Abstract:

In order to utilize the available frequencies efficiently in cellular mobile communication network, a frequency assignment method based on microcanonical annealing algorithm was proposed. In this approach, the optimal frequency deployment was achieved by random walks of demon in the state space when the potential interference was defined as an energy function. Simulations show that if available frequencies are sufficient, the proposed algorithm can achieve equivalent results compared with simulated annealing algorithm. However when available frequencies are reduced, this new heuristic algorithm offers better results with higher probability to hit the global optimum at the cost of less function evaluations. In addition, its fast convergence is also demonstrated by typical optimization tracks.

Key words: microcanonical annealing, simulated annealing, frequency assignment, combinatorial optimization