A method for analyzing the effectiveness of on-duty personnel deployment based on graph neural network
Through the analysis of the distribution data of duty personnel based on graph neural network, the problem of the inability to quickly and accurately analyze the distribution plan of duty personnel in the existing technology is solved, and the rapid and accurate analysis of the distribution efficiency of duty personnel and the formulation of the optimal distribution plan are realized. .
Patent Information
- Application Number
- CN202311162043.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-09-11
- Publication Date
- 2024-11-05
- Estimated Expiration
- 2043-09-11
AI Technical Summary
The prior art cannot quickly and accurately perform the effectiveness analysis of the on-duty personnel deployment plan, resulting in the inability to make the optimal on-duty personnel deployment plan.
Using a graph neural network-based method, the topological structure diagram is constructed by obtaining and preprocessing the distribution data of on-duty personnel, and performing graph embedding operations. The graph neural network model and deep learning model are used to build the point distribution efficiency analysis model of on-duty personnel, and then backpropagation is carried out. The algorithm optimizes the model to achieve rapid and accurate analysis of the distributed data of on-duty personnel.
It realizes targeted analysis of the distribution data of duty personnel, improves the accuracy of the analysis of the on-duty personnel's point distribution efficiency and the generalization ability of the model, and helps the on-duty department to make the best point distribution plan.
Smart Images

Figure CN117236753B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of computer information processing, and in particular to a method for analyzing the efficiency of on-duty personnel deployment based on a graph neural network. Background Art
[0002] In the process of dispatching resources for duty tasks, the rationality of the deployment of duty personnel is an important factor. According to the location, time and number of cases, the distribution of duty personnel and patrol routes must be reasonably arranged. Therefore, it is necessary to analyze the effectiveness of the deployment plan of duty personnel and then make the optimal deployment plan of duty personnel.
[0003] However, since the information and data on duty personnel, cases, etc. are huge and complex, it is impossible to analyze them quickly and accurately by human analysis. Therefore, the present invention proposes a method for analyzing the efficiency of duty personnel deployment based on graph neural networks. Summary of the invention
[0004] In order to solve the technical problems existing in the prior art, the present invention provides a method for analyzing the efficiency of on-duty personnel deployment based on a graph neural network, which is characterized by comprising the following steps:
[0005] S1. Obtain the first duty personnel distribution data to generate an original training data set;
[0006] S2, preprocessing the original training data set to obtain a training data set;
[0007] S3, constructing the training data set into a topological structure diagram;
[0008] S4, performing a graph embedding operation on the topological structure graph to obtain graph embedding vector data;
[0009] S5. Use the graph neural network model and the deep learning model to build a duty personnel deployment efficiency analysis model, use the graph embedding vector data to train the duty personnel deployment efficiency analysis model, and optimize the duty personnel deployment efficiency analysis model through a learning algorithm;
[0010] S6. Input the second duty personnel distribution data into the duty personnel deployment efficiency analysis model to obtain a predicted value, and obtain an efficiency analysis result based on the predicted value.
[0011] Furthermore, the first on-duty personnel distribution data and the second on-duty personnel distribution data both include: the location of the case, the type of case, the time of the case, the distribution of on-duty personnel and the patrol route; the first on-duty personnel distribution data is the model training data; the second on-duty personnel distribution data is the model analysis data.
[0012] Furthermore, the preprocessing of the original training data set is specifically to perform data cleaning, deduplication and standardization on the original training data set.
[0013] Furthermore, the topological structure diagram uses the locations where the cases occurred as nodes and the patrol routes as edges.
[0014] Furthermore, the topological structure graph is subjected to a graph embedding operation, specifically, converting nodes and edges into vector representations.
[0015] Furthermore, the graph neural network model includes a GAT model.
[0016] Furthermore, the deep learning model is a GAN model.
[0017] Furthermore, the GAN model includes a generation model and a discrimination model.
[0018] Furthermore, the use of the graph embedding vector data to train the on-duty personnel deployment efficiency analysis model is specifically as follows: the GAT model assigns weights to each node according to different case types; the generative model learns to generate similar data based on the graph embedding vector data, and the discriminant model distinguishes between the graph embedding vector data and the similar data; the generative model and the discriminant model compete with each other and are repeatedly trained iteratively.
[0019] Furthermore, the learning algorithm is a back propagation algorithm.
[0020] Compared with the prior art, the beneficial effects of the present invention are: by converting the on-duty personnel distribution data into graph embedded vector data, the on-duty personnel deployment efficiency analysis model based on the graph neural network is trained, so as to obtain the on-duty personnel deployment efficiency analysis results for different on-duty personnel distribution data in a targeted, fast and accurate manner, so as to enable the on-duty department to make the optimal on-duty personnel deployment plan; the on-duty personnel deployment efficiency analysis model is based on the graph neural network, and the graph neural network model is used to improve the effectiveness of training and the accuracy of testing; the on-duty personnel deployment efficiency analysis model is optimized by the back propagation algorithm to improve the generalization ability of the model. BRIEF DESCRIPTION OF THE DRAWINGS
[0021] Figure 1 It is a flow chart of a method for analyzing the effectiveness of on-duty personnel deployment based on graph neural network of the present invention. DETAILED DESCRIPTION
[0022] In order to facilitate the understanding of the present invention, the present invention will be described more fully below with reference to the relevant drawings. Preferred embodiments of the present invention are provided in the drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present invention more thorough and comprehensive.
[0023] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used herein in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more related listed items.
[0024] See also Figure 1 As shown, the present invention provides a method for analyzing the effectiveness of on-duty personnel deployment based on a graph neural network, comprising the following steps:
[0025] S1. Obtain the first duty personnel distribution data to generate an original training data set;
[0026] S2, preprocessing the original training data set to obtain a training data set;
[0027] S3, constructing the training data set into a topological structure diagram;
[0028] S4, performing a graph embedding operation on the topological structure graph to obtain graph embedding vector data;
[0029] S5. Use the graph neural network model and the deep learning model to build a duty personnel deployment efficiency analysis model, use the graph embedding vector data to train the duty personnel deployment efficiency analysis model, and optimize the duty personnel deployment efficiency analysis model through a learning algorithm;
[0030] S6. Input the second duty personnel distribution data into the duty personnel deployment efficiency analysis model to obtain a predicted value, and obtain an efficiency analysis result based on the predicted value.
[0031] The first duty personnel distribution data and the second duty personnel distribution data both include: the location of the case, the type of case, the time of the case, the distribution of duty personnel and the patrol route;
[0032] The first on-duty personnel distribution data is model training data, i.e., the on-duty personnel distribution data used to train the on-duty personnel distribution efficiency analysis model; the second on-duty personnel distribution data is model analysis data, i.e., the on-duty personnel distribution data that needs to be analyzed for efficiency in actual application after the on-duty personnel distribution efficiency analysis model has been trained.
[0033] The preprocessing of the original training data set includes data cleaning, deduplication and standardization of the original training data set.
[0034] Data cleaning includes data consistency checking, removing noise data and irrelevant data in the original training data set, and processing missing data and missing values.
[0035] The consistency check is to check whether the data meets the requirements based on the reasonable value range and mutual relationship of each data, and to find data that exceeds the normal range, is logically unreasonable or contradictory. For example, there are no on-duty personnel distributed in a patrol route, which means there is a contradiction.
[0036] In this solution, noise data and irrelevant data are interference data outside the distribution data of the first on-duty personnel, which can be screened and deleted through manual inspection; they can also be screened and deleted through clustering method, for example, the location of the case, case type, time of the case, distribution of on-duty personnel and patrol route are divided into five categories of data, and the five categories of data are grouped together to find data isolated points outside the set. These data isolated points are noise data and irrelevant data, and then they are accurately removed.
[0037] Data deduplication. In this solution, when the location where the case occurred, the corresponding case type and the time when the case occurred are all the same, it is determined to be duplicate data. This avoids the situation where different cases occurring at the same location are deduplicated, thereby ensuring data diversity and improving the accuracy of model training.
[0038] The topological structure diagram uses the location where the case occurred as a node and the patrol route as an edge to reflect the spatial relationship of the distribution data of on-duty personnel during on-duty activities.
[0039] The graph embedding operation of the topological structure graph is specifically: converting nodes and edges into vector representations. The graph embedding operation can select any one of the methods based on factor decomposition, random walk, and deep learning.
[0040] Graph Neural Network (GNN) is a general term for algorithms that use neural networks to learn graph structured data, extract and discover features and patterns in graph structured data, and meet the needs of graph learning tasks such as clustering, classification, prediction, segmentation, and generation.
[0041] The graph neural network model includes the GAT model, namely the Graph Attention Network model. The attention mechanism allows a neural network to focus only on the information required for task learning, and it can select specific inputs. Introducing the attention mechanism in the graph neural network allows the neural network to focus on nodes and edges that are more relevant to the task, thereby improving the effectiveness of training and the accuracy of testing.
[0042] In this solution, the GAT model, i.e. the graph attention network model, can assign different weights to different case types, that is, assign different weights to each node of the topological structure diagram, so that the graph neural network focuses on nodes with larger weights, thereby allowing the on-duty personnel deployment efficiency analysis model to be trained more accurately.
[0043] The deep learning model is a GAN model, i.e., a Generative Adversarial Networks model; the GAN model includes a generative model and a discriminative model. The generative model is responsible for capturing the distribution of sample data, while the discriminative model is generally a binary classifier that determines whether the input is real data or a generated sample. The optimization process of this model is to fix the discriminative model or the generative model during training, update the parameters of the other model, and iterate alternately. Finally, the generative model can estimate the distribution of sample data.
[0044] In this scheme, the GAN model can use the graph embedding vector data obtained through the first on-duty personnel distribution data as real data, and generate similar sample data similar to the real sample data based on the real sample data through the generative model, and then the discriminant model determines whether the similar sample data is real sample data. After continuous generation, discrimination, iteration and parameter adjustment, the similar sample data generated by the generative model is infinitely close to the real data, which makes it impossible for the discriminant model to make a judgment, thereby obtaining a large amount of similar sample data that can be used for the training of the on-duty personnel distribution efficiency analysis model, and ensuring the accuracy of the performance analysis of the on-duty personnel distribution efficiency analysis model after training with sufficient training data.
[0045] That is, the use of the graph embedding vector data to train the on-duty personnel deployment efficiency analysis model is specifically as follows: the GAT model assigns weights to each node according to different case types; the generative model learns to generate similar data based on the graph embedding vector data, and the discriminant model distinguishes between the graph embedding vector data and the similar data; the generative model and the discriminant model compete with each other and are repeatedly trained iteratively.
[0046] The learning algorithm is a back propagation algorithm, which specifically adjusts the weights continuously through the reverse transfer of errors, so that the predicted value output by the on-duty personnel distribution efficiency analysis model is as close to the target value as possible. Specifically, in each forward calculation, the input data is passed into the on-duty personnel distribution efficiency analysis model, and the calculation is performed according to the preset weights to obtain the predicted value. Then, by comparing the error between the predicted value and the true value, it is determined which model parameters need to be adjusted to minimize the error. In the back propagation process, the derivatives of each node with respect to the respective input variables and weight parameters are calculated in turn according to the chain rule, and the error gradient is returned along the connection path to the corresponding nodes in all hidden layers and input layers to update their parameter values. Therefore, iterative optimization on multiple samples can continuously reduce the error between the predicted output and the actual value and improve the generalization ability of the model.
[0047] The second duty personnel distribution data is input into the duty personnel deployment efficiency analysis model to obtain a predicted value, and an efficiency analysis result is obtained according to the predicted value, for example, the distribution of duty personnel or the rationality of patrol routes are analyzed according to the predicted value.
[0048] When the performance analysis results are analyzed, a visual display is performed, and the visual display adopts any method such as a heat map, a bar graph, etc.
[0049] In other embodiments, the graph neural network model may also choose to adopt a graph convolutional network model (Graph Convolutional Network, GCN), and the convolution operation method includes based on spectral decomposition, i.e. spectral decomposition graph convolution; based on node space transformation, i.e. spatial graph convolution.
[0050] The beneficial effects of the present invention are as follows: by converting the on-duty personnel distribution data into graph embedded vector data, the on-duty personnel deployment efficiency analysis model based on the graph neural network is trained, so as to obtain the on-duty personnel deployment efficiency analysis results for different on-duty personnel distribution data in a targeted, fast and accurate manner, so as to enable the on-duty department to make the optimal on-duty personnel deployment plan; the on-duty personnel deployment efficiency analysis model is based on the graph neural network, and the graph neural network model is used to improve the effectiveness of training and the accuracy of testing; the on-duty personnel deployment efficiency analysis model is optimized by the back propagation algorithm to improve the generalization ability of the model.
[0051] For those skilled in the art, various other corresponding changes and deformations can be made according to the technical solutions and concepts described above, and all of these changes and deformations should fall within the protection scope of the claims of the present invention.
[0052] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.
Claims
1. A method for analyzing the effectiveness of on-duty personnel deployment based on graph neural networks. It is characterized in that The following steps are involved: S1. Obtain the first duty personnel distribution data to generate an original training data set; S2, preprocessing the original training data set to obtain a training data set; S3, constructing the training data set into a topological structure diagram; S4, performing a graph embedding operation on the topological structure graph to obtain graph embedding vector data; S5. Use the graph neural network model and the deep learning model to build a duty personnel deployment efficiency analysis model, use the graph embedding vector data to train the duty personnel deployment efficiency analysis model, and optimize the duty personnel deployment efficiency analysis model through a learning algorithm; S6, inputting the second duty personnel distribution data into the duty personnel distribution efficiency analysis model to obtain a predicted value, and obtaining an efficiency analysis result according to the predicted value; The first duty personnel distribution data and the second duty personnel distribution data both include: the location of the case, the type of case, the time of the case, the distribution of duty personnel and the patrol route; The first on-duty personnel distribution data is model training data; the second on-duty personnel distribution data is model analysis data; The topological structure diagram has the locations where the cases occurred as nodes and the patrol routes as edges; The graph embedding operation of the topological structure graph is specifically: converting nodes and edges into vector representations; The graph neural network model includes a GAT model; The deep learning model is a GAN model; The GAN model includes a generation model and a discrimination model; The use of the graph embedding vector data to train the on-duty personnel deployment efficiency analysis model is specifically as follows: the GAT model assigns weights to each node according to different case types; the generative model learns to generate similar data based on the graph embedding vector data, and the discriminant model distinguishes between the graph embedding vector data and the similar data; the generative model and the discriminant model compete with each other and are repeatedly trained iteratively.
2. According to the graph neural network-based on-duty personnel deployment efficiency analysis method of claim 1, It is characterized in that The preprocessing of the original training data set is specifically to perform data cleaning and standardization on the original training data set.
3. According to the graph neural network-based on-duty personnel deployment efficiency analysis method of claim 1, It is characterized in that The learning algorithm is a back propagation algorithm.
