Abstract: Vehicle scheduling problem is a constrained combinatorial optimization problem，the traditional optimization algorithm has a defect of premature convergence.Aiming at this issue，this paper improves the standard genetic algorithm in terms of chromosome encoding，the adaptive operator mechanisms and dealing with constraints.Test results show that the method is effective in improving the computational convergence and overcoming the premature phenomena.

摘要： 车辆调度优化问题是一个有约束的组合优化问题，属于NP难题（Nondeterministic Polynomial Problem）。随着问题输入规模的扩大，求解时间呈几何级数上升，传统的优化算法本身存在着过早收敛于局部值的问题。针对这一问题在染色体编码、算子的自适应机制和约束的处理等方面对标准遗传算法进行了改进。测试结果表明，该算法提高了优化算法的质量和搜索效率，具有良好的效果。