Research On The Online-Merge-Offline Retail Store Location Problem

In recent years,the retail industry is going through a new round of upgrading.Some companies have proposed some new business models,such as online merge offline(OMO).This kind of stores not only includes functions similar to those of traditional supermarkets,but also can get demands from online.For the online demands,retail stores need to distribute the goods to the consumers.This paper adapts these new changes and formulate a bi-objective location model for this kind of retail stores.We illustrate the use of model using actual date from Beijing,which is got from LIANJIA.COM and GAODE MAP.This paper formulates a bi-objective location model based on the Huff’s law for traditional retail stores first.This model has two objective functions:(i)maximizing total demands got by the system and(ii)minimizing the total cost.To solve this problem,a non-dominated sorting genetic algorithm is used.Then,a more complicated bi-objective location model is formulated for the OMO retail stores.this model also maximizing the total demands got by the system and minimizing the total cost.The maximizing function can be divided into two parts:online-part and offline-part.The former is similar to the traditional retail stores location model,we use a new parameter"_#,which means the probability that travel time is less than or equal to time threshold,to replace the fixed time threshold.We use actual date from Beijing to illustrate the advantage of incorporating"_#.For the latter part,a///queueing system is incorporated in this model to determine how many riders are arranged at each retail store.In order to make sure the feasibility of the busy fraction iterative algorithm,we choose two iterative algorithms as a contrast,one is station-specific busy fraction iterative algorithm,the other one is system average busy fraction iterative algorithm.The difference between these two iterative algorithms is less than 8%.The station-specific busy fraction iterative algorithm gets a more practical result than system average busy fraction iterative algorithm.Finally,we use the OMO retail stores location model to explore the property of this kind of retail stores.Our result shows that efficiency is still the key to the retail industry.The OMO retail store holders should try their best to improve the efficiency of distribution.