Robust Optimization Of Logistics Nodes Location Problem With Curved Demands

The scale of China's transportation network is expanding gradually,which is accompanied by huge maintenance demand.Due to the road damage which occurs frequently during the service life,the road maintenance can ensure the smooth flow and extend the service life of the road,so as to boost the national economy.Although road maintenance is gradually gaining attention,the maintenance cost is still high.One of the reasons is that the maintenance stations are not well located.At present,the location of maintenance stations is mostly based on experience and lacks support from effective theoretical.This paper takes road maintenance as the background with the considering of the demands uncertainty,which ensures that maintenance stations can withstand demand changes within a certain range,and avoid large cost increases to reduce cost risks.There are some certain characteristics with road maintenance demands.From the perspective of demand form,the demand is continuously distributed along the road from a long-term perspective,which is called curved demand.From the perspective of demand uncertainty,the demand usually depends on road conditions,temperature,and load capacity,and so on.Therefore the demand will change due to the changes in these factors,making future demands impossible to predict accurately.Considering the linearity and uncertainty of the demand,this paper uses the demand density function to describe the continuous demand,and the interval number is applied to represent the uncertainty.Within the given ellipsoidal uncertainty set,this paper aims to minimize the worst-case transportation cost.Finally,a robust optimization model for logistics nodes location problems with curved demand is constructed.Robust optimization models for single logistics node and multiple logistics nodes are constructed,and the robust counterpart models are calculated.Based on the analysis of the model properties,the conditions that the optimal solution should meet are determined,and an iterative algorithm is designed.In order to verify the validity of the model,case studies are performed.The comparative analysis with the deterministic model proves that the robust optimization model can bring greater stability at a small cost increase.And the robust optimization model can bring great cost savings and reduce cost risks when the demand greatly deviates from the predicted value.At the same time,it is found that the larger the robust budget,the greater the risk that the facility can withstand,which indicates that the trade-off between robust budget and cost needs to be made.The influence of the fluctuation radius on the location decision was also analyzed.It was found that the demand changes greatly,the robust optimization model can significantly reduce the cost risk.The thesis contains 12 figures,10 charts and 64 references.