes and hybrid schemes. Distributed schemes did not take into consideration the load unbalance problem in the whole heterogeneous system, which was not solved successfully in hybrid schemes either. In this paper, we propose a low complexity, centralized network selection scheme, aiming to optimally distribute the end users to the networks of the heterogeneous wireless system, in the sense of maximizing the global spectrum efficiency. Two suboptimal algorithms are proposed to solve this optimization problem, which are based on the measures of absolute bandwidth request (ABR) and relative bandwidth request (RBR) respectively. The overall bandwidth occupation results and the outage probabilities of the two algorithms are evaluated by means of numerical simulation.