In Heterogeneous Wireless Networks, different overlapped Radio Access Technologies (RATs) can coexist with each other in the same geographical area. In such environment, a challenge is to select in which available RATs a user can be connected upon making an incoming service request. In this thesis, this challenge is investigated by proposing a Joint Call Admission Control (JCAC) -based approach that uses the framework of Semi-Markov Decision Process for initial RAT selection in two co-located wireless networks supporting two different service classes. The optimization problem involves the design of a cost function that weights the blocking cost and the energy consumption cost. The JCAC optimal policy is derived using the Value Iteration Algorithm. Simulations results show that the system capacity is maximized while selecting the less energy consuming RAT.