Scheduling And Optimization Of Large-Scale Sorting System Based On Risk Model

As an important part of the warehousing system,the sorting platform has long faced problems such as large fluctuations in task demand,strong subjectivity in scheduling decisions,and low utilization of human resources.How to overcome the strong uncertainty of task requirements,improve the automation level of the scene,and formulate a scientific scheduling decision-making plan has high research value and practical significance.In this paper,a series of researches are carried out on the operation process of the storage sorting platform,focusing on the optimization of personnel scheduling,and a technical framework including process modeling,digital twins,risk estimation,scheduling optimization,and equipment scheduling has been established.The main research work are:1.Process modeling and digital twin of the warehouse sorting platform.Based on the actual production logic of the storage and sorting system,a process model is established to solve the integer programming problem with the lowest total labor cost as the optimization goal.In view of the opacity of on-site production data and the management based on subjective experience,a modular simulation platform was established to realize a digital twin of on-site operation processes.The simulation platform can restore the workload of each operation unit according to the operation logic,and provide a theoretical basis for on-site management personnel to monitor the production progress,perform in-process scheduling,and explore the potential of the operation.2.The task volume risk modeling of the warehouse sorting platform.Accurate task volume forecasting is the prerequisite for lean scheduling of storage and sorting systems and reduction of labor costs.The task volume is affected by complex internal and external factors such as upstream sales,weather and road conditions,time period,etc.,and the prediction results output by scalar often have large deviations.In order to overcome the uncertainty in the task volume forecasting process,this paper proposes a risk modeling forecasting method based on GPR.The prediction algorithm predicts the probability density distribution of the proportion of tasks in a day at each time in the future based on the existing historical task volume prediction values.3.Optimized algorithm for personnel scheduling of warehouse sorting platform.A two-stage stochastic optimization model was designed to formulate a lean shift schedule for warehousing and sorting with the lowest total labor cost.First,the scenario tree is constructed by sampling in the risk model,and then the optimal personnel scheduling scheme is calculated considering the uncertainty of the task volume.This sampling-based optimization method makes full use of the probability density information described by the risk model,and solves the problem of scarce historical data and large noise in practical applications.Design experiments to compare the labor costs of various scheduling schemes and verify the effectiveness of the optimization algorithm.4.The flexible manpower distribution model of the storage and sorting platform.Lean scheduling is part of the task planning level,and at the production management level,enterprises can also improve the utilization of human resources and reduce production labor costs by formulating flexible management methods.In this paper,a flexible management method is designed in terms of the type of labor,shift duration,and salary of the warehouse sorting platform,and the relevant personnel distribution model is established.The research results have been applied and verified in the real production scenarios of logistics enterprises.