针对装卸混合的车辆路径问题这一类典型的NP难题,采用四位数的遗传编码,并对解的可行性进行验证,降低对交叉算子和变异算子的要求,有效提高解的质量.最后对二十个客户点的装卸混合的问题作了数值试验,结果表明遗传算法作为一种有效的随机型全局搜索算法,体现出群体智能的分布型、鲁棒性和快速性的特点. A typical four-digit genetic code is used to solve the typical NP problem of handling and mixing vehicle routing problems. The feasibility of the solution is verified, the requirements of crossover operator and mutation operator are reduced, and the quality of solution is effectively improved. Finally, a numerical experiment on the handling and mixing of twenty customer points is carried out. The results show that genetic algorithm, as an effective random global search algorithm, reflects the distribution, robustness and rapidity of population intelligence.