Abstract: An algorithm based on genetic algorithm for optimizing LSP route programming with the objective of load balancing is proposed, which introduces adaptive probabilities of crossover and mutation, and furthermore, makes use of heuristic information to improve search efficiency effectively. The simulation results show our algorithm is effective, in which it dramatically improves the unbalanced traffic distribution condition brought about by traditional SPF routing policy. In addition it has a superior performance over the standard genetic algorithm in aspects such as stability, optimization performance and efficiency.

Key words: