Intelligent analysis method, device and equipment for work order card, and medium

Abstract

The invention provides an intelligent analysis method, device and equipment for work orders and cards, and a medium, and is used for solving the problems that the difference of processing time required by different links in the execution process of different types of work orders is neglected in the related technology, and the identification result of the cards is easily influenced by human subjective factors; the method comprises the steps that historical work order data are acquired, and the historical work order data comprise the residence time of different types of work orders in each link of a work order execution process; performing model iteration training on a preset neural network model on the basis of the historical work order data to obtain an order jamming duration threshold value of different types of work orders output by the trained neural network model in each link in a work order execution process; monitoring the stay duration of the to-be-executed work order in the link of the work order execution process; and if the stay duration exceeds an order jamming duration threshold value corresponding to the link where the to-be-executed work order is located, determining that the to-be-executed work order has an order jamming fault.

Description

Technical Field

[0001] The present invention relates to the field of operation and maintenance management, and in particular to an intelligent analysis method, device, equipment and medium for work order cards. Background Art

[0002] Due to the increase in the complexity of operation and maintenance, the operation and maintenance management scenarios have the phenomenon of lengthening the call chain between services, difficulty in service positioning, unstable calls between services, etc. In the process of carrying out business, work orders may be stuck due to various problems such as system errors and service delays.

[0003] Most existing work order stuck warning solutions are implemented through manual monitoring or label and text classification models. Manual monitoring is done by staff members subjectively judging whether a work order is stuck based on their experience, or by manually setting a threshold for the processing time of a work order. When the processing time of the work order exceeds the threshold, it is judged to be stuck. Classification models mostly use probability calculations, and their output results are easily affected by human factors. In addition, there is a lack of error correction learning mechanism in the classification model, and the training results of one iteration have large errors.

[0004] The existing solutions ignore the differences in processing time required for different types of work orders at different stages of the execution process, and the card order recognition results are easily affected by human subjective factors, resulting in low efficiency and accuracy of card order recognition. Summary of the invention

[0005] The present invention provides an intelligent analysis method, device, equipment and medium for work order and card order, so as to solve the technical problems that the related technology ignores the differences in processing time required for different links in the execution process of different types of work orders, and the card order recognition results are easily affected by human subjective factors, resulting in low efficiency and accuracy of card order recognition.

[0006] In a first aspect, an embodiment of the present invention provides an intelligent analysis method for a work order, the method comprising:

[0007] Obtain historical work order data, which includes the duration of different types of work orders at each stage of the work order execution process;

[0008] The preset neural network model is iteratively trained based on historical work order data to obtain the stuck time thresholds of different types of work orders in each link of the work order execution process output by the trained neural network model;

[0009] Monitor the length of time that pending work orders stay in the work order execution process;

[0010] If the duration of stay exceeds the stuck time threshold corresponding to the link where the work order to be executed is located, it is determined that a stuck fault has occurred in the work order to be executed.

[0011] In a possible implementation, in the method provided by the embodiment of the present invention, iterative model training is performed on a preset neural network model based on historical work order data, including:

[0012] Perform data labeling on historical work order data to label each work order data in the historical work order data as a stuck order state or a non-stuck order state to obtain a labeling result;

[0013] Perform data normalization on historical work order data, and divide the processed historical work order data into data sets including training set, test set and validation set;

[0014] The data set is used to perform iterative training on the preset neural network model.

[0015] In a possible implementation, in the method provided in the embodiment of the present invention, data labeling of historical work order data includes:

[0016] For each work order data in the historical work order data, the work order data is converted into data points according to the residence time of the work order data in each link of the work order execution process;

[0017] For each data point, determining a first number of other data points existing within a preset range around the data point;

[0018] If the proportion of the first quantity in the total number of data points is less than a preset proportion, the work order data corresponding to the data point is marked as a stuck state, otherwise the work order data corresponding to the data point is marked as a non-stuck state.

[0019] In a possible implementation manner, in the method provided by the embodiment of the present invention, for each data point, determining a first number of other data points existing within a preset range around the data point includes:

[0020] For each data point, the distance between the data point and other data points is calculated. The distance indicates the difference in the residence time of different work order data in each link of the work order execution process.

[0021] The number of target data points whose distance from the data point is less than the preset radius is taken as the first number.

[0022] In a possible implementation, in the method provided in the embodiment of the present invention, data normalization processing is performed on historical work order data, including:

[0023] Digitally encode the attribute features that need to be normalized in the historical work order data to obtain the feature sequence corresponding to the attribute features;

[0024] Each eigenvalue in the feature sequence is scaled according to the maximum and minimum values ​​in the feature sequence to obtain a normalized feature sequence.

[0025] In a possible implementation, in the method provided by the embodiment of the present invention, the preset neural network model includes a BP neural network;

[0026] Use the data set to iteratively train the preset neural network model, including:

[0027] Initialize the learning parameters and hyperparameters of the BP neural network;

[0028] Take the training set as input and perform forward propagation in the BP neural network to obtain the output result of the BP neural network;

[0029] Calculate the error between the output result and the labeled result, and back-propagate the error in the BP neural network to adjust the learning parameters of the BP neural network;

[0030] Repeat the above steps of forward propagation input and back propagation error until the BP neural network reaches a preset number of training rounds or the error meets a preset threshold, and obtain a trained BP neural network model;

[0031] Use the validation set to verify the performance of the trained BP neural network model to adjust the hyperparameters of the trained BP neural network model;

[0032] After the performance verification of the trained BP neural network model is completed, the trained BP neural network model is evaluated using the test set.

[0033] In a possible implementation, in the method provided by an embodiment of the present invention, the BP neural network includes a BP neural network with an additional momentum method.

[0034] In a second aspect, an embodiment of the present invention provides an intelligent analysis device for a work order, including:

[0035] An acquisition unit is used to acquire historical work order data, which includes the residence time of different types of work orders at various stages of the work order execution process;

[0036] A training unit is used to perform iterative model training on a preset neural network model based on historical work order data, so as to obtain the stuck time thresholds of different types of work orders in each link of the work order execution process output by the trained neural network model;

[0037] A monitoring unit is used to monitor the length of time that a work order to be executed stays in a link of the work order execution process;

[0038] The determination unit is used to determine that a stuck order failure occurs in the work order to be executed if the dwell time exceeds the stuck order time threshold corresponding to the link where the work order to be executed is located.

[0039] In a possible implementation manner, in the device provided by the embodiment of the present invention, the training unit is specifically used for:

[0040] Perform data labeling on historical work order data to label each work order data in the historical work order data as a stuck order state or a non-stuck order state to obtain a labeling result;

[0041] Perform data normalization on historical work order data, and divide the processed historical work order data into data sets including training set, test set and validation set;

[0042] The data set is used to perform iterative training on the preset neural network model.

[0043] In a possible implementation, in the device provided by the embodiment of the present invention, when the training unit performs data labeling on the historical work order data, it is specifically used to:

[0044] For each work order data in the historical work order data, the work order data is converted into data points according to the residence time of the work order data in each link of the work order execution process;

[0045] For each data point, determining a first number of other data points existing within a preset range around the data point;

[0046] If the proportion of the first quantity in the total number of data points is less than a preset proportion, the work order data corresponding to the data point is marked as a stuck state, otherwise the work order data corresponding to the data point is marked as a non-stuck state.

[0047] In a possible implementation manner, in the device provided by the embodiment of the present invention, when the training unit determines, for each data point, the first number of other data points existing within a preset range around the data point, the training unit is specifically configured to:

[0048] For each data point, the distance between the data point and other data points is calculated. The distance indicates the difference in the residence time of different work order data in each link of the work order execution process.

[0049] The number of target data points whose distance from the data point is less than the preset radius is taken as the first number.

[0050] In a possible implementation, in the device provided by the embodiment of the present invention, when the training unit performs data normalization processing on the historical work order data, it is specifically used to:

[0051] Digitally encode the attribute features that need to be normalized in the historical work order data to obtain the feature sequence corresponding to the attribute features;

[0052] Each eigenvalue in the feature sequence is scaled according to the maximum and minimum values ​​in the feature sequence to obtain a normalized feature sequence.

[0053] In a possible implementation, in the device provided by the embodiment of the present invention, the preset neural network model includes a BP neural network;

[0054] When the training unit uses the data set to perform model iterative training on the preset neural network model, it is specifically used to:

[0055] Initialize the learning parameters and hyperparameters of the BP neural network;

[0056] Take the training set as input and perform forward propagation in the BP neural network to obtain the output result of the BP neural network;

[0057] Calculate the error between the output result and the labeled result, and back-propagate the error in the BP neural network to adjust the learning parameters of the BP neural network;

[0058] Repeat the above steps of forward propagation input and back propagation error until the BP neural network reaches a preset number of training rounds or the error meets a preset threshold, and obtain a trained BP neural network model;

[0059] Use the validation set to verify the performance of the trained BP neural network model to adjust the hyperparameters of the trained BP neural network model;

[0060] After the performance verification of the trained BP neural network model is completed, the trained BP neural network model is evaluated using the test set.

[0061] In a possible implementation, in the device provided by an embodiment of the present invention, the BP neural network includes a BP neural network with an additional momentum method.

[0062] In a third aspect, an embodiment of the present invention provides an electronic device, comprising: at least one processor, at least one memory, and computer program instructions stored in the memory, which, when executed by the processor, implement the method provided in the first aspect of the embodiment of the present invention.

[0063] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having computer program instructions stored thereon, which, when executed by a processor, implements the method provided in the first aspect of the embodiment of the present invention.

[0064] In the embodiment of the present invention, historical work order data including the residence time of different types of work orders in each link of the work order execution process is first collected, and then the neural network model is iteratively trained using the historical data, so that the neural network model learns the characteristics of the residence time of different work orders in each link in the historical work order data, thereby determining the corresponding stuck order time thresholds of different types of work orders in each link through the neural network model, and finally monitoring whether the residence time of the work order to be executed in the link of the work order execution process exceeds its corresponding stuck order time threshold, and judging whether the work order to be executed is stuck. Compared with the related art, the neural network model can be used to learn and summarize the differences in the processing time required for different types of work orders in different execution links, as well as the differences in the residence time of work orders when the work orders are executed normally and when the execution is abnormal in different execution links, so that the model can accurately and finely predict the corresponding stuck order time thresholds of different types of work orders in different execution links, and avoid the stuck order recognition results being easily affected by human subjective factors, thereby improving the accuracy of stuck order recognition, reducing labor costs, and improving the efficiency of stuck order recognition. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 A schematic diagram of a flow chart of an intelligent analysis method for a work order card list provided in an embodiment of the present invention;

[0066] Figure 2 A schematic diagram of data point annotation provided in an embodiment of the present disclosure;

[0067] Figure 3 A schematic diagram of the algorithm flow of the BP neural network provided in the embodiment of the present disclosure;

[0068] Figure 4 A schematic diagram of the structure of an intelligent analysis device for a work order card provided by an embodiment of the present invention;

[0069] Figure 5 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0070] In order to make the purpose, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0071] The following are explanations of some of the terms that appear in the text:

[0072] 1. BP Neural Network: BP Neural Network (Backpropagation Neural Network) is a multi-layer feedforward neural network trained according to the error back propagation algorithm. It consists of an input layer, a hidden layer, and an output layer. Each neuron is connected to all neurons in the previous layer and processed by weights and activation functions. BP Neural Network is trained by the back propagation algorithm, calculating the error and adjusting the weights to make the output of the network as close to the target value as possible. In each iteration, the neural network calculates the gradient of the loss function and optimizes the weights according to the gradient descent algorithm.

[0073] 2. BP neural network with additional momentum method: It is an improved BP neural network algorithm. It introduces momentum terms based on the standard BP neural network, aiming to accelerate the convergence speed of the network and reduce oscillations during training.

[0074] The following is a more detailed description of the intelligent analysis method, device, equipment and medium for the work order card list provided by the present invention in conjunction with the accompanying drawings and embodiments.

[0075] The embodiment of the present invention provides an intelligent analysis method for a work order card order, such as Figure 1 As shown, including:

[0076] Step S101, obtaining historical work order data, the historical work order data including the stay time of different types of work orders in each link of the work order execution process.

[0077] For example, historical work order data can be obtained by collecting log information of each link of existing work orders. The historical work order data includes the time each existing work order stays at each link in the execution process of the work order system. The time a work order stays at a certain link is affected by factors such as the time required for business execution and whether the system is operating normally.

[0078] When a work order is executed normally, the complexity of processing in different execution links causes the work order to stay in different execution links for different periods of time. When an abnormal situation occurs during the execution of a work order, it will cause the work order to stay in a certain execution link for too long. For example, when a work order flows to a certain execution link, a failure caused by abnormal system operation causes the business to be unable to be processed normally, or it cannot flow normally to the next execution link, causing the work order to stay in a certain execution link for a long time and the work order is stuck. In addition, due to the unavailability of database services and errors in database tables, the business cannot be processed normally; due to some business rules and restrictions, the business requirements are not met, resulting in the inability to process the business; due to abnormalities caused by middleware services, the business cannot continue to be processed, etc., the work order stays in a certain link for too long and becomes stuck.

[0079] The historical work order data also includes some characteristic information indicating the attributes or execution links of the work order, such as product type, operation type, responsible system, etc.

[0080] Step S102, iteratively training the preset neural network model based on the historical work order data to obtain the stuck time thresholds of different types of work orders in each link of the work order execution process output by the trained neural network model.

[0081] To iteratively train the preset neural network model using historical work order data, it is necessary to first label the historical work order data to label each work order data in the historical work order data as stuck or non-stuck to obtain the labeling results. Then, the historical work order data is normalized and the processed historical work order data is divided into a data set including a training set, a test set, and a validation set. Finally, the processed data set is used to iteratively train the preset neural network model.

[0082] Training Set is a data set used to train the model. The training set is used to train the model and fit the data distribution law, that is, to determine the parameters such as the model weight and bias, which are called learning parameters.

[0083] Validation Set is a data set used to verify model performance. During model training, the validation set is used to adjust model parameters and hyperparameters to optimize model performance and avoid overfitting. That is, the validation set is used for model selection and does not participate in the determination of learning parameters. Instead, it is used to select model parameters and hyperparameters with smaller model errors.

[0084] The test set is a data set used to evaluate the performance of the model. After the model training is completed, the test set is used to evaluate the generalization ability of the model (generalization ability is the performance of the model on unknown data). That is, the test set is only used once after the training is completed to evaluate the effect of the final model.

[0085] Using historical work order data to train the neural network model can enable the neural network model to learn the differences in processing time required for different types of work orders in different execution links, as well as the differences in work order residence time when work orders are executed normally and when abnormalities occur in different execution links. As a result, the trained neural network model can output the processing time thresholds for different types of work orders in each link of the work order execution process, that is, the stuck order time thresholds. By confirming whether the work order stays in a certain execution link exceeds the stuck order time threshold, it can be determined whether the work order is stuck.

[0086] In one possible implementation, historical work order data is labeled, including: for each work order data in the historical work order data, converting the work order data into data points according to the residence time of the work order data in each link in the work order execution process; for each data point, determining a first number of other data points existing within a preset range around the data point; if the first number accounts for less than a preset proportion of the number of all data points, labeling the work order data corresponding to the data point as being in a stuck order state, otherwise labeling the work order data corresponding to the data point as being in a non-stuck order state.

[0087] In order to ensure the quality of annotation, a distance-based method can be used to evaluate the quality of annotation. Specifically, the work order can be regarded as a vector according to the length of time the work order stays in each link of the work order execution process, so that each work order data can be converted into a data point, and the distance between each work order data point and the surrounding work order data points is calculated. This distance indicates the difference in the residence time of different work order data in each link of the work order execution process. The first number of other data points existing within a preset range around each data point is determined by the distance between each work order data point and the surrounding work order data points. By judging whether the proportion of the first number corresponding to each data point in the number of all data points is less than the preset proportion, it is judged whether a work order data point is an abnormal data point. If so, the work order data corresponding to the work order data point is marked as a stuck order state, otherwise it is a non-stuck order state.

[0088] For example, based on a given radius ε and ratio π, for each point p to be detected, if the proportion of points whose distance from point p is less than the radius ε in all points is less than π, point p is marked as an abnormal point. Figure 2 A schematic diagram of data point annotation provided in an embodiment of the present disclosure, in Figure 2 In the example data shown, p201 and p202 are marked as abnormal data points by the model when there are no points within the radius ε.

[0089] The distance calculation formula of data point vector is as follows:

[0090] For example, there are two data point vectors X = {x1, x2, 3...xn}, = {y1, 2, y3...yn}, and the distance between them is defined as

[0091]

[0092] In one possible implementation, historical work order data is normalized, including: digitally encoding attribute features in the historical work order data that need to be normalized to obtain a feature sequence corresponding to the attribute features; scaling each feature value in the feature sequence according to the maximum and minimum values ​​in the feature sequence to obtain a normalized feature sequence.

[0093] According to the feature attribute rules, the attributes in the historical work order data, such as product type, operation type, execution link, link dwell time, responsible system, etc., are preprocessed. Among them, product type, operation type and execution link are discrete data, and label encoding can be used to encode different values ​​with integers for easy calculation. For example, the same product type, operation type, and execution link correspond to an integer code (such as a value between 1-100), and finally all features are normalized using the minimum-maximum scaling method. In this way, the historical work order data is converted into a feature sequence that can be directly processed by the neural network model.

[0094] Specifically, the normalization process steps are as follows: 1. Determine the feature column that needs to be normalized, such as the feature column formed by the integer encoding of the label of the discrete feature, and the feature column of the link execution time. 2. Find the minimum value (min) and maximum value (max) of the feature column. 3. For the feature value x of each sample in the feature column, scale it by the following formula: x_scaled = (x-min) / (-min), and the scaled feature value x_scaled will be in the range of [0,1].

[0095] In a possible implementation, the preset neural network model may be a BP neural network; further, the BP neural network may be a BP neural network with an additional momentum method.

[0096] BP neural network is a neural network model that includes an input layer, an output layer, and one or more layers of mysterious hidden units. These hidden units do not communicate directly with the outside world, but as their states change, they can affect the relationship between input and output. Each layer includes several nodes, and then the training data set is placed in the model for learning and iteration, and the optimization results are output. Figure 3 A schematic diagram of the algorithm flow of the BP neural network provided in an embodiment of the present disclosure.

[0097] In order to solve the problem of slow convergence or local convergence of high-dimensional data, the BP neural network algorithm (Backpropagation with Momentum) with the introduction of additional momentum value can be used. This method accelerates convergence and increases stability by using momentum terms when training the neural network.

[0098] The following is an example of the algorithm flow of the BP neural network with the additional momentum method:

[0099] S301, initializing the learning parameters and hyperparameters of the BP neural network.

[0100] Randomly initialize the network's learning parameters such as weights, biases, and momentum terms, usually small random numbers. Set hyperparameters such as the learning rate η and momentum factor α.

[0101] S302, using the training set as input to perform forward propagation in the BP neural network to obtain the output result of the BP neural network; calculating the error between the output result and the labeled result, and performing back propagation of the error in the BP neural network to adjust the learning parameters of the BP neural network.

[0102] For each training data sample taken as input in the training set, the following steps are performed:

[0103] S3021. The features of the input samples are propagated forward and the output of each neuron is calculated.

[0104] S3021. Calculate the output error:

[0105] Calculate the error of the output layer: δ_output = (target-output)*f'(output), where δ_output is the error of the output layer, output is the output of the sample, target is the target output of the sample, and f'(output) is the derivative of the output layer activation function.

[0106] S3023, back propagation error: Starting from the output layer, back propagate the error along the network to update the error term of each neuron:

[0107] δ_hidden=f'(hidden)*sum(W*δ_output)

[0108] Where δ_hidden is the error term of the neuron, f'(hidden) is the derivative of the hidden layer activation function, W is the weight matrix from this layer to the next layer, * represents matrix multiplication, and sum represents column summation.

[0109] S3024, update weight:

[0110] Update the weights from the input layer to the hidden layer: ΔW_input_hidden = η*X'*δ_hidden + α*ΔW_input_hidden_prev, where ΔW_input_hidden_prev is the momentum term from the input layer to the hidden layer, and X' is the transpose of the input sample features.

[0111] Update the weight from the hidden layer to the output layer ΔW_hidden_output = η*H'*δ_output + α*ΔW_hidden_output_prev, where ΔW_hidden_output_prev is the momentum term from the hidden layer to the output layer, and H' is the transpose of the output of the hidden layer.

[0112] Update model weights W_input_hidden=W_input_hidden+ΔW_input_hidden, W_hidden_output=W_hidden_output+ΔW_hidden_output.

[0113] Update the model momentum terms ΔW_input_hidden_prev=ΔW_input_hidden, ΔW_hidden_output_prev=ΔW_hidden_output.

[0114] S303, repeating step S302 until the BP neural network reaches a preset number of training rounds or the error meets a preset threshold, thereby obtaining a trained BP neural network model.

[0115] S305, using the validation set to verify the performance of the trained BP neural network model to adjust the hyperparameters of the trained BP neural network model; after the performance verification of the trained BP neural network model is completed, using the test set to evaluate the trained BP neural network model.

[0116] In the BP neural network with the additional momentum method, the momentum term affects the current weight update by accumulating the past weight adjustments, which can speed up the convergence speed and stability. The momentum factor α controls the degree of momentum influence, and the common range is 0.1 to 0.9. By adjusting the learning rate η and the value of the momentum factor α, the convergence speed and accuracy of the network can be affected. When updating the weights, the weights may also be restricted or regularized to avoid overfitting or excessive weights.

[0117] With the training data set as input, the algorithm continuously performs layer-by-layer recursion to calculate the optimal connection weights, and then performs iterative calculations by continuously adjusting the error value to achieve the optimal weights of the interconnected neurons in the neural network.

[0118] The disclosed embodiment annotates the historical work order record data and then normalizes it to obtain the training set, validation set and test set required for the training model. The BP neural network with momentum method is used for model training and iterative optimization, and finally the optimal stuck time threshold for different types of work orders in different execution links is obtained. This threshold can be used to make a high-precision prediction of whether the upcoming new work order will be stuck in the execution link, thereby improving the accuracy and effectiveness of functions such as work order automation monitoring, stuck order identification, early warning, and analysis and statistics.

[0119] Step S103, monitoring the length of time that the work order to be executed stays in the link of the work order execution process.

[0120] Real-time monitoring of the length of time each pending work order in the execution process of the work order system stays in the current execution link.

[0121] Step S104: If the dwell time exceeds the stuck time threshold corresponding to the link where the work order to be executed is located, it is determined that a stuck fault occurs in the work order to be executed.

[0122] According to the work order type of the work order to be executed and the execution link of the work order to be executed, the different types of work orders output by the model in S102 are searched for the corresponding stuck time threshold in the stuck time threshold of each link. If the stay time of the work order to be executed in the current execution link has exceeded the corresponding stuck time threshold, it can be determined that the work order to be executed has a stuck order failure, and then a stuck order warning can be issued to the operation and maintenance system to remind the operation and maintenance personnel to deal with the stuck order situation of the work order to be executed.

[0123] like Figure 4 As shown, based on the same inventive concept of the intelligent analysis method of the work order card list, the present invention also provides an intelligent analysis device for the work order card list, including:

[0124] The acquisition unit 401 is used to acquire historical work order data, where the historical work order data includes the residence time of different types of work orders at various stages in the work order execution process;

[0125] The training unit 402 is used to perform model iterative training on the preset neural network model based on the historical work order data, so as to obtain the stuck time thresholds of different types of work orders in each link of the work order execution process output by the trained neural network model;

[0126] The monitoring unit 403 is used to monitor the length of time that the work order to be executed stays in the link of the work order execution process;

[0127] The determination unit 404 is used to determine that a stuck order failure occurs in the work order to be executed if the dwell time exceeds the stuck order time threshold corresponding to the link where the work order to be executed is located.

[0128] In a possible implementation, in the device provided by the embodiment of the present invention, the training unit 402 is specifically used for:

[0129] Perform data labeling on historical work order data to label each work order data in the historical work order data as a stuck order state or a non-stuck order state to obtain a labeling result;

[0130] Perform data normalization on historical work order data, and divide the processed historical work order data into data sets including training set, test set and validation set;

[0131] The data set is used to perform iterative training on the preset neural network model.

[0132] In a possible implementation, in the device provided by the embodiment of the present invention, the training unit 402 is specifically used to:

[0133] For each work order data in the historical work order data, the work order data is converted into data points according to the residence time of the work order data in each link of the work order execution process;

[0134] For each data point, determining a first number of other data points existing within a preset range around the data point;

[0135] If the proportion of the first quantity in the total number of data points is less than a preset proportion, the work order data corresponding to the data point is marked as a stuck state, otherwise the work order data corresponding to the data point is marked as a non-stuck state.

[0136] In a possible implementation manner, in the apparatus provided by an embodiment of the present invention, when the training unit 402 determines, for each data point, the first number of other data points existing within a preset range around the data point, the training unit 402 is specifically configured to:

[0137] For each data point, the distance between the data point and other data points is calculated. The distance indicates the difference in the residence time of different work order data in each link of the work order execution process.

[0138] The number of target data points whose distance from the data point is less than the preset radius is taken as the first number.

[0139] In a possible implementation, in the device provided by the embodiment of the present invention, the training unit 402 is specifically used to:

[0140] Digitally encode the attribute features that need to be normalized in the historical work order data to obtain the feature sequence corresponding to the attribute features;

[0141] Each eigenvalue in the feature sequence is scaled according to the maximum and minimum values ​​in the feature sequence to obtain a normalized feature sequence.

[0142] In a possible implementation, in the device provided by the embodiment of the present invention, the preset neural network model includes a BP neural network;

[0143] When the training unit 402 uses the data set to perform model iterative training on the preset neural network model, it is specifically used to:

[0144] Initialize the learning parameters and hyperparameters of the BP neural network;

[0145] Take the training set as input and perform forward propagation in the BP neural network to obtain the output result of the BP neural network;

[0146] Calculate the error between the output result and the labeled result, and back-propagate the error in the BP neural network to adjust the learning parameters of the BP neural network;

[0147] Repeat the above steps of forward propagation input and back propagation error until the BP neural network reaches a preset number of training rounds or the error meets a preset threshold, and obtain a trained BP neural network model;

[0148] Use the validation set to verify the performance of the trained BP neural network model to adjust the hyperparameters of the trained BP neural network model;

[0149] After the performance verification of the trained BP neural network model is completed, the trained BP neural network model is evaluated using the test set.

[0150] In a possible implementation, in the device provided by an embodiment of the present invention, the BP neural network includes a BP neural network with an additional momentum method.

[0151] In addition, the intelligent analysis method and device of the work order card sheet according to the embodiment of the present invention can be implemented by an electronic device. Figure 5 A schematic diagram of the hardware structure of an electronic device provided by an embodiment of the present invention is shown.

[0152] The following specific reference Figure 5 , which shows a structural schematic diagram of an electronic device 500 suitable for implementing the embodiments of the present disclosure. Figure 5 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0153] like Figure 5As shown, the electronic device 500 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 501, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 502 or a program loaded from a storage device 508 to a random access memory (RAM) 503 to implement the intelligent analysis method of the work order card list of the embodiment described in the present disclosure. In the RAM 503, various programs and data required for the operation of the electronic device 500 are also stored. The processing device 501, the ROM 502, and the RAM 503 are connected to each other via a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0154] Typically, the following devices may be connected to the I / O interface 505: an input device 506 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 507 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 508 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 509. The communication device 509 may allow the electronic device 500 to communicate with other devices wirelessly or by wire to exchange data. Although Figure 5 The electronic device 500 is shown with various devices, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead.

[0155] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program contains a program code for executing the method shown in the flowchart, thereby implementing the intelligent analysis method of the work order card as described above. In such an embodiment, the computer program can be downloaded and installed from the network through the communication device 509, or installed from the storage device 508, or installed from the ROM 502. When the computer program is executed by the processing device 501, the above-mentioned functions defined in the method of the embodiment of the present disclosure are executed.

[0156] It should be noted that the computer-readable medium disclosed above may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, device or device. In the present disclosure, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, in which a computer-readable program code is carried. This propagated data signal may take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The computer readable signal medium may also be any computer readable medium other than a computer readable storage medium, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0157] In some embodiments, the client and the server may communicate using any currently known or future developed network protocol such as HTTP (HyperText Transfer Protocol), and may be interconnected with any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network ("LAN"), a wide area network ("WAN"), an internet (e.g., the Internet), and a peer-to-peer network (e.g., an ad hoc peer-to-peer network), as well as any currently known or future developed network.

[0158] The computer-readable medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.

[0159] The computer-readable medium carries one or more programs. When the one or more programs are executed by the electronic device, the electronic device:

[0160] Obtain historical work order data, which includes the duration of different types of work orders at each stage of the work order execution process;

[0161] The preset neural network model is iteratively trained based on historical work order data to obtain the stuck time thresholds of different types of work orders in each link of the work order execution process output by the trained neural network model;

[0162] Monitor the length of time that pending work orders stay in the work order execution process;

[0163] If the duration of stay exceeds the stuck time threshold corresponding to the link where the work order to be executed is located, it is determined that a stuck fault has occurred in the work order to be executed.

[0164] Optionally, when the above one or more programs are executed by the electronic device, the electronic device may also execute other steps described in the above embodiments.

[0165] Computer program code for performing the operations of the present disclosure may be written in one or more programming languages ​​or a combination thereof, including, but not limited to, object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0166] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present disclosure. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some implementations as replacements, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0167] The units involved in the embodiments described in the present disclosure may be implemented by software or hardware, wherein the name of a unit does not, in some cases, limit the unit itself.

[0168] The functions described above herein may be performed at least in part by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), and the like.

[0169] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0170] In the embodiment of the present invention, historical work order data including the residence time of different types of work orders in each link of the work order execution process is first collected, and then the neural network model is iteratively trained using the historical data, so that the neural network model learns the characteristics of the residence time of different work orders in each link in the historical work order data, thereby determining the corresponding stuck order time thresholds of different types of work orders in each link through the neural network model, and finally monitoring whether the residence time of the work order to be executed in the link of the work order execution process exceeds its corresponding stuck order time threshold, and judging whether the work order to be executed is stuck. Compared with the related art, the neural network model can be used to learn and summarize the differences in the processing time required for different types of work orders in different execution links, as well as the differences in the residence time of work orders when the work orders are executed normally and when the execution is abnormal in different execution links, so that the model can accurately and finely predict the corresponding stuck order time thresholds of different types of work orders in different execution links, and avoid the stuck order recognition results being easily affected by human subjective factors, thereby improving the accuracy of stuck order recognition, reducing labor costs, and improving the efficiency of stuck order recognition.

[0171] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0172] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0173] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1A function specified in one or more boxes.

[0174] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0175] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0176] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. An intelligent analysis method for work order, characterized in that: include: Acquire historical work order data, wherein the historical work order data includes the residence time of different types of work orders at various stages of the work order execution process; Based on the historical work order data, a preset neural network model is iteratively trained to obtain the stuck time thresholds of different types of work orders in each link of the work order execution process output by the trained neural network model; Monitor the length of time that pending work orders stay in the work order execution process; If the dwell time exceeds the stuck time threshold corresponding to the link where the work order to be executed is located, it is determined that a stuck fault occurs in the work order to be executed.

2. The intelligent analysis method of work order card list according to claim 1 is characterized in that: The performing iterative model training on a preset neural network model based on the historical work order data includes: Performing data labeling on the historical work order data to label each work order data in the historical work order data as a stuck order state or a non-stuck order state, and obtaining a labeling result; Performing data normalization processing on the historical work order data, and dividing the processed historical work order data into a data set including a training set, a test set, and a validation set; The data set is used to perform model iterative training on a preset neural network model.

3. The intelligent analysis method of work order card list according to claim 2 is characterized in that: The data labeling of the historical work order data includes: For each work order data in the historical work order data, convert the work order data into a data point according to the residence time of the work order data in each link of the work order execution process; For each data point, determining a first number of other data points existing within a preset range around the data point; If the first number accounts for less than a preset proportion of the total number of data points, the work order data corresponding to the data point is marked as a stuck state; otherwise, the work order data corresponding to the data point is marked as a non-stuck state.

4. The intelligent analysis method of work order card list according to claim 3 is characterized in that: The step of determining, for each data point, a first number of other data points existing within a preset range around the data point comprises: For each data point, the distance between the data point and other data points is calculated, and the distance indicates the difference in the residence time of different work order data in each link of the work order execution process; The number of target data points whose distance from the data point is less than a preset radius is taken as the first number.

5. The intelligent analysis method of work order card list according to claim 2 is characterized in that: The performing data normalization processing on the historical work order data includes: Digitally encoding the attribute features that need to be normalized in the historical work order data to obtain a feature sequence corresponding to the attribute features; Each feature value in the feature sequence is scaled according to the maximum value and the minimum value in the feature sequence to obtain a normalized feature sequence.

6. The intelligent analysis method of work order card list according to claim 2 is characterized in that: The preset neural network model includes a BP neural network; The adopting the data set to perform model iterative training on a preset neural network model includes: Initialize the learning parameters and hyperparameters of the BP neural network; Using the training set as input to perform forward propagation in the BP neural network to obtain an output result of the BP neural network; Calculating the error between the output result and the labeled result, and back-propagating the error in the BP neural network to adjust the learning parameters of the BP neural network; Repeat the steps of forward propagation input and back propagation error until the BP neural network reaches a preset number of training rounds or the error meets a preset threshold, thereby obtaining a trained BP neural network model; Using the validation set to validate the performance of the trained BP neural network model to adjust the hyperparameters of the trained BP neural network model; After the performance verification of the trained BP neural network model is completed, the trained BP neural network model is evaluated using the test set.

7. The intelligent analysis method of work order card list according to claim 6 is characterized in that: The BP neural network includes a BP neural network with an additional momentum method.

8. An intelligent analysis device for work order cards, characterized in that: include: An acquisition unit, used to acquire historical work order data, wherein the historical work order data includes the residence time of different types of work orders at various stages in the work order execution process; A training unit, used to perform iterative model training on a preset neural network model based on the historical work order data, so as to obtain the stuck time thresholds of different types of work orders in each link of the work order execution process output by the trained neural network model; A monitoring unit is used to monitor the length of time that a work order to be executed stays in the link of the work order execution process; A determination unit is used to determine that a stuck order failure occurs in the work order to be executed if the stay time exceeds a stuck order time threshold corresponding to the link where the work order to be executed is located.

9. An electronic device, characterized in that: include: At least one processor, at least one memory and computer program instructions stored in the memory, when the computer program instructions are executed by the processor, implement the method according to any one of claims 1 to 7.

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

Images