Abstract: According to the mathematical model of Voyage scheme optimization, an improved ant colony algorithm is presented. Three strategies to improve the algorithm, it first uses selection of two-stage method to reduce the computational complexity, then uses small group of optimal search method to initialize the pheromone intensity, finally designs a kind of dynamic interval mutation operator to avoid getting into the local optimal. Different sizes of numerical experiments show that the algorithm has good computational performance.

Key words: voyage scheme, ant colony algorithm, integer programming, mutation operator

摘要： 针对航次方案优选问题的数学模型，提出一种改进的蚁群算法。算法采取了三种改进策略，用两阶段比选降低计算复杂度，利用小群最优试探初始化信息素强度，设计一种动态间隔变异算子避免陷入局部最优。不同规模的数值实验表明，该算法具有良好的计算性能。

关键词: 航次方案, 蚁群算法, 整数规划, 变异算子