Laboratory man-hour data analysis and resource optimization method and system

Abstract

The invention provides a laboratory man-hour data analysis and resource optimization method and system, and the method comprises the steps: obtaining man-hour data and task data of laboratory personnel, and carrying out the preprocessing of the man-hour data and task data, and obtaining man-hour distribution features and task time consumption features; clustering the man-hour data and the task data through a clustering algorithm to generate a personnel man-hour mode cluster and a task feature cluster; and based on the personnel working hour mode cluster and the task feature cluster, constructing a personnel and task matching cluster, and optimizing resource configuration. Aiming at the problems that the matching efficiency of working hours and tasks of laboratory personnel is low and the decision accuracy is influenced by data exception, efficient resource allocation and management are realized through multi-dimensional data acquisition, clustering analysis and dynamic matching.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of information technology, more particularly, to a laboratory work hour data analysis and resource optimization method and system. BACKGROUND

[0002] Under the background of increasingly fierce market competition and continuously strengthened industry regulation, the requirements for fine and standardized management of laboratories are continuously improving. In order to fully exert the team effectiveness, laboratories urgently need to obtain objective and quantifiable personnel effectiveness evaluation data to support management decisions. However, the existing personnel effectiveness evaluation methods have obvious limitations, mainly reflected in data collection lag and fuzzy evaluation indexes, which leads to the lack of practical application value of the evaluation results, and it is difficult to meet the needs of all parties for the modernization of laboratory management.

[0003] At present, laboratories generally use manual reporting of work hours to perform performance evaluation. The specific process is as follows: after completing the daily testing tasks, the employee manually fills in the date, sample variety, testing project and batch number, etc. in the form, and then the team leader verifies and summarizes the work hours monthly to rank. This method relies on manual recording and summarizing, lacks a scientific and systematic personnel effectiveness evaluation mechanism, and is difficult to accurately reflect the matching efficiency between personnel work hours and tasks in real time, which affects the effective judgment of personnel work load and the rationality of resource allocation.

[0004] Therefore, it is an urgent need to develop a special laboratory personnel effectiveness evaluation system, integrate existing data resources, build a scientific evaluation index system, and realize the automation and intelligentization of the evaluation process with the help of information technology, which has become an urgent need to improve the management effectiveness of laboratories. SUMMARY

[0005] The purpose of the present application is to solve the problems of manual reporting of work hours in laboratories, and to provide a laboratory work hour data analysis and resource optimization method and system.

[0006] The technical scheme of the present application is: The present application provides a laboratory work hour data analysis and resource optimization method, which comprises the following steps: S1, obtaining work hour data and task data of laboratory personnel, respectively preprocessing to obtain work hour distribution characteristics and task time consumption characteristics; S2, clustering the work hour data and the task data by a clustering algorithm to generate personnel work hour mode clusters and task feature clusters; S3, constructing a matching cluster of personnel and tasks based on the personnel work hour mode clusters and the task feature clusters, and optimizing resource allocation.

[0007] Further, S1 comprises: S11, acquiring the work hour data of the laboratory personnel, including personnel identification, work hour length and task type; preprocessing the work hour data to obtain work hour distribution features, including daily average work hour, weekly fluctuation coefficient and overtime frequency; S12, acquiring task data, including task identification and single task time consumption; preprocessing the task data to obtain task time consumption features, including time consumption mean and time consumption fluctuation coefficient; S13, standardizing the work hour data and the task data to generate a feature set of uniform magnitude.

[0008] Further, S13 comprises: for the work hour distribution features and the task time consumption features, respectively adopting Z-score standardization method for processing; performing dimension reduction processing on the standardized feature set, and retaining dimensions strongly correlated with work hour mode and task features.

[0009] Further, S2 comprises: S21, determining the optimal cluster number of the clustering algorithm by adopting elbow method or contour coefficient method; S22, generating personnel work hour mode clusters of fixed stable type, elastic high intensity type and low load fluctuation type based on the feature set of the work hour data; S23, generating task feature clusters of fast standardized task, auxiliary task and high difficulty long time consumption task based on the feature set of the task data; S24, performing feature verification on the personnel work hour mode clusters and the task feature clusters to confirm that the cluster features are consistent with the business scenario.

[0010] Further, S3 comprises: S31, based on the personnel work hour mode clusters and the task feature clusters, constructing a two-dimensional feature space, and performing clustering to generate matching clusters of personnel and tasks; S32, determining matching rules and priorities according to the matching clusters, which are used to guide the allocation of personnel posts and task types; the matching relationship from high to low priority is as follows: allocation combination of fixed stable type personnel and fast standardized task, allocation combination of elastic high intensity type personnel and high difficulty task, and allocation combination of low load fluctuation type and auxiliary task.

[0011] Further, in S32, the matching rules are specifically: for fixed stable type personnel, preferentially allocating fast standardized tasks with low time consumption fluctuation coefficient; for elastic high intensity type personnel, preferentially allocating high difficulty tasks with high time consumption fluctuation coefficient; for low load fluctuation type personnel, allocating auxiliary tasks; Based on the matching rules, a priority list is generated to guide laboratory resource allocation.

[0012] Further, the method further comprises S4, identifying abnormal work time records by calculating the distance between the sample and the cluster centroid, specifically: S41, for each work time record, calculating the Euclidean distance between it and the centroid of the personnel work time mode cluster; if the Euclidean distance of the work time record is greater than the average distance within the cluster and exceeds the first preset threshold, the work time record is marked as an abnormal point; S42, for each task record, calculating the Euclidean distance between it and the centroid of the task feature cluster; if the Euclidean distance of the task record is greater than the average distance within the cluster and exceeds the second preset threshold, the task record is marked as an abnormal point; S43, classifying the records marked as abnormal points to distinguish false positive work time and invalid overtime.

[0013] Further, the method further comprises S5, generating management decision suggestions according to the matching cluster and the abnormal detection result, specifically: S51, generating a differentiated scheduling strategy according to the personnel work time mode cluster; generating a standardized work time quota according to the task feature cluster; generating a personnel post adjustment suggestion according to the matching rule; S52, regularly updating the clustering result and the matching rule to generate dynamic management decision suggestions.

[0014] A laboratory work time data analysis and resource optimization system, the system is configured to perform the method.

[0015] A computer readable storage medium having stored thereon a computer program which, when executed, implements the method.

[0016] The beneficial effects of the present application are: The present application discloses a laboratory work time data analysis and resource optimization method, aiming at the problems of low efficiency of personnel work time and task matching and data abnormality affecting decision accuracy, through multi-dimensional data collection, clustering analysis and dynamic matching, efficient resource allocation and management are realized.

[0017] The present application first acquires and preprocesses personnel work time and task data, calculates daily work time, time period distribution, task time consumption and other features, and ensures data quality through Z-score standardization; determines the optimal clustering number based on the K-means algorithm and the elbow method, generates clusters reflecting work rhythm and task difficulty, and then builds a person-task matching model to preferentially allocate fixed stable personnel and fast standardized tasks to improve efficiency.

[0018] The present application identifies abnormal work time by calculating the Euclidean distance between the work time record and the cluster centroid to ensure data authenticity; regularly updates the clustering and matching rules to dynamically generate scheduling adjustment, task quota and training suggestions, significantly improving laboratory resource utilization and the accuracy of management decisions.

[0019] Other features and advantages of the present application will be described in detail in the subsequent specific embodiments. BRIEF DESCRIPTION OF DRAWINGS

[0020] The above and other objects, features and advantages of the present application will become more apparent from the following detailed description when taken in conjunction with the accompanying drawings in which like reference characters refer to like parts throughout the figures, and wherein:

[0021] Figure 1 A flow chart of the laboratory man-hour data analysis and resource optimization method of the present application is shown. DETAILED DESCRIPTION

[0022] Preferred embodiments of the present application will be described in more detail with reference to the drawings. Although the preferred embodiments of the present application are shown in the drawings, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments set forth herein.

[0023] Example 1

[0024] Figure 1 A flow chart of the laboratory man-hour data analysis and resource optimization method of the present application is shown.

[0025] As Figure 1 shown, the present application provides a laboratory man-hour data analysis and resource optimization method, which comprises the following steps: S1, obtaining man-hour data and task data of laboratory personnel, and pre-processing to obtain man-hour distribution features and task time consumption features, respectively; Specifically, it comprises: S11, obtaining man-hour data of laboratory personnel, including personnel identification, man-hour duration and task type; pre-processing the man-hour data to obtain man-hour distribution features, including daily man-hour, weekly fluctuation coefficient and overtime frequency; S12, obtaining task data, including task identification and single task time consumption; pre-processing the task data to obtain task time consumption features, including time consumption mean and time consumption fluctuation coefficient; S13, for man-hour distribution features and task time consumption features, respectively, Z-score standardization method is used for processing; dimension reduction processing is performed on the standardized feature set, and dimensions strongly related to man-hour mode and task features are retained.

[0026] In an embodiment, personnel work hour records in the past 3 months are extracted from the laboratory management system, including personnel ID, date, work hour length, task type and time period information, and task data such as single-time consumption records of sample detection, data arrangement, etc. are also extracted. The daily average work hours, weekly fluctuation coefficient and overtime frequency are calculated from the work hour data. The weekly average work hours, monthly work hour fluctuation coefficient, proportion of morning, noon and night shifts, overtime length, and proportion of work hours for different task types can also be included. The average single-task consumption time and consumption fluctuation coefficient are calculated from the task data. The median single-task consumption time and personnel completion difference can also be included, and missing values or extreme values such as work hours less than 0.5 hours or more than 16 hours are removed. The features are Z-score standardized, i.e. each feature is subtracted by the mean and divided by the standard deviation, so that the work hour distribution features and task consumption features are in the same order of magnitude, avoiding weight bias. The correlation between the features is obtained, and the dimensions that are strongly correlated with the work hour pattern and task features are retained, such as the daily average work hours and the average consumption time.

[0027] S2, clustering the work hour data and the task data by a clustering algorithm to generate personnel work hour pattern clusters and task feature clusters; Specifically, it includes: S21, determining the optimal cluster number of the clustering algorithm by elbow method or contour coefficient method; S22, generating personnel work hour pattern clusters of fixed stable type, high intensity type with flexibility and low load fluctuation type based on the feature set of the work hour data; S23, generating task feature clusters of fast standardized task, auxiliary task and high difficulty long consumption time task based on the feature set of the task data; S24, verifying the features of the personnel work hour pattern clusters and the task feature clusters to confirm that the cluster features are consistent with the business scenario.

[0028] In an embodiment, the K-means algorithm (which can be implemented using the Python sklearn library) is used to cluster the preprocessed work hour distribution features. First, the “elbow method” or “contour coefficient method” is used to select the optimal K value (for example, when K=3 is found by the elbow method, the downward trend of the within-cluster sum of squares slows down significantly), so that the personnel are divided into 3 work hour pattern clusters, such as fixed stable type, high intensity type with flexibility and low load fluctuation type.

[0029] For example: clustering feature dimensions: select quantitative indicators that are strongly correlated with “work hour pattern” as clustering inputs, such as: basic work hours: daily average work hours, weekly average work hours, monthly work hour fluctuation coefficient (reflecting stability); time period distribution: proportion of work hours in morning, noon and night shifts, overtime length and frequency, proportion of weekend work hours; task association: proportion of work hours for different types of tasks (such as sample detection, data arrangement, instrument calibration).

[0030] Cluster 1 (fixed stable type): daily working hours of about 8 hours, small fluctuations, mainly for routine detection tasks (such as physical and chemical detection post personnel); Cluster 2 (flexible high intensity type): daily working hours of 10+ hours, high overtime frequency, special experiments / emergency tasks accounting for more than 60% (such as scientific research project team personnel); Cluster 3 (low load fluctuation type): daily working hours less than 6 hours, task type is scattered (such as part-time personnel or auxiliary post personnel). Avoid "one-size-fits-all" scheduling, match rest resources for high-intensity groups, and allocate additional tasks for low-load groups; identify "overwork risk" personnel (such as individuals in cluster 2 with abnormally high working hour fluctuations), and intervene in advance.

[0031] In an embodiment, the K-means algorithm is also applied to the task time consumption characteristics, and the K value is determined as 3 to generate task characteristic clusters, such as a fast standardized task cluster, a medium fluctuation task cluster, and a high difficulty long time consumption task cluster. The K-means algorithm is used to divide data points by iteratively minimizing the Euclidean distance of samples to cluster centroids, and the centroid is the mean of points in the cluster.

[0032] The time consumption of laboratory tasks often has "implicit rules", and K-means can group tasks according to "working hour characteristics" through clustering. After clustering the working hour data of researchers, it can provide a basis for task quota formulation and efficiency assessment.

[0033] For example: clustering characteristic dimensions: for task-level data, select single task time consumption (mean, median); time consumption fluctuation coefficient (reflecting task difficulty stability); personnel proficiency correlation (such as the time consumption difference of different personnel completing the same task).

[0034] After clustering "sample detection" type tasks, the following clusters may be obtained: Cluster A (fast standardized task): time consumption mean 20 minutes, fluctuation coefficient <10% (such as routine water quality pH detection); Cluster B (medium fluctuation task): time consumption mean 60 minutes, fluctuation coefficient 20%-30% (such as heavy metal spectrophotometric detection, affected by sample pretreatment); Cluster C (high difficulty long time consumption task): time consumption mean 180 minutes, fluctuation coefficient >50% (such as unknown sample mass spectrometric qualitative analysis).

[0035] Formulate standardized working hour quotas for different cluster tasks (such as 20 minutes per sample for cluster A and 3 hours per sample for cluster C); analyze the time consumption differences of personnel within the same task cluster (such as a personnel in cluster A with a time consumption 30% higher than the mean), and carry out targeted skill training. The clustering results are used to identify task difficulty stability, which is beneficial to subsequent efficiency evaluation.

[0036] S3, based on the personnel working hour mode cluster and the task characteristic cluster, construct a matching cluster of personnel and tasks, and optimize resource allocation; Specifically, it comprises: S31, based on personnel work time mode cluster and task feature cluster, constructing a two-dimensional feature space, performing clustering to generate personnel and task matching clusters; S32, determining matching rules and priorities according to the matching clusters, for guiding the allocation of personnel posts and task types; the matching relationship from high to low priority is: the allocation combination of fixed stable personnel and rapid standardized tasks, the allocation combination of elastic high intensity personnel and high difficulty tasks, and the allocation combination of low load fluctuation personnel and auxiliary tasks. In S32, the matching rule is specifically: for fixed stable personnel, preferentially allocate rapid standardized tasks with low time consumption fluctuation coefficient; for elastic high intensity personnel, preferentially allocate high difficulty tasks with high time consumption fluctuation coefficient; for low load fluctuation personnel, allocate auxiliary tasks; generate a priority list based on the matching rule to guide laboratory resource allocation.

[0037] In an embodiment, a two-dimensional feature space is constructed in combination with personnel work time mode cluster and task feature cluster, that is, the cluster label of each personnel is paired with the task cluster label to form a matching matrix, wherein the matching matrix is a matrix generated by calculating the similarity between clusters, such as cosine similarity, and each element represents the matching degree of the personnel cluster and the task cluster. Cosine similarity is the value of vector inner product divided by the length, which is used to quantify the directional similarity of feature vectors. On the basis of the matching matrix, K-means is applied for comprehensive clustering to generate optimal matching clusters, such as the matching cluster of fixed stable personnel and rapid standardized tasks, and the matching degree is optimized by minimizing the distance between cluster centers. According to the matching cluster, rules are formulated, such as allocating high intensity personnel to high difficulty tasks to avoid resource mismatch, thereby optimizing the overall resource allocation of the laboratory, such as reducing the time waste of high skill personnel handling low difficulty tasks.

[0038] For example, the clustering feature dimension: input “personnel work time mode” (such as cluster 1 - fixed type, cluster 2 - high intensity type) and “task work time feature” (such as cluster A - rapid type, cluster C - high difficulty type) at the same time, construct a two-dimensional feature space for clustering.

[0039] After clustering, the “optimal matching cluster” may be obtained: “fixed stable personnel + rapid standardized task”: the matching degree is the highest, and the efficiency is the best (such as letting regular detection post personnel focus on cluster A task); “elastic high intensity personnel + high difficulty task”: the matching degree is higher (such as letting research post personnel focus on cluster C task); “low load fluctuation personnel + auxiliary task”: avoid resource waste (such as letting auxiliary post personnel undertake low time consumption tasks such as instrument maintenance and sample receiving).

[0040] Formulate "person-task" matching rules to reduce resource mismatch of "high-skilled personnel performing low-difficulty tasks"; based on matching efficiency data, adjust personnel job division (such as reassigning personnel who frequently make matching errors to more suitable task types), optimize resource utilization, and effectively support experimental needs.

[0041] S4, identify abnormal work hour records by calculating the distance between the sample and the cluster center, specifically: S41, for each work hour record, calculate the Euclidean distance between it and the cluster center of the personnel work hour mode; if the Euclidean distance of the work hour record is greater than the average distance within the cluster and exceeds the first preset threshold, mark the work hour record as an abnormal point; S42, for each task record, calculate the Euclidean distance between it and the cluster center of the task feature; if the Euclidean distance of the task record is greater than the average distance within the cluster and exceeds the second preset threshold, mark the task record as an abnormal point; S43, classify the records marked as abnormal points to distinguish false work hours and invalid overtime.

[0042] In one embodiment, for each work hour record sample, the Euclidean distance between it and the cluster center of the personnel work hour mode is calculated, and if the distance exceeds 2 times the average distance within the cluster, it is marked as abnormal, such as a record with daily average work hours of 18 hours far exceeding the center of the fixed stable cluster by 8 hours.

[0043] Similarly, the distance between the task data sample and the task feature cluster center is calculated to identify a single time consumption of 120 minutes for a "rapid standardized task" (cluster A), which is far beyond the cluster average of 20 minutes, which may be a task classification error (actually a cluster C task) or an operation error causing time consumption anomaly. This distance calculation process helps to automatically screen false work hours, reduces the cost of manual review, and traces the cause such as operation error (such as personnel who frequently misrecord need to be trained in system operation, and teams with frequent invalid overtime need to optimize task planning).

[0044] S5, generate management decision suggestions according to the matching cluster and abnormal detection results, specifically: S51, generate differentiated scheduling strategies according to the personnel work hour mode cluster; generate standardized work hour quota according to the task feature cluster; generate personnel position adjustment suggestions according to the matching rules; S52, update the clustering results and matching rules regularly to generate dynamic management decision suggestions.

[0045] In one embodiment, the matching cluster such as the optimal person-task matching and the abnormal record is integrated to generate suggestions such as arranging rest for high-intensity cluster personnel or conducting training for abnormal personnel. Output decision reports including differentiated scheduling and task supplement plans to achieve human resource optimization and efficiency improvement. This generation process is based on matching cluster data to allocate supplemental tasks for low-load clusters, resulting in reduced management costs.

[0046] Embodiment 2

[0047] The present application provides a system for laboratory man-hour data analysis and resource optimization, configured to perform the method.

[0048] Example 3

[0049] The present application provides a computer readable storage medium having stored thereon a computer program which, when executed, implements the method.

[0050] The above has described embodiments of the present application, the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes are obvious to those skilled in the art without departing from the scope and spirit of the described embodiments.

Claims

1. A method for laboratory labor data analysis and resource optimization, comprising: The method comprises the following steps: S1, obtaining the work hour data and task data of laboratory personnel, respectively preprocessing to obtain the work hour distribution characteristics and task time consumption characteristics; S2, clustering the work hour data and task data by a clustering algorithm to generate personnel work hour mode clusters and task feature clusters; S3, constructing personnel and task matching clusters based on the personnel work hour mode clusters and the task feature clusters, and optimizing resource allocation.

2. The laboratory labor data analysis and resource optimization method of claim 1, wherein S1 comprises: S11, obtaining the work hour data of laboratory personnel, including personnel identification, work hour length and task type; preprocessing the work hour data to obtain work hour distribution characteristics, including daily average work hour, weekly fluctuation coefficient and overtime frequency; S12, obtaining task data, including task identification and single task time consumption; preprocessing the task data to obtain task time consumption characteristics, including time consumption mean and time consumption fluctuation coefficient; S13, standardizing the work hour data and task data to generate a unified magnitude feature set.

3. The laboratory labor data analysis and resource optimization method of claim 2, wherein S13 comprises: For work hour distribution characteristics and task time consumption characteristics, Z-score standardization method is used for processing respectively; The standardized feature set is processed by dimension reduction, and the dimensions strongly related to work hour mode and task characteristics are retained.

4. The laboratory man-hour data analysis and resource optimization method of claim 1, wherein S2 Comprise: S21, determine the optimal cluster number of the clustering algorithm by elbow method or contour coefficient method; S22, based on the feature set of the work hour data, generate personnel work hour mode clusters of fixed stable type, high intensity type of flexibility and low load fluctuation type; S23, based on the feature set of the task data, generate task feature clusters of fast standardized task, auxiliary task and high difficulty long time consumption task; S24, verify the features of the personnel work hour mode clusters and the task feature clusters to confirm that the cluster features are consistent with the business scenario.

5. The laboratory man-hour data analysis and resource optimization method of claim 1, wherein S3 Comprise: S31, based on the personnel work hour mode clusters and the task feature clusters, construct a two-dimensional feature space, and execute clustering to generate personnel and task matching clusters; S32, determine the matching rules and priorities according to the matching clusters, which are used to guide the allocation of personnel posts and task types; The matching relationship from high to low priority is as follows: allocation combination of fixed stable type personnel and fast standardized task, allocation combination of high intensity type personnel and high difficulty task, and allocation combination of low load fluctuation type and auxiliary task.

6. The laboratory labor data analysis and resource optimization method of claim 5, wherein In S32, The matching rules are as follows: for fixed stable type personnel, preferentially allocate fast standardized tasks with low time consumption fluctuation coefficient; for high intensity type personnel, preferentially allocate high difficulty tasks with high time consumption fluctuation coefficient; for low load fluctuation type personnel, allocate auxiliary tasks; Generate a priority list based on the matching rules to guide laboratory resource allocation.

7. The laboratory labor data analysis and resource optimization method of claim 1, wherein The method further comprises S4, identifying abnormal work hour records by calculating the distance between samples and cluster centroids, specifically: S41, for each work hour record, calculate the Euclidean distance between it and the centroid of the personnel work hour mode cluster to which it belongs; if the Euclidean distance of the work hour record is greater than the average distance within the cluster and exceeds the first preset threshold, mark the work hour record as an abnormal point; S42, for each task record, calculate the Euclidean distance between it and the centroid of the task feature cluster to which it belongs; If the Euclidean distance of the task record is greater than the average distance within the cluster and exceeds a second preset threshold, the task record is marked as an outlier point; S43, classifying the records marked as outlier points to distinguish false positive work hours and invalid overtime.

8. The laboratory labor data analysis and resource optimization method of claim 1, wherein The method further comprises S5, generating management decision suggestions according to the matching cluster and the anomaly detection result, specifically: S51, generating a differentiated scheduling strategy according to the personnel work hour mode cluster, generating a standardized work hour quota according to the task feature cluster, and generating a personnel post adjustment suggestion according to the matching rule; S52, periodically updating the clustering result and the matching rule to generate dynamic management decision suggestions.

9. A system for laboratory labor data analysis and resource optimization, characterized by, The system is configured to perform the method of any one of claims 1-8.

10. A computer readable storage medium having stored thereon a computer program, characterized in that When the program is executed, the method of any one of claims 1-8 is implemented.

Images