In this paper, the problem about power control in multiuser cognitive networks is addressed. The scenario we study is a heterogeneous network, where multiple cognitive (without the spectrum license) users transmit simultaneously to their corresponding receivers through a common frequency channel in the presence of a primary (with the spectrum license) user communicating with its base station. The goal of this paper is to study how to minimize the sum-power of the above cognitive network given certain QoS requirements of all cognitive users. The power optimization problem is formulated as a nonlinear and non-convex problem. We first apply the classical power control algorithm to find the pareto optimal efficient power for each cognitive user and formulate the achievable power region of the cognitive networks. Then, the optimization problem is converted into a convex one and Lagrange multiplier method is utilized to solve it. Our work provides a theoretical lower bound on the power contr...