LEARNING THE COMPATIBILITY MATRIX FOR ADAPTIVE RESOURCE-MANAGEMENT INCELLULAR RADIO NETWORKS

To satisfy the growing demand for mobile communications micro-cellular systems are expected to be introduced in the near future. For these, effective procedures not only for channel allocation but also for load balancing and call admission control still have to be developed. Here , the knowledge of the actual interference relations between radio cel ls is a key issue for both the more traditional channelized systems an d the recently proposed spread spectrum techniques. Realizing that the interference relations among the cells are only slowly varying, the a uthors generalize the existing local and shea term approaches to inter ference detection by including a memory trace in the evaluation proced ure. A compatibility matrix for the cells is learned by statistically correlating carrier and interference power levels with call activities . In this paper, the overall system design is presented along with the detailed description of the learning procedure which is thought of as both an integrated component of future radio network management syste ms as well as a possible enhancement of already existing radio network planning methods. Simulations show that the adaptive strategy offers a considerable capacity increase that results from an increased spectr um efficiency.