In this thesis, we consider an URN type frequency-hopped spread-spectrum multiple access communication system. We devise a near optimum access set control policy which changes the access set size adaptively as a function of the system state such that channel throughput can be maximized. Our access set control policy is shown to provide a significant performance improvement over the adaptive ALOHA scheme which changes the retransmission probability according to system state. We propose a system state estimation algorithm and analyze the system performance with the proposed estimation scheme.