Research On Many-Objective Optimization Of Distributing Emergency Relief Supplies

In recent years,various natural disasters have occurred frequently in China,causing huge losses to the national economy and social development.After a disaster,how to efficiently and scientifically distribute multiple types of emergency relief supplies from multiple supply locations to multiple demand locations in the disaster area is a very difficult problem in emergency response.It should be noted that the existing research is limited to three optimization objectives,and often only considers the distribution of emergency relief supplies in a specific situation.This simple multiobjective optimization idea and a single emergency scenario are difficult to fit the complex emergency environment and meet the complex emergency needs in reality.To improve the flexibility and scientificity of emergency decision-making,this dissertation focuses on many-objective optimization of distributing emergency relief supplies for post-disaster response.The main research work of this dissertation is as follows:(1)The research background and significance of the problem of emergency relief supply distribution are summarized.Specifically,the related work and drawbacks are discussed on the basis of the analysis of the current research literature.Then,the relevant knowledge of multi-objective optimization theory is introduced.Particularly,the Strength Pareto Evolutionary Algorithm 2(SPEA2)and Shift-based Density Estimation(SDE)based SPEA2+SDE algorithm is described in detail.(2)A clear mathematical model for many-objective emergency relief supply distribution problem is developed in terms of multiple types of emergency relief supplies,multiple supply locations,multiple demand locations,and various influencing factors in the complex post-disaster emergency scenarios.In the proposed model,six objectives,including the deployment time,the panic level,the dissatisfaction degree,the envy degree,the expectation loss,and the operation cost,are considered simultaneously.Additionally,three constraints,including no overwork constraints,no oversupply constraints,and maximum contribution constraints,are also taken into consideration.Simulation experiments verified the effectiveness of the proposed model.(3)A many-objective emergency relief supply distribution algorithm is proposed based on SPEA2+SDE,two-dimensional integer vector encoding,and Supply-Oriented Technique(SOT)under the extreme shortage of emergency relief supplies in the early post-disaster period.Unlike the Demand-Oriented Technique(DOT),SOT requires all supply locations to contribute all emergency relief supplies at the current period.The proposed algorithm,called SPEA2+SDE+SOT,is compared with SPEA2+SDE+DOT,Encoding Repair and None-dominated Sorting based Differential Evolution(ERNSDE)and Greedy-Search-based Multi-Objective Genetic Algorithm(GSMOGA).The experimental results show that the proposed SPEA2+SDE+SOT has better overall performance.(4)A many-objective adaptive distribution algorithm for emergency relief supplies is proposed based on SPEA2+SDE and Adaptive Individual Repairs(AIR)under the situation that some types of emergency relief supplies can be satisfied in the middle and later periods of emergency response.AIR can adaptively determine the corresponding distribution schemes according to the supply and demand of each emergency relief supply.Finally,the effectiveness of the proposed algorithm,named SPEA2+SDE+AIR,is verified in comparison with ERNS-DE and GSMOGA.