Optimized planning of frequency hopping in cellular networks

We consider a generalization of the classical frequency assignment problem. The generalization arises when frequency hopping is used in a cellular network. The planning problem concerns assigning lists of frequencies to blocks of transceivers, such that the total interference is minimized. This problem is considerably more difficult than the classical frequency assignment problem, because of the large number of possible frequency lists. We provide the technical background that motivates our study, and present a mathematical model which includes the classical frequency assignment problem as a special case. We describe a simulated annealing algorithm. The algorithm explores the solution space by solving an integer program in each iteration. We report computational results for real-life and synthesized networks.