Optimizing The Daily Milk-run System Of An Automobile Manufacturer With The Consideration Of Dock Occupation

In order to achieve just-in-time production,manufacturing enterprises has widely used milk-run system in their inbound logistics systems.In the milk-run system,manufacturing enterprises feed materials needed by the production line through the small-and-multiple batches.Due to limited number of docks,it is possible for the vehicles to queue when frequently loading and unloading at the depot.In this context,this study considers vehicle capacity,available time at the loading and unloading docks,etc,as constraints,integrates the selection of transportation vehicles,vehicle routing,and pickup interval,takes the minimum number of vehicles as objective,establishes planning model of cyclic vehicle routing problem to optimize scheduling program of the manufacturer.Based on it,a hybrid genetic algorithm integrating route checking module and dock scheduling module is proposed to generate vehicle routing plan.In the proposed method,5 crossover operators and 3 mutation operators are randomly selected to generate offspring and a local education method is applied to enhance the fitness of offspring.Furthermore,in order to improve the efficiency of the algorithm,a route-check module is integrated to verify preliminary feasibility and a schedulecheck module is conducted by Flexsim to simulate the loading,unloading and queuing at the depot and generates schedule of the vehicles.The method proposed in this paper is applied to a real example of a Sino-foreign joint venture auto company to show the efficiency of the proposed algorithm.It founds that the method can effectively optimize the pickup system,reduce the number of vehicles needed from the original 21 cars to 15 cars,then the cost was reduced by 28.57%.In addition,the results show that in the milk-run system,the queuing time at the depot cannot be ignored when the number of docks is limited.Moreover,it’s more inclined to choose suppliers closed to the depot when selecting suppliers.Facing different solutions with the same number of vehicles,decision makers may tend to choose the solution that has less routes.The hybrid genetic algorithm and a series of management decision-making suggestions proposed by this research can assist the logistics department of manufacturing enterprises to optimize their strategy of inbound logistics,so as to decrease the number of vehicles used and reduce the logistics cost.It is of great practical value to improve their market competitiveness.