Study on routing optimization method of logistics distribution based on SOM

Rational delivery route can improve the efficiency of logistics distribution, SOM neural network can keep the topological characteristics between the neurons, so SOM is used to solve the shortest delivery route of logistics. One delivery node is selected as input vector of the network randomly and the out put nodes of the network is 2 times than the number of delivery nodes. When training, the network will identify the weight vector closest to the delivery node and move it and its neighbors closer to the delivery node. After training, every delivery node has the closest node respectively. The steps using SOM to find the shortest route is brought out. Test our method with different delivery nodes, results show that SOM algorithm can obtain faster convergence rate and get the better optimal solution.