Abstract
In this paper, we use the stochastic control technique Markov Decision Process (MDP) to study and examine the relationship between optimal decision, which the service provider should apply, and traffic parameters in wireless heterogeneous networks. The MDP model is formulated as a linear programming problem with objectives of maximizing the system utilization while ensuring that mobile users are experiencing low dropping rates which is translated in satisfying their Quality of Set-vice (QoS) requirements. Simulation results show the strength of our contribution.