Research On Vehicle Logistics Order Allocation Based On Demand Forecasting And Hierarchical Decision-making

Vehicle logistics order allocation is the one of the key decisions in vehicle logistics,and the quality of the decision-making will have a direct impact on vehicle delivery level and vehicle delivery cost.In recent years,competition in the automotive market becomes fiercer,putting forward higher requirements for the delivery level;road transport regulations are tightened,putting delivery cost under greater pressure.At present,the intelligent decision-making system of vehicle logistics is in its initial stage,focusing on the current road order allocation problem,lacking a forward-looking scientific solution for the whole process of order allocation,and it is difficult to take into account both delivery level requirements and delivery cost objectives.Therefore,from the perspective of vehicle logistics service supply chain management,this thesis analyzes the constraints and objectives of the whole vehicle logistics order allocation problem,constructs models from the perspective of demand forecasting and hierarchical decision-making,and proposes the solution.The effectiveness of the models and algorithm is verified by combining the example data and numerical experiments.Through a case study,the practical application value of the models and algorithm is tested,and further allocation decision-making suggestions are proposed.Firstly,from the perspective of demand forecasting,in view of the difference in the use of forecasted orders in allocation decisions,a ten-day order demand forecasting model and a daily demand forecasting model are established respectively,and the validity of the forecasting models is verified by practical forecast examples.For the ten-day order demand forecasting problem,on the basis of the grey forecasting model,a grey seasonal index forecasting model is established,and the fitting accuracy is improved from 69.76% to 89.22%,and the forecast accuracy is improved from 42.80%to 90.13%,which can provide a reference basis for subsequent capacity preparation.For the daily demand forecasting problem,on the basis of ARIMA model.EMDARIMA model is established,the Mean Absolute Error is reduced from 297.83 to204.24,the forecast accuracy is improved from 61.76% to 84.92%,which can improve the success rate of daily order allocation.Secondly,from the perspective of hierarchical decision-making,this thesis analyses the constraints and objectives of the whole process of vehicle logistics order allocation,constructs a three-layer decision-making model of “transport mode selection-loading plan selection-vehicle FLSP selection” and design a solving algorithm,and verifies the effectiveness of the model algorithm through numerical experiments.The upper layer is a selection model of transportation mode aiming at minimizing delayed delivery time and cost.The middle layer is a selection model of auto-carriers loading plan aiming at maximizing the loading value.The lower layer is a selection model of FLSP aiming at maximizing global satisfaction of functional logistics service providers.There is still little relevant research on the solution of the upper layer of the model,so a heuristic algorithm is proposed.The numerical experiments find: the heuristic algorithm has good solving quality and solving efficiency under different scale cases compared with GUROBI,NSGA-2 and MOPOS.Based on the general loading strategy,the middle layer of the model can further optimize the urgent order loading,space utilization,and number of auto-carriers’ destinations and other indicators.Compared with the current situation of vehicle order allocation,the lower layer of the model can effectively improve the satisfaction of global logistics service providers.Finally,the practical application of the proposed models and algorithm is verified by the case study of Company Z,using the actual allocation solution as a reference,and relevant decision-making recommendations are made.The results of the case show that both the hierarchical decision-making model solution and the model solution based on demand forecasting and hierarchical decision-making outperform the actual allocation solution.When the forecast accuracy of order demand is high enough,the model solution based on demand forecasting and hierarchical decision-making performs better.When the forecast accuracy of the order demand is difficult to reach a high level,the model solution based on demand forecasting and hierarchical decision-making performs better in term of cost,and the hierarchical decision-making model solution performs better in term of time.Based on the study results,the following recommendations are made: when the order demand tends to be stable,the model solution based on demand forecasting and hierarchical decision-making is recommended.When order demand is highly stochastic,if the decision-makers prefer cost-first strategy,the model solution based on demand forecasting and hierarchical decision-making is recommended;if the decision-makers prefer time-first strategy,the hierarchical decision-making model solution is recommended.The research work in this thesis theoretically enriches the research results in the field of whole vehicle logistics decision-making and logistics service supply chain order allocation;practically it provides decision-making reference and basis for decision makers and related managers in the whole vehicle logistics industry.