This paper presents a multiple-objective routing optimization scheme based on genetic algorithm (GA) methodology to the metric tradeoff optimization problem between two routing metric cost (delay and bandwidth) in overlay network. Besides an effective tradeoff routing feature, the scheme applies genetic- crossover to the local-search operations in the optimized iterative course for better performance. Simulation experiments demonstrate, compared with non-GA algorithm, the scheme achieves mainly about 2.17%~11.76% optimal gains.