Abstract: To prolong the fatigue life of asphalt pavement，the fatigue performance of asphalt mixtures must be improved by changging the constituent struettu'ed mixture primarily．Inthispaper，the efect of constituent strueture on fatigue performance of asphalt mixture was analyzed． In succe．ssion，for the optimizing target as fatigue performance，the asphalt mixture were optimized combining BP nerve net with Genetic Algorithins．As a result，the relation between 4．75ram passing percent，O．075 real passing percent，filling—bitumen ratio and voids percent and fatigue performance of asphalt mixture，which can be considered in asphalt mixture design based on performances

Key words: road engineering, asphalt mixture, fatigue performance, optimization, genetic algorithms

摘要： 分析了沥青混合料组成结构对其疲劳性能的影响，采用BP神经网络和遗传算法相结合的方法，以疲劳性能为目标进行了沥青混合料性能优化设计，获得了4．75ram筛孔通过率、0．075ram筛孔通过率、粉胶比、空隙率与疲劳性能之间的关系，为更好地以性能为中心进行沥青混合料设计提供了参考．

关键词: 道路工程, 沥青混合料, 疲劳性能, 优化, 遗传算法