基于人工鱼群算法的共享汽车停放网点优化算法的研究

Research on optimization algorithm of shared car parking network based on artificial fish swarm algorithm

针对共享汽车运营中车辆与客户匹配率低的问题,文中对汽车停放网点的选址位置及网点建设规模进行了优化设计。通过引入差分进化算法,对传统的人工鱼群算法进行改进,将人工鱼群划分为开发集与探索集。开发集与探索集间使用不同的步长和视野参数,实现了鱼群差异化的并行进化,提升了算法对于最优值的搜索能力。通过引入判定概率,改变原有的聚群、追尾方式,实现开发集与共享集间信息的高效互通,降低算法的运算复杂度。最终基于实际共享汽车的运营数据,测试网点的优化方案。结果表明,在有限的网点建设成本下,该算法可以改善67%的供求差。此外,改进算法的平均运行时间为56.25 s,而传统算法的运行时间为89.35 s,运行效率显著提升。

In view of the low matching rate between vehicles and customers in the operation of shared vehicles,this paper optimizes the location and construction scale of parking outlets.By introducing the differential evolution algorithm,the traditional artificial fish swarm algorithm is improved.The artificial fish swarm is divided into development set and exploration set.Different step length and field parameters are used between development set and exploration set to realize the parallel evolution of fish swarm differentiation and improve the search ability of the algorithm to the optimal value.By introducing the decision probability,the original clustering and tailgating methods are changed,and the efficient information exchange between the development set and the shared set is realized,which reduces the computational complexity of the algorithm.Finally,based on the actual operation data of the shared car,the optimization scheme of the network is tested.The results show that the algorithm can improve the supply⁃demand gap by 67%under the limited network construction cost.In addition,the average running time of the improved algorithm is 56.25 s,while the running time of the traditional algorithm is 89.35 s,the running efficiency is significantly improved.