Research On A Distributionally Robust Approach For Dynamic Containers Deployment

In the past 20 years,under the impetus of reform and opening-up policy,China's national economy and foreign trade have developed rapidly.Containers are used in multi-commodity trades more and more extensively as they are convenient,fast,strong and pretty good in multimode transport.But the imbalance of trade flows between countries and regions make it a big challenge for shipping companies to reposition the empty containers.Which led to the import-oriented port has accumulated a lot of empty containers,and export-oriented port urgent need for a large number of empty containers.The deployment of empty containers is a huge operating cost for shipping companies.Therefore,to solve the problem of empty container deployment,can help shipping companies to reduce costs and improve the competitiveness of enterprises.In this thesis,we use the distributional ly robust optimization to study the empty container replacement problem,and considering the laden containers,the innovative introduction of the foldable container.The standard containers discussed in this thesis are 20 feet,and the foldable containers are identical in size to the standard container when they are opened,and the volume is one quarter of the standard container when folded.Foldable containers not only save storage space,but also can save space,which means that will greatly reduce the cost of inventory and increase the ship's container loading.Foldable containers can also improve handling efficiency.Firstly,this thesis establishes the deterministic model of container deployment problem.Secondly,based on the deterministic model,we establish a two-stage stochastic programming model and a distributionally robust model when the supply and demand of containers are uncertain,and transform the distributionally robust model into a solvable second-order cone programming model.Finally,we use the rolling horizon simulation methods to compare the three model,we can find that the distributionally robust model is more advantageous in solving the problem of containers deployment,which can reduce the cost for the enterprise.