Research On Location Routing Problem Of Multi-type Precooling Facilities Based On Bi-level Programming Model

With the improvement of resident consumption level,the requirements for the circulation quality of fresh agricultural products are getting higher,and the cold chain market of agricultural products has been rapidly expanded.But the decay loss rate of fresh agricultural products in China is still higher than developed countries,and the key precooling problem in the first mile of the post-harvest stage has been widely concerned by the government and enterprises.At the same time,compared with the traditional architectural precooling station having the characteristics of being far away from the origin,low utilization rate,and long construction period,new precooling facilities such as modular temporary precooling warehouse and mobile precooling vehicles have become effective tools to solve the precooling problem.Combined with the national conditions of dispersed agricultural production and the majority of smallholders in China,it is necessary to unify the layout of both fixed and mobile precooling facilities considering the complex field environment.Starting from three aspects of background,core problem and solution method,this paper reviewed the relevant research progress and theoretical basis of precooling link,location-routing problem and bi-level programming solution method.Through summarizing domestic and foreign research status and shortcomings,this paper identified multi-type precooling facilities location routing optimization problem which was studied.Then,a bi-level programming model from the perspective of the precooling service provider was formulated and a two-stage heuristic algorithm was designed.Finally,the proposed model and algorithm were applied to a practical example to test the performance.The main work done in this study are as follows:(1)Formulating a bi-level programming model for the location routing problem of multi-type precooling facilities.Considering the basic characteristics of multi-type precooling facilities in the first mile of agricultural products' post-harvest stage,the location routing model of multi-type precooling facilities was formulated based on a bi-level programming method.The upper model was used to solve the facility location problem with uncertain capacity,and the objective was to minimize the total cost of precooling facilities while maximizing their comprehensive service demand.The lower model was used to solve the multi-center multi-vehicle routing problem in which the objective was to minimize the total cost of vehicle transportation.Besides,the personalized customer demand for service time,uneven distribution of customer's precooling demand and precooling delay and other factors were considered in the model,contributing to improving the reliability of the layout optimization of the actual precooling facility.(2)Designing a two-stage heuristic algorithm based on clustering.First,the clustering algorithm was improved considering the demand factors except for spatial distance.It was used to analyze the distribution of precooling demand and to determine the candidate location of precooling warehouse from known customer points.Second,the NSGA-? algorithm with elite strategy was used to solve the facility location problem in the upper model.In the lower model,the temporal-spatial distance factor was introduced according to the customer differences in spatial distance and service time requirements,and the corresponding customer point allocation algorithm and genetic algorithm were designed to solve the vehicle routing problem.Finally,based on numerical experiments,it was found that the combination of the maximum capacity,the number and the coverage distance of temporary precooling warehouses could have an important impact on the location routing decision of multi-type precooling facilities.To sum up,this study extends the theory and methods of the cooperative layout of multi-type precooling facilities in the first mile of agricultural products' post-harvest stage to a certain extent,enriches the research on the location routing problem of precooling facilities,and provides a reference for related research.Also,it is helpful for e-commerce enterprises to optimize the layout of precooling facilities,reduce the decay loss rate of agricultural products and improve their economic benefits.