Application Of The Distributionally Robust Optimization In Containers Deployment

In this thesis,we mainly study a class of deployment problems considering the integrated laden containers routing and empty containers repositioning in the liner shipping industry.firstly,the problem and relevant parameters are introduced.We propose the deterministic model for integrated deployment problem of self-owned containers,laden containers and empty containers.Secondly,according to the uncertainty of some parameters in reality,random variables are introduced and the two-stage stochastic programming model is established.Then,we extend it to the distributionally robust optimization model.Furthermore,we transform it in to a solvable second-order cone programming model.Moreover,a tangent plane algorithm is introduced to solve the second-order conical programming problem.Finally,the data experiment is carried out.We take a liner shipping route as an example for numerical calculation and compare the deterministic model with the distributionally robust optimization model.It shows that the distributionally robust optimization model is more advantageous,which can help the industry reduce costs and form corporate competitiveness.