Abstract: Many real-life networks such as logistics and communication network and so on can be regard as a stochastic-flow network.A multi-objectives optimization problem in a stochastic-flow network is studied in this paper.The objectives are the probability of transmission is successful and the cost that transmission costs.The authors build a model using the conception of MPs，so greatly reduce the complexity of model.A multi-objectives genetic algorithm for computing the optimal flow allocation policy is proposed.Tested by an example，this algorithm commendably solves the problem of flow control and allocation in a stochastic-flow network.

Key words: stochastic-flow network, multi-objectives genetic alogrithm, integer programming

摘要： 现实世界的网络比如：物流网络、通信网络、交通网络，电网等可以被抽象成一个随机流量网络。以传输成功率和整个传输所花费的成本为目标，对随机流量网络上流量的分配控制的多目标优化问题进行了研究。采用MPs的概念对问题建模，大大简化了模型的复杂程度。最后提出一个多目标遗传算法，通过实例验证，该算法较好地解决了随机流量网络上的流量分配控制问题。

关键词: 随机流量网络, 多目标遗传算法, 整数规划