基于蚁群禁忌混合算法的成品油多舱配送路径优化研究

依据油品运输策略的不同,成品油二次配送可划分为两种模式:同一加油站的不同油品需求可拆分配送和不可拆分配送.在两种模式的基础上,根据带时间窗的成品油多舱配送基本模型,衍生出两类配送子模型;综合考虑蚁群算法较强的全局搜索能力和禁忌搜索算法的局部搜索能力,设计蚁群禁忌混合算法(ACO-TS),并提出相应策略用于两类子模型的求解.采用12组不同类型的算例进行数值实验,实验结果表明,混合算法能有效的求解两类配送子模型,并且针对第二类子模型设计的特有邻域能够加快算法求解速度;此外,两种配送模式中,同一加油站不同油品需求可拆分的模式在降低配送成本的同时,能够大幅提高车载率,减少车辆使用量。

Route optimization for the refined oil multi-compartment distribution based on ant colony and tabu search hybrid algorithm

With different transport strategies, the refined oil secondary distribution could be divided into two modes:one claims that the different products required by a gas station must be entirely delivered by one single vehicle, and the other is permitted to use several vehicles. We formulate each mode as an integer programming model based on a multi compartment refined oil distribution model with the time window. An ant colony and tabu search hybrid algorithm (ACO-TS) with different solution strategies is proposed to solve those two models. Computational experiments are performed on 12 instances. The results show that the hybrid algorithm is highly effective and special neighborhood structures for the second model are advantageous in increasing solving speed. In addition, the second mode that different products required by a gas station may be brought by several vehicles could reduce the distribution cost and significantly increase the load rate with less vehicles.