Network selection plays an important role in load balancing and radio resources management. In this paper, we use the evolutionary game theory to address the network selection problem in the constrained heterogeneous networks. An indication function is introduced into the payoff function to represent the punishment term.

The population dynamics can be obtained by the replicator equations. A centralized population evolution algorithm is proposed to implement the evolutionary game. At last, the validity of our analysis and the proposed algorithm is verified by numerical results and Monte Carlo simulations.