融合算法改善后在物流配送中的应用

Application of Improved Fusion Algorithm in Logistics Distribution

论文对粒子群算法(PSO)进行改进,引入了非线性递减的惯性权重,比之线性权重,不易陷入局部最优,搜索精度更高。通过Griewank函数测试后比较发现,收敛速度虽优于PSO,但是在多峰值的算法中,依然容易陷入局部最优解,为此引入了模拟退火算法,跳出了局部最优解,在Griewank函数测试中表现为精度更高,明显优于PSO算法。在此基础上,使其应用在二级区域配送网络中,以上海市地图为例,来找到最优解。实验结果表明,该算法在此应用中表现良好,可以寻得最优解决方案,并且优于其他三种算法。此外,所提出的方法在其他研究领域有很大的应用潜力改进的。

In this paper,the particle swarm optimization(PSO)is improved,and the non-linear decreasing inertia weight is introduced.Compared with the linear weight,it is not easy to fall into the local optimum and the search accuracy is higher.After comparing the Griewank function test,it is found that although the convergence speed is better than PSO,it is still easy to fall into the local optimal solution in the multi-peak algorithm.For this reason,the simulated annealing algorithm is introduced,which jumps out of the local optimal solution,and is tested in the Griewank function.The medium performance is higher accuracy,which is significantly better than the PSO algorithm.On this basis,it is applied to the secondary regional distribution network,taking the map of Shanghai as an example,to find the optimal solution.Experimental results show that the algorithm performs well in this appli-cation,can find the optimal solution,and is better than the other three algorithms.In addition,the proposed method has great appli-cation potential for improvement in other research fields.