Examination and approval task allocation method and device

Abstract

The embodiment of the invention provides an approval task allocation method and device. The method comprises the steps of obtaining current approval task data of a current approval task; obtaining current approver data of a current approver corresponding to the current approval task; inputting the current examination and approval task data and the current examination and approval personnel data into a preset completion duration prediction model, and predicting examination and approval completion time of the current examination and approval personnel on the current examination and approval task; and according to the examination and approval completion time and the current examination and approval task data, determining a current examination and approval task allocation scheme, the current examination and approval task allocation scheme being the current examination and approval task allocated to the current examination and approval personnel. According to the embodiment of the invention, the examination and approval completion time is predicted based on the completion time prediction model, and the current examination and approval task is allocated to the corresponding appropriate current examination and approval personnel based on the examination and approval completion time and the current examination and approval task, so that the examination and approval efficiency and examination and approval quality of the examination and approval task are improved.

Description

Technical Field

[0001] Embodiments of the present invention relate to the field of IT and software development technology, and in particular, to a method for allocating approval tasks, an apparatus for allocating approval tasks, an electronic device, and a computer-readable storage medium. Background Art

[0002] In traditional workflow systems, the allocation of approval tasks usually follows preset approval personnel allocation rules or relies on static allocation methods such as manual processing.

[0003] However, as the business continues to grow and change, this static allocation method has gradually shown its limitations. For example, when the approver takes leave, the approval time is too long, or the workload is overloaded, the workflow system lacks effective response measures and cannot adjust the allocation of approval tasks in time, resulting in delays or interruptions in the approval process. In addition, relying on manual allocation of approval tasks also has the problem of inefficiency: the allocation personnel of approval tasks need to spend a lot of time to understand the current work status and professional capabilities of each approver, as well as the specific requirements of each approval task. This allocation method is not only time-consuming and labor-intensive, but also due to the intervention of human factors, there may be risks of unfair or incorrect allocation, resulting in low approval efficiency and quality of approval tasks.

[0004] Therefore, it is urgent to propose an allocation plan for approval tasks to reasonably allocate approval tasks and improve the approval efficiency and quality of approval tasks. Summary of the invention

[0005] The purpose of the embodiment of the present invention is to provide a method for allocating approval tasks to solve the problem that the approval tasks cannot be reasonably allocated, resulting in low approval efficiency and approval quality. The specific technical solution is as follows:

[0006] In a first aspect of the present invention, a method for allocating approval tasks is provided, the method comprising:

[0007] Get the current approval task data of the current approval task;

[0008] Obtaining the current approver data of the current approver corresponding to the current approval task;

[0009] Input the current approval task data and the current approval personnel data into a preset completion time prediction model to predict the approval completion time of the current approval personnel for the current approval task;

[0010] A current approval task allocation scheme is determined according to the approval completion time and the current approval task data, wherein the current approval task allocation scheme is the current approval task allocated to the current approval personnel.

[0011] In one embodiment of the present invention, before inputting the current approval task data and the current approval personnel data into a preset completion time prediction model to predict the approval completion time of the current approval personnel for the current approval task, the method further includes:

[0012] Acquire approval personnel data and approval task data from the approval database; wherein the approval task data includes the actual approval completion time;

[0013] Build a neural network model;

[0014] The neural network model is trained using the approval personnel data and the approval task data to obtain a training approval completion time;

[0015] Calculating the loss value of the neural network model using a preset loss function according to the training approval completion time and the actual approval completion time;

[0016] When the loss value meets the preset convergence condition, the neural network model is used as the completion time prediction model.

[0017] In an embodiment of the present invention, the current approval task data at least includes a priority, and according to the approval completion time and the current approval task data, determining the current approval task allocation scheme includes:

[0018] Obtaining a random approval task allocation scheme; the random approval task allocation scheme is the current approval task randomly allocated to the current approval personnel, the random approval task allocation scheme includes a sub-allocation scheme, and the sub-allocation scheme is the current approval personnel corresponding to the current approval task;

[0019] Inputting the approval completion time and the priority into a preset fitness function to obtain the fitness of the sub-allocation scheme;

[0020] Determining a first target sub-allocation scheme from the sub-allocation schemes according to the fitness by a roulette wheel method;

[0021] Taking the sub-allocation scheme with the largest fitness as the second target sub-allocation scheme;

[0022] Performing a mutation operation on the random approval task allocation scheme to obtain a mutated approval task allocation scheme; the mutated approval task allocation scheme includes the first target sub-allocation scheme and the second target sub-allocation scheme;

[0023] Using the variant approval task allocation scheme as a random approval task allocation scheme, and returning to the step of determining the random approval task allocation scheme;

[0024] When the step of obtaining a random approval task allocation scheme is executed for a preset number of iterations, the sub-allocation schemes with the largest fitness in the iteration process are combined into the current approval task allocation scheme.

[0025] In an embodiment of the present invention, the preset number of iterations is determined according to the number of the approval tasks, the number of the approval personnel, and preset performance parameters.

[0026] In an embodiment of the present invention, after determining a current approval task allocation scheme according to the approval completion time and the current approval task data, wherein the current approval task allocation scheme is the current approval task allocated to the current approval personnel, the method further comprises:

[0027] Real-time monitoring of the approval progress and work status of the current approver;

[0028] The current approval task allocation plan is re-determined according to the approval progress and the work status.

[0029] In an embodiment of the present invention, the approval database is updated at a preset frequency, and a current approval task allocation scheme is determined according to the approval completion time and the current approval task data, and the current approval task allocation scheme is the current approval task allocated to the current approval personnel. After that, the method further includes:

[0030] When the preset optimization time is reached, the approval personnel data and the approval task data are obtained from the approval database;

[0031] The approval personnel data and the approval task data are used to update the completion time prediction model.

[0032] In one embodiment of the present invention, the characteristic factors of the approval personnel data include at least the professional field, historical approval performance and workload of the approval personnel, and the characteristic factors of the approval task data include at least the urgency, complexity, priority and actual completion time of the approval task; the neural network model includes input layers and hidden layers, the number of input layers corresponds to the number of characteristic factors in the approval database, and the number of hidden layers corresponds to the number of input layers.

[0033] The embodiment of the present invention also discloses a device for allocating approval tasks, the device comprising:

[0034] The approval task data acquisition module is used to obtain the current approval task data of the current approval task;

[0035] An approving personnel data acquisition module is used to acquire the current approving personnel data of the current approving personnel corresponding to the current approval task;

[0036] An approval completion time prediction module is used to input the current approval task data and the current approval personnel data into a preset completion time prediction model to predict the approval completion time of the current approval personnel for the current approval task;

[0037] The allocation scheme determination module is used to determine the current approval task allocation scheme according to the approval completion time and the current approval task data. The current approval task allocation scheme is the current approval task allocated to the current approval personnel.

[0038] In another aspect of the present invention, a computer-readable storage medium is provided, wherein instructions are stored in the computer-readable storage medium, and when the computer-readable storage medium is executed on a computer, the computer executes any of the above-mentioned approval task allocation methods.

[0039] In another aspect of the present invention, a computer program product comprising instructions is provided, which, when executed on a computer, enables the computer to execute any of the above-mentioned methods for allocating approval tasks.

[0040] Compared with the related art, the embodiments of the present invention have at least the following advantages:

[0041] In an embodiment of the present invention, the current approval task data of the current approval task is obtained, and the current approval personnel data of the current approval personnel corresponding to the current approval task is obtained, and the current approval task data and the current approval personnel data are input into a preset completion time prediction model to predict the approval completion time of the current approval personnel for the current approval task, and then, the current approval task allocation scheme can be determined according to the approval completion time and the current approval task data to complete the allocation of the current approval task. The embodiment of the present invention predicts the approval completion time based on the completion time prediction model, and allocates the current approval task to the corresponding appropriate current approval personnel based on the approval completion time and the current approval task, thereby improving the automation level of approval task allocation, reducing manual intervention, and improving the approval efficiency and approval quality of the approval task. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art.

[0043] Figure 1 A flowchart of a method for allocating approval tasks provided in an embodiment of the present invention;

[0044] Figure 2 A structural block diagram of an approval task allocation device provided in an embodiment of the present invention;

[0045] Figure 3 The present invention is a block diagram of an electronic device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0046] The technical solutions in the embodiments of the present invention will be described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0047] One of the core concepts of the embodiments of the present invention is to provide an intelligent allocation method for approval tasks based on neural networks and genetic algorithms, which can provide an efficient, intelligent and adaptive approval task allocation solution to meet the growing demand for approval efficiency and approval quality in modern workflows.

[0048] Reference Figure 1 , is a flowchart of a method for allocating approval tasks provided in an embodiment of the present invention, such as Figure 1 As shown, the method may specifically include the following steps:

[0049] Step 101, obtaining the current approval task data of the current approval task;

[0050] Step 102: Acquire the current approver data of the current approver corresponding to the current approval task;

[0051] In the specific implementation, the workflow (approval task) is an abstract and general description of the business rules between the workflow and its various operation steps, such as the approval task for video and other businesses. There are usually multiple approval tasks, or one. The approver is the staff member who handles the approval task, and there are usually multiple approvers, or one approver.

[0052] In an embodiment of the present invention, the approval task that needs to be processed currently is the current approval task, and the approval personnel who processes the current approval task is the current approval personnel. An intelligent scheduling system can be deployed in the machine. When there are multiple current approval tasks that need to be assigned to the current approval personnel, the intelligent scheduling system obtains the current approval task data (approval task data) of the current approval task, and obtains the current approval personnel data (approval personnel data) of the current approval personnel corresponding to the current approval task. Specifically, the characteristic factors of the approval task data may include but are not limited to the urgency, complexity, priority, and actual completion time of the approval task. The characteristic factors of the approval personnel data may include but are not limited to the professional field, historical approval performance, and workload of the approval personnel.

[0053] Step 103: input the current approval task data and the current approval personnel data into a preset completion time prediction model to predict the approval completion time of the current approval personnel for the current approval task;

[0054] Among them, the completion time prediction model can be pre-deployed in the intelligent scheduling system, and the completion time prediction model can be a neural network model. Specifically, the neural network model is a computational model that simulates the neural network structure of the human brain. The neural network model consists of multiple layers, including input layer, hidden layer and output layer. Each layer contains multiple neurons (nodes). The neural network model automatically adjusts the weights and biases in the network by learning the relationship between input data and output data, thereby realizing the prediction and classification of unknown data.

[0055] In an embodiment of the present invention, historical approval task data and approval personnel data are used to train a completion time prediction model, so that the completion time prediction model can predict the approval completion time required for the current approval personnel to handle the current approval task based on the input current approval task data and current approval personnel data.

[0056] Step 104: Determine a current approval task allocation plan based on the approval completion time and the current approval task data. The current approval task allocation plan is the current approval task allocated to the current approval personnel.

[0057] Among them, the current approval task allocation plan refers to allocating the current approval task to the appropriate current approval personnel. In the embodiment of the present invention, the current approval task data includes the urgency, complexity, priority and other information of the current approval task. Therefore, according to the approval completion time predicted by the completion time prediction model and the current approval task data of the current approval task, the current approval task allocation plan is determined, so that the current approval task can be reasonably allocated to the current approval personnel for processing, so that the current approval task can be completed efficiently and accurately.

[0058] For example, assuming that there are current approval tasks A1, A2, and A3, and current approvers B1, B2, and B3; assuming that the urgency, complexity, and priority of current approval task A1 are higher than those of current approval task A2 and A3, and the urgency, complexity, and priority of current approval task A2 are higher than those of current approval task A3; the professional field, historical approval performance, and workload of current approver B3 are better than those of current approver B1 and B2, and the professional field, historical approval performance, and workload of current approver B2 are better than those of current approver B1; then the allocation plan for the current approval task can be determined as assigning the current approval task A1 to current approver B3, the current approval task A2 to current approver B2, and the current approval task A3 to current approver B1, so that the current approval task is assigned to a suitable current approver, thereby improving the approval efficiency and quality of the current approval task.

[0059] It should be noted that the embodiments of the present invention may involve the use of user data. In actual applications, user-specific personal data can be used in the scheme described herein within the scope permitted by applicable laws and regulations, subject to the requirements of applicable laws and regulations of the country where the user is located (for example, with the user's explicit consent, effective notification to the user, etc.).

[0060] In the above-mentioned approval task allocation method, the current approval task data of the current approval task is obtained, and the current approval personnel data of the current approval personnel corresponding to the current approval task is obtained, and the current approval task data and the current approval personnel data are input into the preset completion time prediction model to predict the approval completion time of the current approval personnel for the current approval task, and then, the current approval task allocation scheme can be determined according to the approval completion time and the current approval task data to complete the allocation of the current approval task. The embodiment of the present invention predicts the approval completion time based on the completion time prediction model, and allocates the current approval task to the corresponding appropriate current approval personnel based on the approval completion time and the current approval task, thereby improving the automation level of approval task allocation, reducing manual intervention, and improving the approval efficiency and approval quality of the approval task.

[0061] In one embodiment of the present invention, before inputting the current approval task data and the current approval personnel data into a preset completion time prediction model to predict the approval completion time of the current approval personnel for the current approval task, the method further includes:

[0062] Acquire approval personnel data and approval task data from the approval database; wherein the approval task data includes the actual approval completion time;

[0063] Build a neural network model;

[0064] The neural network model is trained using the approval personnel data and the approval task data to obtain a training approval completion time;

[0065] Calculating the loss value of the neural network model using a preset loss function according to the training approval completion time and the actual approval completion time;

[0066] When the loss value meets the preset convergence condition, the neural network model is used as the completion time prediction model.

[0067] In the embodiment of the present invention, an approval database is constructed that includes features such as the professional field of the approval personnel, historical approval performance, workload, etc., as well as features such as the urgency, complexity, priority, and actual completion time of the approval task. When it is necessary to train a completion time prediction model for predicting the approval completion time, a neural network model is constructed, and the approval personnel data and approval task data are converted into features as input data to construct a completion time prediction model for the approval personnel for the approval task.

[0068] Exemplarily, access is made to a company's existing OA workflow database to collect the approver data and approval task data of the workflow tasks (approval tasks). The collection frequency may be once a day or once a week. The approver data and approval task data may specifically include the type of approval task, its importance, complexity, the approver's professional field, position, workload, the approver's degree of completion of each approval task, completion time, and other data.

[0069] After obtaining the data of the approver data and the approval task data, the data is cleaned and preprocessed to obtain the workflow data training set and the approver data training set. Feature extraction is performed on the workflow data training set and the approver data training set to obtain the workflow data feature set and the approver data feature set. For example, the feature factors of the workflow data feature set can be: workflow category, process complexity, business complexity, etc. The feature factors of the approver data feature set can be: professional field, position, rank, workload, organizational team, etc.

[0070] In a specific example, the workflow data feature set and the reviewer data feature set are used as input data to construct a neural network model, wherein the number of nodes in the input layer of the neural network model corresponds to the number of selected feature factors. Two hidden layers are added to the neural network model, and the number of hidden layers can be set to twice the number of input layers based on the complexity of the problem.

[0071] In a specific example, the mean square error MSE (Mean Squared Error) can be selected as the loss function of the neural network model, and Adam (an optimization algorithm with adaptive learning rate) can be selected as the optimization algorithm of the neural network model.

[0072] In some examples, the mean square error MSE may be:

[0073]

[0074] Where n is the number of samples, y i is the true value of the i-th sample (actual approval completion time), y p red iis the predicted value of the i-th sample (training approval completion time).

[0075] In some examples, the Adam optimization algorithm is:

[0076]

[0077] After the neural network model is constructed, historical data can be obtained to train the neural network model, where the historical data can be the approval personnel data and approval task data obtained from the approval database, and the weight of the neural network model is adjusted through the back propagation algorithm to minimize the prediction error. Among them, minimizing the prediction error means adjusting the weight of the neural network model through optimization algorithms, such as gradient descent, Adam and other optimization algorithms, so that the loss value output by the loss function according to the actual approval completion time and the training approval completion time, such as the loss value output by the mean square error MSE, is as small as possible. When the loss value output by the mean square error MSE is as small as possible, it can be regarded as meeting the convergence condition of the neural network model, and the neural network model at this time can be used as a completion time prediction model to predict the approval completion time of the approval personnel handling the approval task.

[0078] In an embodiment of the present invention, the current approval task data at least includes a priority, and according to the approval completion time and the current approval task data, determining the current approval task allocation scheme includes:

[0079] Obtaining a random approval task allocation scheme; the random approval task allocation scheme is the current approval task randomly allocated to the current approval personnel, the random approval task allocation scheme includes a sub-allocation scheme, and the sub-allocation scheme is the current approval personnel corresponding to the current approval task;

[0080] Inputting the approval completion time and the priority into a preset fitness function to obtain the fitness of the sub-allocation scheme;

[0081] Determining a first target sub-allocation scheme from the sub-allocation schemes according to the fitness by a roulette wheel method;

[0082] Taking the sub-allocation scheme with the largest fitness as the second target sub-allocation scheme;

[0083] Performing a mutation operation on the random approval task allocation scheme to obtain a mutated approval task allocation scheme; the mutated approval task allocation scheme includes the first target sub-allocation scheme and the second target sub-allocation scheme;

[0084] Using the variant approval task allocation scheme as a random approval task allocation scheme, and returning to the step of determining the random approval task allocation scheme;

[0085] When the step of obtaining a random approval task allocation scheme is executed for a preset number of iterations, the sub-allocation schemes with the largest fitness in the iteration process are combined into the current approval task allocation scheme.

[0086] In an embodiment of the present invention, based on the predicted completion time of the current approval task by the current approval personnel, the approval personnel are intelligently allocated through a genetic algorithm. Among them, the genetic algorithm (GA) is designed and proposed based on the evolution law of organisms in nature. It is a computational model of the biological evolution process that simulates the natural selection and genetic mechanism of Darwin's theory of biological evolution. It is a method for searching for the optimal solution by simulating the natural evolution process. The genetic algorithm uses a mathematical method and computer simulation operation to convert the problem-solving process into a process similar to the crossover and mutation of chromosome genes in biological evolution. When solving more complex combinatorial optimization problems, it can usually obtain better optimization results faster than some conventional optimization algorithms. Genetic algorithms have been widely used in combinatorial optimization, machine learning, signal processing, adaptive control and artificial life.

[0087] In the embodiment of the present invention, the workflow set of the current approval personnel to be assigned and the set of the current approver pool are obtained, that is, the approval task data of the current approval task and the approval personnel data of the current approval personnel are obtained, and the approval completion time predicted by each approval personnel for different approval personnel is calculated based on the training completion time prediction model, and the fitness value is calculated according to the approval completion time and the priority of the current approval task. Based on the characteristics of the approval scenario, the fitness function of the population (for the current approval personnel and the random approval task allocation scheme for the current approval personnel) is set as:

[0088] f=∑t×α

[0089] Among them, t is the approval time (the approval completion time predicted by the completion time prediction model for each approver to different approvers), and α is the priority of the workflow. The individual (chromosome) with the smaller calculated fitness in the population is optimal. Among them, the individual refers to the sub-allocation scheme of the population. For example, the first current approval task is assigned to the current approver 1, which can be regarded as an individual in the population.

[0090] The initial population realizes the allocation of approval tasks by randomly generating a random approval task allocation scheme, that is, the current approval task can be randomly assigned to the current approver to generate it. For example, assuming that there are 10 current approval tasks and there are 15 current approvers in the approver pool, the random approval task allocation scheme can be: 1,5,14,3,8,10,4,11,9,10, that is, the first current approval task is assigned to the current approver 1, the second current approval task is assigned to the current approver 5, the third current approval task is assigned to the current approver 14, and so on.

[0091] The chromosomes are selected according to the fitness score by means of roulette, and the chromosomes with high scores have a greater probability of being selected for reproduction, that is, the first target sub-allocation scheme is selected from the sub-allocation schemes in the random approval task allocation scheme according to the fitness. Among them, roulette is a selection operation method used in genetic algorithms. Roulette selects individuals for reproduction according to their fitness by simulating roulette. Among them, the higher the fitness of the individual, the greater the probability of being selected.

[0092] Ensure that the best chromosomes in each generation can directly enter the next generation, that is, the sub-allocation scheme with the largest fitness corresponding to the sub-allocation scheme in the random approval task allocation scheme is used as the second target sub-allocation scheme.

[0093] Next, the random approval task allocation scheme (population) can be mutated with a certain probability to obtain a mutated approval task allocation scheme, which increases the diversity of the population. Approval personnel from different populations can be selected to exchange positions to avoid premature convergence. It should be noted that the mutated approval task allocation scheme includes the first target sub-allocation scheme and the second target sub-allocation scheme.

[0094] In one embodiment of the present invention, the preset number of iterations of the genetic algorithm is determined according to the number of the approval tasks, the number of the approval personnel, and the preset performance parameters. Exemplarily, the number of iterations is: the number of workflows*the number of approval personnel*a, where a is the performance parameter preset according to the current machine.

[0095] The variant approval task allocation scheme is used as the random approval task allocation scheme, and the step of determining the random approval task allocation scheme is returned, that is, the above steps are repeated for iteration. When the step of obtaining the random approval task allocation scheme reaches the preset number of iterations, the sub-allocation scheme with the largest fitness in the entire iteration process is combined into the current approval task allocation scheme. For example, the random approval task allocation scheme can be: 1, 5, 14, 3, 8, 10, 4, 11, 9, 10, and the current approval task allocation scheme determined by the genetic algorithm can be 2, 5, 4, 7, 15, 1, 3, 9, 8, 10, which satisfies the intelligent allocation of approval tasks for the approvers.

[0096] In an embodiment of the present invention, after determining a current approval task allocation scheme according to the approval completion time and the current approval task data, wherein the current approval task allocation scheme is the current approval task allocated to the current approval personnel, the method further comprises:

[0097] Real-time monitoring of the approval progress and work status of the current approver;

[0098] The current approval task allocation plan is re-determined according to the approval progress and the work status.

[0099] In the embodiment of the present invention, after the current approval task allocation scheme is determined, the current approval task can be allocated to the corresponding current approval personnel for processing based on the current approval task allocation scheme. At this time, the approval progress of the current approval personnel in processing the current approval task and the working status of the current approval personnel can be monitored in real time. Through the approval progress and the working status of the current approval personnel, the allocation of approval tasks can be dynamically adjusted to cope with emergencies and changes.

[0100] Specifically, if the approval progress of the current approver is delayed for a certain period of time, a delay reminder will be triggered. In the case of further delay, the system will reallocate the current approval task assigned to the current approver, and increase the priority of the current approval task of the current approver to avoid the current approval task of the current approver being unprocessed for a long time. The time for triggering the delay processing can be preset according to different approval tasks.

[0101] In the above embodiment, when an abnormal situation occurs during the processing of the current approval task, such as the absence of the current approval personnel or the backlog of approval tasks, the intelligent scheduling system can quickly reallocate the current approval task to minimize the impact on the current approval task in the entire approval process.

[0102] In an embodiment of the present invention, the approval database is updated at a preset frequency, and a current approval task allocation scheme is determined according to the approval completion time and the current approval task data, and the current approval task allocation scheme is the current approval task allocated to the current approval personnel. After that, the method further includes:

[0103] When the preset optimization time is reached, the approval personnel data and the approval task data are obtained from the approval database;

[0104] The approval personnel data and the approval task data are used to update the completion time prediction model.

[0105] In an embodiment of the present invention, when the preset optimization time is reached, the completion time prediction model and the genetic algorithm model can be continuously optimized based on the continuously updated approval personnel data and approval task data in the approval database, so as to continuously improve the approval efficiency and approval quality of the intelligent scheduling system.

[0106] In order to enable those skilled in the art to better understand the embodiments of the present invention, a complete and specific embodiment is used below to illustrate the implementation process of the method for intelligently reallocating approval tasks of the present invention:

[0107] 1. Access the company's existing OA workflow database to collect workflow data (approval task data) and approval personnel data of workflow tasks. The collection frequency can be once a day. Among them, the approval personnel data and approval task data can specifically include the type of approval tasks, importance, complexity, professional field of the approval personnel, position, workload, completion degree of the approval personnel for each approval task, completion time and other data.

[0108] 2. Clean and preprocess the data to obtain the workflow data training set and the approver data training set.

[0109] 3. Perform feature extraction on the workflow data training set and the approval personnel data training set to obtain the workflow data feature set and the approval personnel data feature set. For example, the feature factors of the workflow data feature set may be: workflow category, process complexity, business complexity, etc. The feature factors of the approval personnel data feature set may be: professional field, position, rank, workload, organizational team, etc.

[0110] 4. Use the workflow data feature set and the approval personnel data feature set to train the neural network model and obtain the approval personnel's completion time prediction model for the approval task.

[0111] The specific steps for building a completion time prediction model for approval tasks are as follows:

[0112] 1. Take the workflow data feature set and the approval personnel data feature set as input data to build a neural network model. The number of nodes in the input layer of the neural network model corresponds to the number of selected feature factors.

[0113] 2. Add two hidden layers to the neural network model. Based on the complexity of the problem, set the number of hidden layers to twice the number of input layers.

[0114] 3. Select mean square error (MSE) as the loss function and Adam as the optimization algorithm.

[0115]

[0116] Where n is the number of samples, yi is the true value of the i-th sample, y p red i is the predicted value of the ith sample.

[0117] Among them, the Adam optimization algorithm is:

[0118]

[0119] 4. Use historical data (such as approval personnel data and approval task data obtained from the approval database) to train the neural network model, adjust the weight of the neural network model through the back propagation algorithm, minimize the prediction error, and when the neural network model reaches the preset convergence conditions, obtain the training completion time prediction model. The completion time prediction model can be used to predict the approval personnel's predicted completion time for the approval task.

[0120] 5. Based on the predicted completion time of the workflow by the approvers predicted by the completion time prediction model, the approvers are intelligently allocated through genetic algorithms.

[0121] Among them, the specific steps of constructing the genetic algorithm are:

[0122] 1. Get the workflow set of the current approver to be assigned and the set of the current approver pool, that is, get the approval task data of the current approval task and the approver data of the current approver.

[0123] 2. Based on the above neural network model, i.e., the completion time prediction model, the predicted completion time of each approver in the set under different workflows (approval tasks) is calculated, and the fitness value is calculated according to the predicted completion time and the current approval task priority. Based on the characteristics of the approval scenario, the fitness function of the population corresponding to the approval scenario is set as:

[0124] f = ∑t × α, where t is the approval time and α is the workflow priority. The individuals (chromosomes) with the smaller calculated fitness in the population are optimal. The population can be a random approval task allocation scheme, or a sub-allocation scheme in a random approval task allocation scheme.

[0125] 3. The initial population generates the current approval task allocation plan by randomly assigning the current approval tasks to the approvers. For example, if there are 10 current approval tasks and the current approver pool is 15, the current approval task allocation plan can be: 1, 5, 14, 3, 8, 10, 4, 11, 9, 10.

[0126] 4. Select chromosomes according to fitness through roulette (i.e., screen the first target sub-allocation scheme). Chromosomes with high fitness have a greater probability of being selected for reproduction, that is, the random approval task allocation scheme generated in the next iteration includes individuals screened through roulette.

[0127] 5. Ensure that the best chromosomes in each generation can directly enter the next generation (i.e., screen the second target sub-allocation scheme), that is, the random approval task allocation scheme generated in the next iteration includes the individual with the largest fitness.

[0128] 6. Perform mutation operations on the random approval task allocation scheme with a certain probability to increase the diversity of the population. Approval personnel from different populations can be selected to exchange positions to avoid premature convergence.

[0129] 7. Set the number of iterations of the genetic algorithm to: number of workflows * number of reviewers * a, where a is the performance parameter preset according to the current machine.

[0130] Then, the final output result in step 5, that is, the current approval task allocation plan can be: 2, 5, 4, 7, 15, 1, 3, 9, 8, 10., completing the intelligent allocation of the current approval task.

[0131] 5. After the current approval task is assigned to the corresponding current approver according to the current approval task allocation plan, the approval progress and work status of the current approver are monitored in real time. If the approval progress of the current approver is delayed for a certain period of time, an extension reminder will be triggered to promptly handle the assigned current approval task. In the case of further delay, the intelligent scheduling system will reallocate the current approval task assigned to the current approver, and the priority of the current approval task assigned to the current approver will be increased.

[0132] In summary, an embodiment of the present invention provides a method for automatic reallocation of approval tasks, which can monitor the working status of approval personnel and approval tasks that need to be approved in real time, and automatically allocate approval tasks to the most suitable approval personnel according to preset performance indicators and priority rules, thereby improving the approval efficiency and approval quality of approval tasks.

[0133] It should be noted that, for the sake of simplicity, the method embodiments are described as a series of action combinations, but those skilled in the art should be aware that the embodiments of the present invention are not limited by the order of the actions described, because according to the embodiments of the present invention, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions involved are not necessarily required by the embodiments of the present invention.

[0134] Reference Figure 2 , is a structural block diagram of an approval task allocation device provided in an embodiment of the present invention, such as Figure 2 As shown, the device may specifically include the following modules:

[0135] The approval task data acquisition module 201 is used to acquire the current approval task data of the current approval task;

[0136] The approval personnel data acquisition module 202 is used to acquire the current approval personnel data of the current approval personnel corresponding to the current approval task;

[0137] The approval completion time prediction module 203 is used to input the current approval task data and the current approval personnel data into a preset completion time prediction model to predict the approval completion time of the current approval personnel for the current approval task;

[0138] The allocation scheme determining module 204 is used to determine the current approval task allocation scheme according to the approval completion time and the current approval task data. The current approval task allocation scheme is the current approval task allocated to the current approval personnel.

[0139] In one embodiment of the present invention, the device further includes: a completion time prediction model training module, which is used to:

[0140] Acquire approval personnel data and approval task data from the approval database; wherein the approval task data includes the actual approval completion time;

[0141] Build a neural network model;

[0142] The neural network model is trained using the approval personnel data and the approval task data to obtain a training approval completion time;

[0143] Calculating the loss value of the neural network model using a preset loss function according to the training approval completion time and the actual approval completion time;

[0144] When the loss value meets the preset convergence condition, the neural network model is used as the completion time prediction model.

[0145] In one embodiment of the present invention, the current approval task data at least includes a priority, and the allocation scheme determination module 204 is used to include:

[0146] Obtaining a random approval task allocation scheme; the random approval task allocation scheme is the current approval task randomly allocated to the current approval personnel, the random approval task allocation scheme includes a sub-allocation scheme, and the sub-allocation scheme is the current approval personnel corresponding to the current approval task;

[0147] Inputting the approval completion time and the priority into a preset fitness function to obtain the fitness of the sub-allocation scheme;

[0148] Determining a first target sub-allocation scheme from the sub-allocation schemes according to the fitness by a roulette wheel method;

[0149] Taking the sub-allocation scheme with the largest fitness as the second target sub-allocation scheme;

[0150] Performing a mutation operation on the random approval task allocation scheme to obtain a mutated approval task allocation scheme; the mutated approval task allocation scheme includes the first target sub-allocation scheme and the second target sub-allocation scheme;

[0151] Using the variant approval task allocation scheme as a random approval task allocation scheme, and returning to the step of determining the random approval task allocation scheme;

[0152] When the step of obtaining a random approval task allocation scheme is executed for a preset number of iterations, the sub-allocation schemes with the largest fitness in the iteration process are combined into the current approval task allocation scheme.

[0153] In an embodiment of the present invention, the preset number of iterations is determined based on the number of the approval tasks, the number of the approval personnel, and preset performance parameters.

[0154] In one embodiment of the present invention, the apparatus further comprises: a reallocation module, configured to:

[0155] Real-time monitoring of the approval progress and work status of the current approver;

[0156] The current approval task allocation plan is re-determined according to the approval progress and the work status.

[0157] In one embodiment of the present invention, the device further includes: a completion duration prediction model updating module, configured to:

[0158] When the preset optimization time is reached, the approval personnel data and the approval task data are obtained from the approval database;

[0159] The approval personnel data and the approval task data are used to update the completion time prediction model.

[0160] In one embodiment of the present invention, the characteristic factors of the approval personnel data include at least the professional field, historical approval performance and workload of the approval personnel, and the characteristic factors of the approval task data include at least the urgency, complexity, priority and actual completion time of the approval task; the neural network model includes input layers and hidden layers, the number of input layers corresponds to the number of characteristic factors in the approval database, and the number of hidden layers corresponds to the number of input layers.

[0161] In an embodiment of the present invention, the current approval task data of the current approval task is obtained, and the current approval personnel data of the current approval personnel corresponding to the current approval task is obtained, and the current approval task data and the current approval personnel data are input into a preset completion time prediction model to predict the approval completion time of the current approval personnel for the current approval task, and then, the current approval task allocation scheme can be determined according to the approval completion time and the current approval task data to complete the allocation of the current approval task. The embodiment of the present invention predicts the approval completion time based on the completion time prediction model, and allocates the current approval task to the corresponding appropriate current approval personnel based on the approval completion time and the current approval task, thereby improving the automation level of approval task allocation, reducing manual intervention, and improving the approval efficiency and approval quality of the approval task.

[0162] As for the above-mentioned device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0163] The embodiment of the present invention further provides an electronic device, such as Figure 3 As shown, it includes a processor 501, a communication interface 502, a memory 503 and a communication bus 504, wherein the processor 501, the communication interface 502, and the memory 503 communicate with each other through the communication bus 504.

[0164] Memory 503, used for storing computer programs;

[0165] The processor 501 is used to implement any of the approval task allocation methods described in the above embodiments when executing the program stored in the memory 503:

[0166] The communication bus mentioned in the above terminal can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The communication bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0167] The communication interface is used for communication between the above terminal and other devices.

[0168] The memory may include a random access memory (RAM) or a non-volatile memory, such as at least one disk memory. Optionally, the memory may also be at least one storage device located away from the aforementioned processor.

[0169] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0170] In another embodiment provided by the present invention, a computer-readable storage medium is provided, in which instructions are stored. When the computer-readable storage medium is executed on a computer, the computer executes the approval task allocation method described in any one of the above embodiments.

[0171] In another embodiment provided by the present invention, a computer program product including instructions is also provided, and when the computer is run on the computer, the computer is enabled to execute the approval task allocation method described in any one of the above embodiments.

[0172] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented by software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present invention is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that includes one or more available media integrated. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state hard disk Solid State Disk (SSD)), etc.

[0173] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or still includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "including one..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0174] Each embodiment in this specification is described in a related manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0175] The above description is only a preferred embodiment of the present invention and is not intended to limit the protection scope of the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention are included in the protection scope of the present invention.

Claims

1. A method for allocating approval tasks, characterized in that: The method comprises: Get the current approval task data of the current approval task; Obtaining the current approver data of the current approver corresponding to the current approval task; Input the current approval task data and the current approval personnel data into a preset completion time prediction model to predict the approval completion time of the current approval personnel for the current approval task; A current approval task allocation scheme is determined according to the approval completion time and the current approval task data, wherein the current approval task allocation scheme is the current approval task allocated to the current approval personnel.

2. The method according to claim 1, characterized in that Before inputting the current approval task data and the current approval personnel data into a preset completion time prediction model to predict the approval completion time of the current approval personnel for the current approval task, the method further includes: Acquire approval personnel data and approval task data from the approval database; wherein the approval task data includes the actual approval completion time; Build a neural network model; The neural network model is trained using the approval personnel data and the approval task data to obtain a training approval completion time; Calculating the loss value of the neural network model using a preset loss function according to the training approval completion time and the actual approval completion time; When the loss value meets the preset convergence condition, the neural network model is used as the completion time prediction model.

3. The method according to claim 1, characterized in that The current approval task data at least includes a priority. According to the approval completion time and the current approval task data, a current approval task allocation scheme is determined, including: Obtaining a random approval task allocation scheme; the random approval task allocation scheme is the current approval task randomly allocated to the current approval personnel, the random approval task allocation scheme includes a sub-allocation scheme, and the sub-allocation scheme is the current approval personnel corresponding to the current approval task; Inputting the approval completion time and the priority into a preset fitness function to obtain the fitness of the sub-allocation scheme; Determining a first target sub-allocation scheme from the sub-allocation schemes according to the fitness by a roulette wheel method; Taking the sub-allocation scheme with the largest fitness as the second target sub-allocation scheme; Performing a mutation operation on the random approval task allocation scheme to obtain a mutated approval task allocation scheme; the mutated approval task allocation scheme includes the first target sub-allocation scheme and the second target sub-allocation scheme; Using the variant approval task allocation scheme as a random approval task allocation scheme, and returning to the step of determining the random approval task allocation scheme; When the step of obtaining a random approval task allocation scheme is executed for a preset number of iterations, the sub-allocation schemes with the largest fitness in the iteration process are combined into the current approval task allocation scheme.

4. The method according to claim 3, characterized in that The preset number of iterations is determined according to the number of the approval tasks, the number of the approval personnel, and preset performance parameters.

5. The method according to claim 1, characterized in that After determining a current approval task allocation scheme according to the approval completion time and the current approval task data, wherein the current approval task allocation scheme is the current approval task allocated to the current approval personnel, the method further comprises: Real-time monitoring of the approval progress and work status of the current approver; The current approval task allocation plan is re-determined according to the approval progress and the work status.

6. The method according to claim 2, characterized in that The approval database is updated at a preset frequency, and a current approval task allocation scheme is determined according to the approval completion time and the current approval task data, wherein the current approval task allocation scheme is the current approval task allocated to the current approval personnel. The method further comprises: When the preset optimization time is reached, the approval personnel data and the approval task data are obtained from the approval database; The approval personnel data and the approval task data are used to update the completion time prediction model.

7. The method according to claim 2, characterized in that The characteristic factors of the approval personnel data include at least the professional field, historical approval performance and workload of the approval personnel, and the characteristic factors of the approval task data include at least the urgency, complexity, priority and actual completion time of the approval task; the neural network model includes input layers and hidden layers, the number of input layers corresponds to the number of characteristic factors in the approval database, and the number of hidden layers corresponds to the number of input layers.

8. A device for allocating approval tasks, characterized in that: The device comprises: The approval task data acquisition module is used to obtain the current approval task data of the current approval task; An approving personnel data acquisition module is used to acquire the current approving personnel data of the current approving personnel corresponding to the current approval task; An approval completion time prediction module is used to input the current approval task data and the current approval personnel data into a preset completion time prediction model to predict the approval completion time of the current approval personnel for the current approval task; The allocation scheme determination module is used to determine the current approval task allocation scheme according to the approval completion time and the current approval task data. The current approval task allocation scheme is the current approval task allocated to the current approval personnel.

9. An electronic device, characterized in that: It includes a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other through the communication bus; Memory, used to store computer programs; A processor, for implementing the method steps described in any one of claims 1 to 7 when executing a program stored in a memory.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images