A dynamic scheduling method for sorting personnel based on their biometric characteristics

Abstract

The present invention discloses a method for dynamic scheduling of personnel based on the biometric characteristics of sorting personnel, comprising the following steps: collecting data on the individual biometric characteristics of personnel at logistics transfer sites, preprocessing the collected data, and constructing a biometric vector; constructing a virtual mapping system for the ternary collaboration of man, machine, and task to achieve predictive optimization and state-driven regulation of scheduling strategies; designing a multi-objective function based on the current sorting task, personnel health, sorting priority, and fatigue control as goals to achieve comprehensive optimization of dynamic scheduling; based on the established multi-objective optimization model, applying a multi-objective evolutionary algorithm to the multi-objective scheduling problem, and performing structural optimization and heuristic enhancement for logistics scenarios. The present invention realizes dynamic perception, intelligent optimization, and closed-loop control of logistics sorting scheduling while taking into account the improvement of operating efficiency, personnel health protection, and task response accuracy.

Description

Technical Field

[0001] The present invention relates to the field of logistics scheduling, and in particular to a personnel dynamic scheduling method based on the biological characteristics of sorting personnel. Background Art

[0002] With the rapid development of e-commerce logistics, sorting operations at logistics transfer stations are facing the dual challenges of improving efficiency and ensuring personnel safety. Traditional static scheduling methods, such as fixed eight-hour shifts and experience-based manual allocation, are no longer able to meet the flexible and intelligent management requirements of modern logistics. These methods often ignore the physiological state of sorters, such as key biometrics like fatigue and heart rate variability, leading to irrational allocation of work intensity, resulting in overwork and even health risks. Furthermore, their scheduling mechanisms lack flexibility and cannot dynamically adjust to real-time workloads and personnel status, resulting in inefficient resource allocation. Furthermore, static algorithm parameters are fixed, making them incapable of adapting to complex scenarios such as peak hours and sudden job vacancies. In recent years, the rapid development of wearable devices (such as smart wristbands and smart ID cards) and biometric sensor technology has enabled the real-time collection of sorter biometric data, providing a technical foundation for building a human-centric, dynamic scheduling system. However, how to deeply integrate these high-frequency, multi-dimensional human physiological data with scheduling algorithms, and achieve accurate modeling and prediction of personnel status through digital twin models, so as to achieve dynamic scheduling that takes into account work efficiency, personnel health and task urgency, is still an important technical difficulty that urgently needs to be overcome. Summary of the Invention

[0003] The purpose of the present invention is to overcome the shortcomings of the existing technology and provide a dynamic personnel scheduling method based on the biological characteristics of sorting personnel. Under the premise of taking into account the improvement of work efficiency, personnel health protection and task response accuracy, it realizes dynamic perception, intelligent optimization and closed-loop control of logistics sorting scheduling, and solves the problems of the traditional static scheduling mode's lack of physiological state perception and poor adaptability to complex environments.

[0004] The purpose of the present invention is achieved through the following technical solution: a dynamic scheduling method for sorting personnel based on the biological characteristics of sorting personnel, comprising the following steps:

[0005] Collect individual biometric data of logistics transfer station staff, pre-process the collected data, and construct biometric vectors;

[0006] Build a virtual mapping system for the collaboration of humans, machines, and tasks to achieve predictive optimization and state-driven control of scheduling strategies;

[0007] Based on the current sorting task, personnel health, sorting priority and fatigue control as goals, a multi-objective function is designed to achieve comprehensive optimization of dynamic scheduling;

[0008] Based on the established multi-objective optimization model, the multi-objective evolutionary algorithm is applied to the multi-objective scheduling problem, and structural optimization and heuristic enhancement are performed for logistics scenarios.

[0009] The beneficial effects of the present invention are as follows: the present invention constructs a dynamic scheduling and dispatching optimization system driven by multi-source biometrics for intelligent logistics transfer sites. By integrating high-frequency physiological data collected by wearable devices, digital twin modeling and multi-objective evolutionary algorithms, the dynamics and intelligence of the scheduling and dispatching strategy are comprehensively improved. Specifically, the present invention establishes a personnel status quantification model with key physiological indicators such as heart rate, fatigue, and attention as core variables, revealing the intrinsic coupling relationship between the physiological status of workers and work efficiency and health risks; at the same time, digital twin technology is introduced to construct a virtual mapping system of "man-machine-task" ternary collaboration, realizing real-time mapping and prediction of the individual status of sorting personnel, and providing high-precision, feedback-enabled dynamic input for the scheduling system. In terms of scheduling optimization, the present invention proposes a multi-objective scheduling optimization model that covers maximizing operational efficiency, minimizing health risks, and task emergency response and fatigue balance. It also performs structural optimization and adaptive enhancement based on the NSGA-II algorithm, and proposes a two-dimensional matrix chromosome encoding method to ensure spatial continuity and temporal coordination of task allocation. By introducing an adaptive crossover mutation strategy, the algorithm's population diversity and local search capabilities are effectively improved, and its robustness and adaptability to complex scenarios such as peak periods and sudden job vacancies are enhanced. Under the premise of taking into account operational efficiency improvement, personnel health protection, and task response accuracy, the present invention realizes dynamic perception, intelligent optimization, and closed-loop control of logistics sorting and scheduling, breaking through the technical bottleneck of the traditional static scheduling model's lack of physiological state perception and poor adaptability to complex environments, and providing a new theoretical basis and technical support for the field of intelligent logistics scheduling. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 Flow chart of the method of the present invention. DETAILED DESCRIPTION

[0011] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings, but the protection scope of the present invention is not limited to the following.

[0012] This patent aims to solve the key technical difficulties existing in the current logistics sorting and scheduling, including: how to scientifically quantify the correlation between the biological characteristics of sorting personnel and their work efficiency, construct a mathematical model with physiological indicators such as heart rate and fatigue as core variables, and reveal the impact mechanism of individual physiological state on work performance; how to build a dynamic virtual mapping body of sorting personnel based on digital twin technology, and achieve accurate modeling and dynamic simulation of individual status by integrating multi-source real-time data collected by wearable devices (such as heart rate, movement trajectory, working hours, etc.), thereby forming a "virtual-real linkage" system that is highly synchronized with real personnel, providing real-time and predictable digital support for scheduling strategies; on this basis, how to achieve dynamic scheduling driven by multiple goals, while ensuring work efficiency, taking into account personnel health, safety risks and task urgency, and completing intelligent optimization of scheduling strategies.

[0013] like Figure 1 As shown, a dynamic scheduling method for sorting personnel based on biometric characteristics includes the following steps:

[0014] Collect individual biometric data of logistics transfer station staff, pre-process the collected data, and construct biometric vectors;

[0015] For the collection and quantification of individual biometric data of logistics transfer workers, we use smart ID cards, smart bracelets and eye trackers to collect the following biometric data in real time: heart rate , unit is times / minute; body temperature , unit is degrees Celsius; range of motion , a normalized value calculated by the acceleration sensor, ranging from 0 to 1; fatigue , a standardized evaluation value based on physiological electromyographic signals, ranging from 0 to 1; eye tracking , the standardized focus of attention determined by an eye tracker, ranging from 0 to 1.

[0016] For the fatigue quantification model, we combine electromyographic signals (EMG) to construct a fatigue function:

[0017]

[0018] in, is the weight; The wearable patch electrode is attached to the skin on the surface of the muscle to collect the electrical signals generated by the muscle during activity. The wearable patch electrode detects the weak potential changes generated by muscle contraction through the skin. The action frequency refers to the peak value detection of the envelope in the action state: each muscle contraction usually forms an obvious peak on the envelope. The action frequency is calculated by counting the number of peaks that appear within a certain period of time, which is recorded as ; The base frequency is the preset frequency: 1 time / 15 seconds;

[0019] Based on the above definition, we integrate the above biometric data into a biometric vector .

[0020] The acceleration modulus is calculated using the acceleration measured by the three-axis acceleration sensor:

[0021]

[0022] 、 Indicates the measured value at time t x , y , z Acceleration in three directions; represents the acceleration modulus at time t;

[0023] Using min-max normalization:

[0024]

[0025] in, is the historical statistical maximum acceleration modulus value, is the acceleration modulus value collected when standing still;

[0026] For the above collected data, data preprocessing is performed: the heart rate ,body temperature The time series data is denoised using the following formula:

[0027]

[0028] in, is the observed value, i.e., heart rate or body temperature , is the filtered value, is the Kalman gain; the heart rate after denoising Recorded as , body temperature after denoising Recorded as ;

[0029] The constructed biometric vector is recorded as:

[0030] .

[0031] To achieve real-time perception, dynamic mapping, and intelligent simulation of the operating status of sorting personnel at logistics transfer sites, this invention constructs a virtual mapping system that integrates the "human-machine-task" ternary collaboration to achieve predictive optimization of scheduling strategies and state-driven regulation. Specifically, each operator corresponds to a unique digital twin, which is dynamically modeled based on the following three types of information:

[0032] (1) Static information layer: contains personnel ID, job type, maximum working time, and health records, which are used to initialize the basic parameter constraints of the twin model;

[0033] (2) Dynamic state layer: real-time acquisition of biometric feature vectors , transmitted to the digital twin of the staff, and based on the edge computing gateway, the data of the staff digital twin is uploaded to the digital twin system of the sorting machine in real time;

[0034] (3) Simulation behavior layer:

[0035] The state professional model is constructed by using the collected historical data and adopting the time series prediction model based on LSTM network: First, based on the historical information, multiple groups of sample data are constructed. Each group of sample data contains the states of two adjacent time windows. After the LSTM network is trained using multiple groups of samples, a trained LSTM network is obtained. Then, the state of the current time window is input into the trained LSTM network to obtain the state of the next time window. The state refers to the state at each moment. ;

[0036] Based on the historical state of the most recent time window, input it into the trained LSTM model to obtain the state information in the next time window;

[0037] Sorting task information: Predefine and clarify the sorting task details, including the sorter's position, the number of packages to be sorted, and the sorter status. Then, establish a task process model for the sorting task, including the interaction between the sorter and the sorter during the sorting process. Connect the task information to the sorter's digital twin model.

[0038] The digital twin model of the sorting machine establishes an accurate three-dimensional model of the equipment based on the CAD model, kinematic and dynamic parameters of the actual equipment, and connects the physical equipment model with the real-time data update interface to achieve two-way mapping between the virtual environment and the physical equipment; the digital twin models and task information of the sorting machine and personnel both require interfaces for synchronous data communication; the digital twin model of the sorting machine is used to accurately evaluate the current sorting task information, and by accessing the biological status data of the sorting personnel's digital twin, it obtains the sorting performance index, namely the package return rate.

[0039] Multi-objective optimization model construction: Based on the current sorting task, personnel health, sorting priority and fatigue control as goals, a multi-objective function is designed to achieve comprehensive optimization of dynamic scheduling;

[0040] (1) Work efficiency goals

[0041] Sorting efficiency It is defined as the number of parcels that can be sorted per minute, which is related to the real-time fatigue and concentration of the sorting personnel. ,in , is the weight coefficient, satisfying .

[0042] (2) Personnel health goals

[0043] Health Risk Index ,in The baseline values ​​for healthy heart rate and body temperature are: are the thresholds for healthy heart rate and body temperature, is the weight.

[0044] (3) Mission urgency goals

[0045] Task urgency ,in is the adjustment coefficient, The remaining time between the current sorting task and the package dispatch time. The time it takes for each package to enter the sorting system and be sorted and shipped.

[0046] (4) Fatigue control target

[0047] Defining fatigue balance , where Q is the total number of sorters in the current sorting shift, is the average fatigue level, is the weight; represents the fatigue level of the qth sorter;

[0048] Combining the above objectives, we construct a multi-objective optimization function:

[0049]

[0050] in, The weight coefficients of each factor are dynamically adjusted through the analytic hierarchy process (AHP), including:

[0051] Constructing a hierarchical model: The typesetting scheduling problem is divided into two layers: (1) the target layer achieves the optimal balance between scheduling efficiency and personnel health; (2) the criterion layer includes four core optimization objectives: , that is, the operating efficiency target, express; , that is, the health risk control target, express; , that is, the task urgency response target, express; , that is, the fatigue control target, express.

[0052] Based on the above four defined criteria, a pairwise comparison matrix is ​​constructed. ,in Indicates the importance of criterion i relative to criterion j.

[0053] Assume that the dimension of the pairwise comparison matrix is ​​n (4 in this invention), and obtain its maximum eigenvalue by eigenvalue decomposition method. , used to measure the consistency of the comparison matrix. Calculate the consistency index And look up the table to obtain the random consistency index (RI is an empirical value proposed by Saaty, which represents the expected consistency index under completely random conditions. It has different values ​​for matrices of different orders. In the current case of n=4, RI=0.9.) Then calculate the consistency ratio , like , then the judgment matrix is ​​considered to have good consistency and the weights are reasonable and reliable; on the contrary, if , it is considered that there is a large deviation in the judgment matrix and the matrix structure needs to be revised to improve consistency.

[0054] Based on the established multi-objective optimization model, the multi-objective evolutionary algorithm is applied to the multi-objective scheduling problem, and structural optimization and heuristic enhancement are performed for logistics scenarios.

[0055] Based on the multi-objective optimization model established above, this paper applies the classic multi-objective evolutionary algorithm NSGA-II to the multi-objective scheduling problem, performing structural optimization and heuristic enhancements specifically for logistics scenarios. To ensure spatial continuity and temporal controllability of task allocation, the following two-dimensional matrix structure is used to encode chromosomes and represent the scheduling solution: ,in is the number of sorting areas, is the number of scheduling time periods, is the sorting personnel id of the i-th sorting area and the j-th scheduling time period. Personnel collection, each person cannot be assigned to multiple areas in the same time period. Each scheduling plan is regarded as an individual, and all individuals constitute a population;

[0056] In order to improve population diversity and avoid premature convergence, the following adaptive probability adjustment formula is designed:

[0057] Crossover probability :

[0058]

[0059] Mutation probability :

[0060]

[0061] in, is the fitness of the current individual (multi-objective weighted value, i.e. ; is the maximum and average fitness of the current population; a trace constant to prevent division by zero; are the upper and lower bounds of the crossover probability, The lower and upper bounds of the mutation probability. The worse the individual (the lower the fitness), the higher the crossover probability, which enhances exploration; the higher the mutation probability, the greater the possibility of mutation to escape the local optimum.

[0062] Implementation of multi-objective evolutionary algorithm NSGA-II:

[0063] 1. Initialization phase

[0064] 1.1 Encoding method: ,in is the number of sorting areas, is the number of scheduling time periods, is the sorting personnel id of the i-th sorting area and the j-th scheduling time period; the personnel scheduling plan used in the current logistics transfer station is used as the initial population .

[0065] 1.2 Parameter setting: maximum number of iterations , population size (number of different scheduling schemes) N

[0066] 2. Fitness Evaluation

[0067] For each individual in the population (scheduling scheme):

[0068] 2.1 Calculate four objective functions:

[0069]

[0070]

[0071]

[0072]

[0073] 2.2 Multi-objective fitness weighting:

[0074]

[0075] 3. Non-dominated sorting and congestion calculation

[0076] 3.1 Non-dominated sorting:

[0077] Divide the population into several levels (Front 1, Front 2, ...);

[0078] Front 1: completely non-dominant individuals;

[0079] Front 2: individuals dominated by Front 1 but not by others;

[0080] And so on.

[0081] Domination means: A dominates B if and only if: A is not inferior to B in all objectives; A is superior to B in at least one objective.

[0082] A solution is called a non-dominated solution if no other solution dominates it.

[0083] NSGA-II tends to retain non-dominated solutions because they represent the most balanced and optimal scheduling scheme combinations among multiple objectives.

[0084] In NSGA-II:

[0085] First layer Front ( ): all non-dominated solutions;

[0086] Second layer Front ( ): Individuals that are dominated only by the solution of Front1;

[0087] And so on.

[0088] 3.2 Crowding distance:

[0089] In the same Front, the congestion degree is calculated based on the objective function value;

[0090] Individuals with high crowding are retained first to ensure population diversity.

[0091] Crowding: Within the same non-dominated layer (e.g., Front1), different solutions may have similar objective values. To maintain diversity, we prioritize those solutions that are "sparsely distributed." NSGA-II uses "crowding" to measure the sparseness of the area around a solution: higher crowding indicates more isolated solutions, which are more likely to be retained; lower crowding indicates more crowded solutions, which are more likely to be eliminated. Crowding ensures that different scheduling strategies are retained, rather than a single solution being retained.

[0092] Calculation scheme: For the kth objective function , the crowding degree of individual i is:

[0093]

[0094] in, , is the value of the adjacent individuals of individual i in the kth target dimension. , is the maximum and minimum value of the objective function.

[0095] 4. Choose an action:

[0096] Parent individuals are selected from the current population using a binary tournament method. This involves randomly selecting two individuals from the population and comparing their non-dominance ranks (fronts): the lower rank is prioritized. If the ranks are tied, the one with the higher crowding is prioritized. The superior individual is selected as the "parent" for crossover and mutation.

[0097] 5. Crossover and Mutation

[0098] Crossover: Utilizes defined adaptive crossover probabilities , perform segment crossover on the time periods or regions of the two parent chromosomes and maintain individual legitimacy (no repeated allocation).

[0099] Mutation: Utilizes defined adaptive mutation probabilities , randomly select a position in the matrix and replace it with other people who are not assigned in the same time period to ensure that the constraints are satisfied.

[0100] 6. Merger

[0101] The current population P t and the offspring population Q t Merge into R t , for R t Perform non-dominated sorting and select N individuals from the first several Fronts to form the next generation P t +1: Prioritize lower levels and higher congestion levels.

[0102] 7. Termination and Output

[0103] If the maximum number of iterations is reached Or if the convergence condition is met, it terminates and outputs the final scheduling measurement plan

[0104] The optimal individual solution is finally output by the above multi-objective evolutionary algorithm (NSGA-II). The optimization goal is constructed in point 5, and the representation method of the scheduling solution (chromosome) is defined in point 7. The multi-objective evolutionary algorithm (NSGA-II) searches for the optimal scheduling solution under the constraints of these objective functions, and finally outputs a two-dimensional matrix ,in:

[0105] Row, sorting area: Each row represents a different predefined sorting area in the logistics transfer yard;

[0106] Column, scheduling time period: each column represents a scheduling period, in units of one hour;

[0107] The content of each cell in the matrix: sorter ID.

[0108] The comprehensive significance of scheduling measurement: 1. Personnel allocation: This specifically reflects the arrangement of sorting personnel in each area during each time period, ensuring that personnel do not overlap and meet regional task requirements. 2. Task volume allocation: Based on the expected sorting task volume, the scheduling system not only determines who will go to which area and when, but also estimates the number of sorting tasks that the personnel should complete within the corresponding time period, which is then used for subsequent operation monitoring and closed-loop feedback control. 3. Dynamic adjustment basis: This scheduling matrix, as the final output, is generated by a multi-objective optimization algorithm (such as NSGA-II) and can be dynamically adjusted based on real-time data to maximize operational efficiency, minimize health risks, ensure timely emergency response to tasks, and balance personnel fatigue.

[0109] The foregoing description is a preferred embodiment of the present invention. It should be understood that the present invention is not limited to the form disclosed herein and should not be construed as excluding other embodiments. Instead, the present invention can be used in other combinations, modifications, and environments and can be modified within the scope of the concept described herein through the above teachings or techniques or knowledge in the relevant field. Modifications and variations made by those skilled in the art that do not depart from the spirit and scope of the present invention are intended to be protected by the appended claims.

Claims

1. A dynamic scheduling method for sorting personnel based on their biometric characteristics, characterized by: The following steps are involved: Collect individual biometric data of logistics transfer station staff, pre-process the collected data, and construct biometric vectors; Build a virtual mapping system for the collaboration of humans, machines, and tasks to achieve predictive optimization and state-driven control of scheduling strategies; Based on the current sorting task, personnel health, sorting priority and fatigue control as goals, a multi-objective function is designed to achieve comprehensive optimization of dynamic scheduling; The multi-objective function is designed based on the current sorting task, personnel health, sorting priority and fatigue control as the goals, including: The design goals are as follows: (1) Work efficiency target, using sorting efficiency Characterization: Sorting efficiency It is defined as the number of parcels that can be sorted per minute. It is related to the real-time fatigue and concentration of the sorting personnel. The current sorting status of the sorting machine is affected by the type of sorting task and the current operating status of the sorting machine. The sorting efficiency is recorded as ,in , , is the weight coefficient, satisfying ; The real-time parcel return rate is output by the digital twin model of the sorting machine combined with the biometric status data of the sorting personnel, which is equal to the number of super-circuit pieces divided by the total number of pieces; For fatigue, For eye tracking; (2) Personnel health goals, using health risk index Characterization: Health Risk Index ,in The baseline values ​​for healthy heart rate and body temperature are: are the thresholds for healthy heart rate and body temperature, is the weight; is the heart rate after denoising; is the body temperature after denoising; (3) Task urgency goal, using task urgency Characterization: Task urgency ,in is the adjustment coefficient, The remaining time between the current sorting task and the package dispatch time. The time it takes for each package to enter the sorting system and be sorted and shipped; (4) Fatigue control target, using fatigue balance Characterization: Defining fatigue balance , where Q is the total number of sorters in the current sorting shift, is the average fatigue level, is the weight; represents the fatigue level of the qth sorter; Combining work efficiency goals, personnel health goals, task urgency goals, and fatigue control goals, a function for constructing a multi-objective optimization model is constructed: in, are their respective weight coefficients; Based on the established multi-objective optimization model, the multi-objective evolutionary algorithm NSGA-II is applied to the multi-objective scheduling problem, and structural optimization and heuristic enhancement are performed for logistics scenarios.

2. The method for dynamic scheduling of sorting personnel based on their biometric characteristics according to claim 1, characterized in that: The data collection of individual biometric characteristics of logistics transfer site staff includes: The following biometric data of staff members are collected in real time: Heart rate , the unit is times / minute; body temperature , the unit is Celsius; Range of motion , a normalized value calculated by the acceleration sensor, ranging from 0 to 1, is calculated as follows: The acceleration modulus is calculated using the acceleration measured by the three-axis acceleration sensor: 、 Indicates the measured value at time t x , y , z Acceleration in three directions; represents the acceleration modulus at time t; Using min-max normalization: in, is the historical statistical maximum acceleration modulus value, is the acceleration modulus value collected when standing still; Fatigue , a standardized evaluation value based on physiological electromyographic signals, ranging from 0 to 1; The calculation of is combined with electromyographic signals to construct the fatigue function: in, is the weight; The wearable patch electrode is attached to the skin on the surface of the muscle to collect the electrical signals generated by the muscle during activity. The wearable patch electrode detects the weak potential changes generated by muscle contraction through the skin. The action frequency refers to the peak value detection of the envelope in the action state: each muscle contraction usually forms an obvious peak on the envelope. The action frequency is calculated by counting the number of peaks that appear within a certain period of time, which is recorded as ; The base frequency is the preset frequency: 1 time / 15 seconds; Gaze Tracking , the standardized focus of attention determined by an eye tracker, ranging from 0 to 1.

3. The method for dynamic scheduling of sorting personnel based on biometric characteristics of sorting personnel according to claim 2, characterized in that: The preprocessing of the collected data includes: Heart rate using Kalman filter algorithm ,body temperature The time series data is denoised using the following formula: in, is the observed value, i.e., heart rate or body temperature , is the filtered value, is the Kalman gain; the heart rate after denoising Recorded as , body temperature after denoising Recorded as ; The constructed biometric vector is recorded as: 。 4. The method for dynamic scheduling of sorting personnel based on biometric characteristics of sorting personnel according to claim 3, characterized in that: The construction of a virtual mapping system for the ternary collaboration of humans, machines, and tasks to achieve predictive optimization and state-driven control of scheduling strategies includes: Each operator has a unique digital twin, which is dynamically modeled based on the following three types of information: (1) Static information layer: contains personnel ID, job type, maximum working time, and health records, which are used to initialize the basic parameter constraints of the twin model; (2) Dynamic state layer: real-time acquisition of biometric feature vectors , transmitted to the digital twin of the staff, and based on the edge computing gateway, the data of the staff digital twin is uploaded to the digital twin system of the sorting machine in real time; (3) Simulation behavior layer: Using the collected historical data, a time series prediction model based on the LSTM network is constructed: First, based on the historical information, multiple sets of sample data are constructed. Each set of sample data contains the states of two adjacent time windows. After the LSTM network is trained using multiple sets of samples, a trained LSTM network is obtained. Then, the state of the current time window is input into the trained LSTM network to obtain the state of the next time window. The state refers to the state at each moment. ; Based on the historical state of the most recent time window, input it into the trained LSTM model to obtain the state information in the next time window; Sorting task information: Predefine and clarify the sorting task details, including the sorter's position, the number of packages to be sorted, and the sorter status. Then, establish a task process model for the sorting task, including the interaction between the sorter and the sorter during the sorting process. Connect the task information to the sorter's digital twin model. The digital twin model of the sorting machine establishes an accurate three-dimensional model of the equipment based on the CAD model, kinematic and dynamic parameters of the actual equipment, and connects the physical equipment model with the real-time data update interface to achieve two-way mapping between the virtual environment and the physical equipment; the digital twin models and task information of the sorting machine and personnel need to have interfaces for synchronous data communication; the digital twin model of the sorting machine is used to accurately evaluate the current sorting task information, and by accessing the biological status data of the sorting personnel's digital twin, it obtains the sorting performance index, namely the package return rate.

5. The method for dynamic scheduling of sorting personnel based on biometric characteristics of sorting personnel according to claim 1, characterized in that: In the process of achieving comprehensive optimization of dynamic scheduling, dynamic adjustment is performed through the analytic hierarchy process, including: A1. Constructing a hierarchical model: The typesetting scheduling problem is divided into two layers: (1) The target layer achieves the optimal balance between scheduling efficiency and personnel health; (2) Criteria layer, including four core optimization objectives: , that is, the operating efficiency target, express; , that is, the health risk control target, express; , that is, the task urgency response target, express; , that is, the fatigue control target, express; A2. Based on the above four defined optimization objectives, perform pairwise comparison matrix ,in Indicates the importance of criterion i relative to criterion j; in, , i=1,2,3,4; j=1,2,3,4; where, Customize the settings based on experience. When i=j, ; A3. Assume that the dimension of the pairwise comparison matrix is ​​n. Since the values ​​of i and j are both integers from 1 to 4, n=4. Use the eigenvalue decomposition method to find the maximum eigenvalue of the comparison matrix. , used to measure the consistency of the comparison matrix; Calculating consistency index , calculate the consistency ratio , where RI is the consistency index. When dimension n=4, RI=0.

9. , then the judgment matrix is ​​considered to have good consistency and the weights are reliable and reasonable; on the contrary, if , it is considered that there is a large deviation in the judgment matrix, and it is revised , to improve consistency.

6. The method for dynamic scheduling of sorting personnel based on biometric characteristics of sorting personnel according to claim 1, characterized in that: Based on the established multi-objective optimization model, the multi-objective evolutionary algorithm NSGA-II is applied to the multi-objective scheduling problem, and structural optimization and heuristic enhancement are performed for logistics scenarios, including: To ensure the spatial continuity and temporal controllability of task allocation, a two-dimensional matrix structure is used to encode chromosomes and represent the scheduling scheme: ,in is the number of sorting areas, is the number of scheduling time periods, is the sorting personnel id of the i-th sorting area and the j-th scheduling time period; for Personnel gathering: Each person cannot be assigned to multiple areas at the same time; Take each scheduling scheme as an individual, and all individuals constitute the population; In order to improve population diversity and avoid premature convergence, the following adaptive probability adjustment formula is designed: Crossover probability : Mutation probability : in, is the fitness of the current individual, using multi-objective weighted values, namely ; is the maximum and average fitness of the current population; a trace constant to prevent division by zero; are the upper and lower bounds of the crossover probability, are the upper and lower limits of the mutation probability; For the goal of operational efficiency, For health risk control goals, To respond to the mission urgency target, Set a target for fatigue control; Apply the multi-objective evolutionary algorithm NSGA-II to the multi-objective scheduling problem and solve and output the final scheduling measurement solution .

Images