Emergency Commodities Supply Network Design Based On A Joint Chance Constraint Programming

The occurrence of natural disasters has always been accompanied by the development of society.In recent decades,the number of disasters that have occurred in the world has gradually increased,which seriously threaten the safety of human beings and the development of ecological environment,and cause huge losses to the society and economy.After a disaster occurs,it is necessary to quickly provide a large amount of emergency commodities to affected people for timely relief and assistance.Therefore,it is very significant to design an effective emergency commodities supply network in pre-disaster phase for responding these disasters.This paper proposes an emergency commodities supply network model considering multisourcing supply strategy,which simultaneously determine decisions of relief facilities location,emergency commodity pre-position in open relief facilities,and service linkages allocation between affected areas and open relief facilities.Taking into account the uncertainty of demand information in affected areas,our model only need mean and interval(the lower and upper bounds)of demand information of each affected area rather than any exact demand distribution.A joint chance constraint is introduced to the model to ensure that a certain level of service can be achieved,which can be approximately transformed into a linear integer programming problem that can be solved by the commercial solver CPLEX,and iteratively solved by cutting plane algorithm.The effectiveness of the emergency commodities supply network model is verified by a case study of Typhoon Maria that occurred in China in 2018.The computational results show that the emergency preparedness network model considering the multi-sourcing supply strategy can achieve a higher service level with less inventory of relief packages and lower total cost,and the average improvement of total inventory level of relief packages(optimal total cost,resp.)of the multi-sourcing model relative to the single-sourcing model is between 2% to 3.5%(between 9% to 11%,resp.),which is significantly better than that considering the single-sourcing supply strategy.The observation is further verified by the numerical experiment on larger-scale network structures.