A Meter Reading Path Planning Method and System Based on Multi-Objective Constraints

Abstract

This application provides a method and system for meter reading route planning based on multi-objective constraints, relating to the field of meter reading route planning technology. The method includes: acquiring a set of meter reading tasks and extracting historical status data of geographical locations and associated metering devices as feature data; predicting the predicted operation time for each meter reading task based on the feature data; spatially clustering the set of meter reading tasks using an improved watershed algorithm to form several task clusters; performing initial route planning with the optimization objective of minimizing the total travel distance and the total predicted operation time, generating an initial operation sequence for each task cluster; executing the initial operation sequence, calculating the deviation between the actual operation time and the predicted operation time, and re-planning the route for subsequent unexecuted meter reading tasks to generate an updated operation sequence. This solves the technical problems of task allocation imbalance, static path fixation, and poor multi-objective coordination in existing meter reading route planning technologies.

Description

Technical Field

[0001] This invention relates to the field of meter reading route planning, and in particular to a meter reading route planning method and system based on multi-objective constraints. Background Technology

[0002] In electricity meter reading, the rational allocation of reading tasks and route planning directly affect reading efficiency, personnel scheduling costs, and user satisfaction.

[0003] However, traditional meter reading route planning often relies on manual experience or simple, single-goal-oriented algorithms, focusing solely on shortening travel distances while neglecting crucial factors such as variations in meter reading task durations, the suitability of personnel skills, and dynamic changes in the actual field environment. As meter reading operations expand and the spatial distribution of work areas becomes increasingly complex, traditional methods are gradually revealing problems such as unbalanced task allocation, high commuting costs due to path redundancy, low operational efficiency, and an inability to adapt to unforeseen circumstances. These issues make it difficult to meet the comprehensive demands of modern meter reading operations for improved efficiency, cost control, and balanced operations.

[0004] Therefore, there is an urgent need for a meter reading path planning method and system based on multi-objective constraints to break through the existing technical bottlenecks. Summary of the Invention

[0005] This invention addresses the technical problems of unbalanced task allocation, static and fixed paths, and poor multi-objective coordination in meter reading path planning in existing technologies, and provides a meter reading path planning method and system based on multi-objective constraints.

[0006] The technical solution of the present invention to solve the above-mentioned technical problems is as follows:

[0007] In a first aspect, the present invention provides a meter reading path planning method based on multi-objective constraints, including:

[0008] Obtain the set of meter reading tasks, extract the geographical location and historical status data of the associated metering device for each meter reading task, and use them as feature data for each meter reading task.

[0009] Based on the aforementioned feature data, the operation time for each meter reading task is predicted, thus obtaining the predicted operation time.

[0010] Based on the geographical location and the predicted operation time, the set of meter reading tasks is spatially clustered using an improved watershed algorithm to form several task clusters.

[0011] For each task cluster, with the optimization objective of minimizing the total travel distance and the total predicted operation time, a path planning algorithm is used for initial path planning to generate the initial operation sequence for each task cluster.

[0012] The initial job sequence is executed, the actual job duration is collected, the deviation between the actual job duration and the predicted job duration is calculated, the optimization parameters of the path planning algorithm are dynamically adjusted, the path planning is re-executed for subsequent unexecuted meter reading tasks, and an updated job sequence is generated.

[0013] Secondly, the present invention provides a meter reading path planning system based on multi-objective constraints, comprising:

[0014] The data acquisition module is used to acquire the set of meter reading tasks, extract the geographical location and historical status data of the associated metering device for each meter reading task, and use them as feature data for each meter reading task.

[0015] The duration prediction module is used to predict the operation duration of each meter reading task based on the feature data, and obtain the predicted operation duration.

[0016] The clustering module is used to spatially cluster the set of meter reading tasks based on the geographical location and the predicted operation duration using an improved watershed algorithm, forming several task clusters.

[0017] The path planning module is used to perform initial path planning for each task cluster, with the optimization goal of minimizing the total travel distance and the total predicted operation time, and to generate the initial operation sequence for each task cluster.

[0018] The correction and update module is used to execute the initial job sequence, collect the actual job duration, calculate the deviation between the actual job duration and the predicted job duration, dynamically adjust the optimization parameters of the path planning algorithm, re-plan the path for subsequent unexecuted meter reading tasks, and generate an updated job sequence.

[0019] The beneficial effects of this invention are:

[0020] Compared to existing technologies, this application first acquires a set of meter reading tasks and extracts the geographical location and historical status data of the associated metering devices for each task as feature data, providing reliable data support for subsequent work duration prediction and path planning. Secondly, based on the feature data, the work duration of each task is predicted, achieving accurate quantitative prediction of the work duration for each task and providing reliable data support for subsequent task clustering and multi-objective path planning based on an improved watershed algorithm. Thirdly, based on geographical location and predicted work duration, the set of meter reading tasks is spatially clustered using an improved watershed algorithm, forming several task clusters. These clusters ensure spatial proximity of tasks within the cluster and are adapted to the work capabilities of meter readers. Furthermore, for each task cluster, with the optimization objective of minimizing the total travel distance and total predicted work duration, a path planning algorithm is used for initial path planning, generating an initial work sequence for each task cluster, achieving optimal planning of the execution order of meter reading tasks within the cluster. Finally, the initial job sequence is executed, the actual job duration is collected, the deviation between the actual job duration and the predicted job duration is calculated, the optimization parameters of the path planning algorithm are dynamically adjusted, the path planning for subsequent unexecuted meter reading tasks is re-planned, and an updated job sequence is generated. This achieves real-time adaptation of path planning to the actual job scenario and effectively avoids the limitations of static planning.

[0021] Through the above technical solutions, this application integrates the entire process of meter reading task feature extraction, accurate prediction of operation time, multi-dimensional constraint clustering, and dynamic path optimization, achieving scientific and intelligent meter reading path planning. On the one hand, an improved watershed algorithm is used to cluster geographical locations and predicted operation time, ensuring that task clusters have spatial concentration and balanced operation load, effectively solving the problem of unbalanced traditional task allocation, reducing ineffective movement of meter readers, and improving the rationality of task allocation. On the other hand, a preset height difference threshold is dynamically set based on the meter reader's business ability coefficient to achieve accurate matching between tasks and personnel, avoiding low operation efficiency due to differences in personnel capabilities. Furthermore, the initial path planning with the dual objectives of minimizing total travel distance and total predicted operation time, combined with a closed-loop mechanism that dynamically adjusts and optimizes parameters based on actual operation deviations, not only reduces commuting redundancy costs and improves overall operation efficiency, but also achieves real-time adaptation of path planning to the actual on-site scenario, avoiding the limitations of static planning. This can meet the comprehensive needs of modern meter reading operations for efficiency improvement, cost control, and operation balance. Attached Figure Description

[0022] Figure 1 This is a flowchart illustrating the meter reading path planning method based on multi-objective constraints provided by the present invention.

[0023] Figure 2This is a schematic diagram of the structure of the meter reading path planning system based on multi-objective constraints provided by the present invention.

[0024] In the attached diagram, the components represented by each number are as follows:

[0025] Data acquisition module 11, duration prediction module 12, clustering and partitioning module 13, path planning module 14, and correction and update module 15. Detailed Implementation

[0026] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0027] In the description of this invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0028] In the description of this invention, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this invention is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use the invention. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that the invention can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid obscuring the description of the invention with unnecessary detail. Therefore, the invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed herein.

[0029] Example 1, as Figure 1 As shown, this embodiment of the invention provides a meter reading path planning method based on multi-objective constraints, including:

[0030] S10: Obtain the set of meter reading tasks, extract the geographical location and historical status data of the associated metering device for each meter reading task, and use them as feature data for each meter reading task.

[0031] Traditional meter reading route planning often relies solely on geographical distance, failing to adequately consider the time window constraints, priority differences, and the impact of metering device operating status on the operation. This results in unreasonable task allocation and poor route planning adaptability. Furthermore, geographical location directly determines the spatial distribution of tasks, and historical status data from associated metering devices can reflect the difficulty and risk of the operation.

[0032] To address the aforementioned issues, this application obtains a set of meter reading tasks and extracts the geographical location and historical status data of the associated metering devices for each meter reading task as feature data for each task.

[0033] Specifically, step S10 in the method includes:

[0034] Based on preset task time window rules and task priority classification rules, the meter reading tasks to be executed are dynamically selected from the meter reading task pool to form a set of meter reading tasks. The task time window rules are used to distinguish between rigid time window tasks and flexible time window tasks.

[0035] For each meter reading task in the set of meter reading tasks, the geographical location information of the meter reading task is parsed and extracted, wherein the geographical location information includes at least latitude and longitude coordinates;

[0036] For each meter reading task in the set of meter reading tasks, the historical status data of the metering device corresponding to the meter reading task is queried and extracted from the electricity metering device operation database. The historical status data includes at least fault records, operation and maintenance records and historical meter reading failure records.

[0037] The geographical location information corresponding to each meter reading task is combined with the historical status data to form the feature data of the meter reading task.

[0038] In this embodiment, based on preset task time window rules and task priority classification rules, meter reading tasks to be executed are dynamically selected from the meter reading task pool to form a set of tasks to be read. The task time window rules distinguish between rigid time window tasks and flexible time window tasks: rigid time window tasks are those that must be completed within a specific time period, such as user-arranged weekend meter readings or monthly fixed-date meter readings for important electricity users; flexible time window tasks are those without strict time restrictions and can be flexibly arranged within a certain period, such as quarterly meter readings for ordinary residential users. The task priority classification rules refer to the criteria for prioritizing tasks based on their urgency; for example, meter reading tasks following a fault report have a higher priority than regular meter reading tasks.

[0039] For example, based on preset task time window rules and task priority classification rules, combined with the current date, such as Monday, all rigid time window tasks that need to be completed within this week and flexible time window tasks marked as high priority, such as meter reading tasks that need to be reviewed after fault reporting, are included in the set of tasks to be read.

[0040] Secondly, for each meter reading task in the set of tasks to be read, the geographic location information of the task is parsed and extracted. This geographic location information includes at least latitude and longitude coordinates, which can be obtained by converting the detailed address using geocoding technology. For example, converting "XXX, Unit X, Building X, XX Community, XX District, XX City" to "116.40°E, 39.90°N". The geographic location information of the meter reading task is crucial spatial data for subsequent spatial clustering and path planning, ensuring accurate spatial positioning of the task.

[0041] Secondly, for each meter reading task in the set of tasks to be read, the historical status data of the corresponding metering device is queried and extracted from the electricity metering device operation database. This database stores the entire lifecycle operational data of all metering devices, covering factory information, installation information, fault records, maintenance records, and historical meter reading data. Historical status data includes at least fault records, maintenance records, and historical meter reading failure records: fault records include the time and handling results of faults such as short circuits or inaccurate metering; maintenance records include records of metering device repair, calibration, and replacement; and historical meter reading failure records include records where data collection was unsuccessful due to the user not being home or device malfunction. For each meter reading task, the corresponding fault records, maintenance records, and historical meter reading failure records are retrieved from the electricity metering device operation database using the unique identifier of the metering device, such as device ID or asset number. This historical status data directly reflects the historical operating status of the electricity metering device and the difficulty of meter reading, serving as a key basis for predicting operation time.

[0042] Finally, the geographical location information and historical status data corresponding to each meter reading task are structurally combined to form the feature data of the meter reading task. For example, the feature data of a certain meter reading task can be represented as: {latitude and longitude: (116.40°, 39.90°), fault record: short circuit fault in March 2024 (repaired), maintenance record: calibration in January 2024, historical meter reading failure record: 1 time (user not at home in February 2024)}. The structured feature data can be directly input into the job duration prediction model to ensure that the model can comprehensively obtain relevant task information.

[0043] In summary, compared to existing technologies, this application acquires a set of meter reading tasks and extracts the geographical location and historical status data of the associated metering devices for each task as feature data for each task. This allows for the selection of the current meter reading tasks and the extraction of key features for each task, providing reliable data support for subsequent operation duration prediction and route planning.

[0044] S20: Based on the feature data, predict the operation time of each meter reading task to obtain the predicted operation time.

[0045] Because the status and geographical location of the metering devices involved in different meter reading tasks vary, the operation time—that is, the total time from arriving at the meter reading location to completing the meter reading data collection—will also differ. For example, if the metering device for a certain meter reading task has multiple historical meter reading failure records and is located in a suburban area, it may require a longer operation time.

[0046] However, in traditional meter reading route planning, the operation time relies heavily on manual experience for estimation, without fully taking into account the geographical differences of the meter reading task and the historical status data of the associated metering devices. This results in a large deviation in the time estimation, which in turn causes the route planning to be out of touch with the actual operation requirements.

[0047] To address the aforementioned issues, this application predicts the operation time for each meter reading task based on the aforementioned feature data, thereby obtaining the predicted operation time.

[0048] Specifically, step S20 in the method includes:

[0049] Retrieve the pre-trained job duration prediction model;

[0050] Input the characteristic data of the meter reading task into the operation duration prediction model;

[0051] The operation duration prediction model is based on the feature data. By analyzing the historical status data of the electricity metering device and combining it with the geographical location information of the meter reading task, it performs reasoning calculations and outputs the predicted operation duration of the meter reading task.

[0052] In this embodiment, a pre-trained job duration prediction model is first retrieved. This job duration prediction model refers to a model built based on machine learning algorithms and trained using historical meter reading data, possessing the ability to predict job duration. The pre-trained job duration prediction model can be stored in the model database of the route planning system and can be directly retrieved through the route planning system interface when job duration prediction is needed.

[0053] Secondly, the characteristic data of the meter reading task is input into the operation time prediction model. Based on the characteristic data, the operation time prediction model analyzes the historical status data of the electricity metering device and combines it with the geographical location information of the meter reading task to perform inference calculations and output the predicted operation time of the meter reading task. For example, if there are multiple meter reading failure records in the historical status data, the operation time prediction model will determine that the meter reading task is more difficult and more communication or retry time should be reserved. If there are fault records, the operation time prediction model will consider the possible device debugging time after fault handling. At the same time, the operation time prediction model combines the geographical location information of the meter reading task to analyze traffic accessibility. For example, commuting time needs to be considered for suburban tasks, and parking time needs to be considered for tasks in densely populated residential areas. After comprehensive calculation through the built-in inference algorithm, the predicted operation time of the meter reading task is output.

[0054] For example, if the characteristic data of a meter reading task after a fault is repaired in a suburb is input into a pre-trained task duration prediction model, the output predicted task duration is 45 minutes, including 15 minutes of commuting time, 20 minutes of meter reading operation time, and 10 minutes of exception handling time.

[0055] Specifically, the construction process of the "job duration prediction model" includes:

[0056] Historical meter reading tasks are obtained from the meter reading work order database and the electricity metering device operation database. The historical meter reading tasks include the geographical location information of the historical meter reading tasks, the historical status data of the associated metering devices, and the corresponding actual operation time.

[0057] After denoising and normalizing the collected geographic location information and historical status data of associated metering devices, they are combined into a sample feature dataset, and the corresponding actual working time is used as the sample label to form a sample label dataset.

[0058] A job duration prediction model was built based on machine learning algorithms.

[0059] The job duration prediction model is trained under supervision using the sample feature dataset and the sample label dataset until it is verified to converge, thus obtaining the trained job duration prediction model.

[0060] In this embodiment, historical meter reading tasks are first obtained from the meter reading work order database and the electricity metering device operation database. These historical tasks include the geographical location information of the historical meter reading tasks, the historical status data of the associated metering devices, and the corresponding actual operation time. The meter reading work order database stores information on all completed meter reading work orders, including the task identifier, geographical location information, executor, and actual operation time for each work order; the electricity metering device operation database stores the historical status data of the corresponding metering devices.

[0061] For example, the data in the meter reading work order database can be associated with the data in the electricity metering device operation database through task identifiers, ensuring that each historical meter reading task contains complete geographical location information, historical status data of the associated metering device, and the corresponding actual operation duration. For instance, 100,000 historical meter reading task data from January 2023 to January 2024 can be obtained as the basic data for model training.

[0062] Secondly, the collected geographic location information and historical status data of associated metering devices are denoised and normalized, and then combined into a sample feature dataset. The corresponding actual operation time is used as the sample label to form a sample label dataset. Specifically, the collected historical meter reading task data is preprocessed. First, missing or abnormal data in the geographic location information and historical status data of associated metering devices are removed. Then, the preprocessed geographic location information and historical status data of associated metering devices are combined to form a sample feature dataset. At the same time, the actual operation time corresponding to each historical meter reading task is used as the sample label for that sample to form a sample label dataset.

[0063] For example, the sample characteristics of a certain historical task are {latitude and longitude: (116.41°, 39.91°), fault records: none, maintenance records: calibration in May 2023, historical meter reading failure records: 0 times}, and the corresponding sample label is 20 minutes.

[0064] Next, a job duration prediction model is constructed based on machine learning algorithms. Ideally, the gradient boosting decision tree algorithm can be chosen to construct the job duration prediction model. This algorithm is characterized by its ability to handle high-dimensional features, high prediction accuracy, and strong resistance to overfitting, making it suitable for job duration prediction problems that include multi-dimensional features such as geographical location and historical status.

[0065] For example, the gradient boosting decision tree algorithm is selected to construct the job duration prediction model, which mainly consists of three parts: a base learner, a loss function, and a gradient boosting iterative framework. The base learner uses a CART regression decision tree, with the input layer dimension of each decision tree matching the dimension of the meter reading task's feature data, such as latitude and longitude, fault records, maintenance frequency, and historical meter reading failures. The decision tree depth is set to 5-8 layers to balance fitting ability and overfitting risk. The loss function uses mean squared error (MSE) to quantify the deviation between the model's predicted job duration and the actual job duration. The gradient boosting iterative framework generates multiple decision trees in a forward step-by-step manner. The initial tree is generated based on the average job duration of all samples. Each subsequent tree is fitted to the prediction residual of the previous round model, i.e., the actual duration minus the predicted duration. The number of iterations is set to 100-200 rounds, and the model accuracy is verified using a validation set after each iteration.

[0066] Finally, the job duration prediction model is trained under supervision using the sample feature dataset and sample label dataset until validation convergence, resulting in a trained job duration prediction model. For example, the job duration prediction model can be trained using the following technical path: 1. Data preparation: Randomly divide the sample feature dataset and sample label dataset into training and validation sets in a 7:3 ratio; 2. Model training: Using feature data containing geographical location information and historical status data of associated metering devices in the training set as input features, and the corresponding actual job duration as the supervision label, the gradient descent optimization algorithm is used to iteratively adjust the parameters of the job duration prediction model. The decision tree depth is set to 5-8 layers, the learning rate to 0.01-0.1, and the minimum number of sample splits per tree to 10. Gradual optimization continuously reduces the error between the model's predicted values ​​and the training set labels. A periodic verification mechanism is introduced during training: every 10 training rounds, the job duration prediction model is verified using a validation set. The mean absolute error between the predicted values ​​and the validation set sample labels is calculated. When the mean absolute error of five consecutive verification rounds is less than a preset threshold, the model is considered to have converged, training is stopped, and the completed job duration prediction model is obtained. The preset threshold can be dynamically configured according to the accuracy requirements of meter reading operations, such as 3 minutes, as long as it meets the tolerance range of the duration error of conventional meter reading tasks, ensuring that the model's prediction results have practical application value.

[0067] In summary, compared to existing technologies, this application predicts the operation time of each meter reading task based on the aforementioned feature data, thus obtaining the predicted operation time. This achieves accurate quantitative prediction of the operation time of each meter reading task, providing reliable data support for subsequent task clustering and multi-objective path planning based on the improved watershed algorithm.

[0068] S30: Based on the geographical location and the predicted operation duration, the set of meter reading tasks is spatially clustered using an improved watershed algorithm to form several task clusters.

[0069] In traditional meter reading route planning, task division often relies solely on a single spatial distance factor without considering the differences in the operation time of meter reading tasks. This results in problems such as spatial dispersion of the generated task clusters, imbalance of operation load, and poor adaptability of personnel scheduling. Geographical location is the core foundation for ensuring the spatial concentration of tasks, and the predicted operation time directly determines the task execution load.

[0070] To address the aforementioned issues, this application uses an improved watershed algorithm to spatially cluster the set of meter reading tasks based on the geographical location and the predicted operation duration, forming several task clusters.

[0071] Specifically, step S30 in the method includes:

[0072] The geographical location information of each meter reading task is transformed into coordinates, and the latitude and longitude coordinates are mapped to coordinate points in a two-dimensional plane coordinate system. The predicted operation time of each meter reading task is converted into the terrain height value of the corresponding coordinate point to generate a task terrain map.

[0073] The current meter reader to be assigned is determined from the set of meter readers to be assigned, and the preset height difference threshold corresponding to the current meter reader to be assigned is calculated.

[0074] In the mission topographic map, the coordinate point with the smallest terrain height value is determined as the starting point of the simulated flooding process;

[0075] Starting from the starting point, the simulation process is carried out by gradually increasing the water level from the terrain height value at the starting point, and the terrain height value of the first coordinate point that the water level rises above is taken as the terrain height value of the first stage.

[0076] Calculate the height difference between the terrain height value at the starting point and the terrain height value in the first stage, and determine whether the height difference meets the preset height difference threshold corresponding to the current meter reader to be assigned. If so, continue to simulate the water level rise process.

[0077] If not, a watershed is generated at the terrain height value in the first stage, and the meter reading tasks corresponding to all coordinate points that have been submerged during this water level rise are grouped into a task cluster and assigned to the currently assigned meter reading personnel, thus completing one clustering.

[0078] Mark the coordinate points corresponding to the terrain height values ​​of the first stage as the new starting points, and remove the coordinate points corresponding to the assigned tasks from the task terrain map.

[0079] Continue simulating the rising water level until all meter reading tasks corresponding to all coordinate points on the task topographic map have been assigned, then end the simulation process.

[0080] In this embodiment, the geographical location information of each meter reading task is first transformed into coordinates, mapping latitude and longitude coordinates to coordinate points in a two-dimensional plane coordinate system. The predicted operation time of each meter reading task is then converted into the corresponding terrain elevation value of the coordinate point, generating a task topographic map. Specifically, the geographical location information of each meter reading task is first transformed into coordinates. For example, the Gauss-Kruger projection method can be used to map latitude and longitude coordinates to unique coordinate points in a two-dimensional plane coordinate system. This projection method effectively reduces the distortion of geographical coordinates during the plane mapping process, ensuring accurate calculation of spatial distances between tasks. Simultaneously, the predicted operation time is used as the terrain elevation value of the corresponding coordinate point. The longer the predicted operation time, the higher the terrain elevation value of the corresponding coordinate point. For example, if the predicted operation time is 30 minutes, the terrain elevation value of the corresponding coordinate point is 30. The two-dimensional plane coordinate points of each meter reading task are associated with the corresponding terrain elevation values, and integrated to generate a task topographic map. The task topographic map is a three-dimensional data model, where the X and Y axes represent two-dimensional plane coordinates corresponding to the actual geographical location, and the Z axis represents the terrain elevation value corresponding to the predicted operation time.

[0081] Secondly, the current meter reader to be assigned is determined from the set of personnel to be assigned, and the preset height difference threshold corresponding to the current meter reader is calculated. The set of personnel to be assigned refers to all meter readers who are currently idle and available to undertake meter reading tasks. This set is pre-stored in the personnel scheduling database and includes information such as personnel identification, operational capability coefficient, and current location. For example, the current meter reader to be assigned is first determined from the set of personnel to be assigned according to preset rules. Then, based on parameters such as the operational capability coefficient of the current meter reader to be assigned and the standard deviation of the terrain height value on the task topography map, the preset height difference threshold corresponding to the current meter reader to be assigned is calculated. The preset height difference threshold is the core basis for determining whether meter reading tasks belong to the same cluster, reflecting the range of differences in working time that the current meter reader to be assigned can withstand. The preset rules can be sorted in descending order based on operational capability coefficient, ascending order based on location distance, etc.

[0082] Secondly, in the task topographic map, the coordinate point with the smallest topographic height value is determined as the starting point of the simulated flooding process. Specifically, in the task topographic map, the topographic height values ​​of all coordinate points are traversed, and the coordinate point with the smallest topographic height value is selected. This coordinate point corresponds to the meter reading task with the shortest predicted operation time, and it is determined as the starting point of the simulated flooding process. The reason for choosing the coordinate point with the smallest topographic height value as the starting point is that starting the clustering from the task with the shortest operation time ensures that the operation load of the initial cluster is low, and gradually expanding to tasks with longer operation times avoids excessive task load in a single cluster.

[0083] Furthermore, starting from the initial point, the simulation demonstrates the gradual rise of water level from the initial terrain elevation value. The terrain elevation value of the first coordinate point submerged by the rising water level is taken as the first-stage terrain elevation value. Specifically, starting from the initial point, a preset fixed step size is used, which can be 1% of the range of terrain elevation values ​​on the task terrain map. The range is the difference between the maximum and minimum terrain elevation values. Simultaneously, during the water level rise, the first unmarked coordinate point encountered is monitored in real time, and the terrain elevation value of that coordinate point is taken as the first-stage terrain elevation value.

[0084] For example, if the terrain height value at the starting point is 10, the corresponding predicted operation time is 10 minutes, and the terrain height value range on the task terrain map is 50, then the preset step size can be set to 0.5. The water level starts from 10 and gradually rises in steps of 0.5. When the water level rises to 12, it first touches a new coordinate point. The terrain height value of this coordinate point is 12, which is the terrain height value of the first stage.

[0085] Further, the height difference between the terrain height value at the starting point and the terrain height value at the first stage is calculated, and it is determined whether the height difference meets the preset height difference threshold corresponding to the meter reader to be assigned. If so, the simulation of the water level rise process continues. Specifically, the height difference between the terrain height value at the starting point and the terrain height value at the first stage is calculated. For example, the height difference between 10 and 12 is 2. The preset height difference threshold corresponding to the meter reader to be assigned is retrieved. For example, 3. It is determined whether the height difference meets the preset height difference threshold: if the height difference ≤ the preset height difference threshold, such as 2≤3, it means that the difference in operation time between the task corresponding to the new coordinate point and the task at the starting point is within the acceptable range for the personnel. And because it is the first contact during the water level rise process, it means that the two meter reading tasks are spatially adjacent. Therefore, no watershed is generated, and the simulation of the water level rise process continues according to the preset step size. The coordinate point is marked as the associated point of the area to which the starting point belongs.

[0086] Conversely, if not, a watershed is generated at the terrain height value in the first stage. All meter reading tasks corresponding to coordinates submerged during this water level rise are grouped into a single task cluster and assigned to the currently assigned meter readers, completing one clustering operation. Specifically, if the height difference exceeds a preset height difference threshold, it indicates a significant difference in work time between the two tasks. Grouping them into the same cluster would lead to uneven workload for the currently assigned meter readers. Therefore, a watershed is generated at the midpoint of the shortest path between the coordinates corresponding to the terrain height value in the first stage and the starting point. After generating the watershed, all meter reading tasks corresponding to coordinates submerged during this water level rise are grouped into a single task cluster and assigned to the currently assigned meter readers, completing one clustering operation.

[0087] Furthermore, the coordinates corresponding to the terrain height values ​​in the first stage are marked as new starting points, while the coordinates corresponding to the assigned tasks are removed from the task terrain map. Specifically, the watershed serves as the spatial boundary between the two clusters of tasks, separating the area to which the starting point belongs from the area to which the coordinates of the first stage belong. The coordinates corresponding to the terrain height values ​​in the first stage are marked as new starting points, the water level is updated to the terrain height values ​​corresponding to the new starting points, and the simulation of water level rise continues. The coordinates corresponding to the assigned tasks are removed from the task terrain map to avoid duplicate allocation during subsequent clustering processes.

[0088] Finally, the simulation of the rising water level continues until all meter reading tasks corresponding to all coordinate points on the task topographic map have been assigned, at which point the simulation ends. Specifically, starting from a new point, the above steps of water level rise, height difference determination, watershed generation or further water level rise, and task cluster assignment are repeated until all meter reading tasks corresponding to all coordinate points on the task topographic map have been assigned, at which point the simulation ends. At this point, several task clusters are obtained, each corresponding to one meter reader.

[0089] Specifically, the process of determining the "preset height difference threshold corresponding to the currently assigned meter reader" includes:

[0090] Based on the terrain height values ​​of all coordinate points in the task terrain map, the standard deviation of the terrain height values ​​is calculated.

[0091] From the set of meter readers to be assigned, determine the current meter reader to be assigned according to preset rules, and obtain the business capability coefficient of the current meter reader to be assigned.

[0092] Based on the standard height difference threshold, the standard deviation of the terrain height value, and the business capability coefficient of the current meter reader to be assigned, the preset height difference threshold for this division process is obtained through weighted calculation.

[0093] Iterate through the set of meter readers to be assigned and calculate the preset height difference threshold for each meter reader in turn.

[0094] In this embodiment, the standard deviation of the terrain height values ​​is first calculated based on the terrain height values ​​of all coordinate points on the mission terrain map. Specifically, the terrain height values ​​of all coordinate points on the mission terrain map are iterated through, and the standard deviation is calculated using the following formula: ,in, Let μ be the terrain elevation value for a single coordinate point, n be the average of all terrain elevation values, and n be the total number of coordinate points. The standard deviation of the terrain elevation values ​​is calculated. The standard deviation of the terrain elevation values ​​reflects the dispersion of the predicted operation time for all meter reading tasks. The larger the standard deviation of the terrain elevation values, the more significant the difference in operation time, and a more refined preset height difference threshold needs to be set.

[0095] Secondly, from the set of meter readers to be assigned, the current meter readers to be assigned are determined according to preset rules, and the business capability coefficient of the current meter readers to be assigned is obtained. The preset rules refer to allocation rules such as prioritizing the assignment of the nearest personnel or prioritizing tasks with large differences in work duration, which can be dynamically set by those skilled in the art according to the actual situation. For example, if the preset rule is to prioritize the assignment of the nearest personnel, the current meter readers to be assigned are determined based on all meter readers currently in an idle state who can undertake meter reading tasks, and the business capability coefficient of the current meter readers to be assigned is calculated and obtained.

[0096] Next, based on the standard height difference threshold, the standard deviation of terrain height, and the business capability coefficient of the meter readers to be assigned, the preset height difference threshold for this division process is obtained through weighted calculation. The standard height difference threshold is a benchmark threshold pre-set based on industry meter reading clustering practices. The unit is consistent with the terrain height value. Preferably, the standard height difference threshold can be set to 0.5-1.2, or it can be dynamically adjusted according to the meter reading area type, such as densely populated urban areas, remote rural areas, etc., and the actual application scenario such as the actual operation mode. The weighted calculation formula can be set as: preset height difference threshold = standard height difference threshold × a + standard deviation of terrain height value × b + business capability coefficient × c, where a, b, and c are preset weight coefficients and satisfy a + b + c = 1. Preferably, a = 0.3, b = 0.4, and c = 0.3. This weight allocation takes into account both the stability of the standard height difference threshold and the impact of the dispersion of operation time and personnel capability. Those skilled in the art can dynamically adjust the values ​​of a, b, and c according to the actual meter reading efficiency requirements, personnel scheduling needs, etc., without being limited to the above preferred example.

[0097] For example, if the preset weights of the standard height difference threshold, the standard deviation of terrain height, and the business ability coefficient of the currently assigned meter reader are a=0.3, b=0.4, and c=0.3 respectively, when the standard height difference threshold is 1.0, the standard deviation of terrain height is 5, and the business ability coefficient is 1.116, the preset height difference threshold = 1.0×0.3+5×0.4+1.116×0.3=2.63.

[0098] Finally, the set of meter readers to be assigned is iterated through, and the preset height difference threshold for each meter reader is calculated sequentially. Specifically, following the steps above, each meter reader in the set of meter readers to be assigned is iterated through, and the preset height difference threshold for each person is calculated sequentially. For example, if there are 3 meter readers in the set of meter readers to be assigned, with business ability coefficients of 1.5, 1.2, and 0.9 respectively, then the calculated preset height difference thresholds are 2.75, 2.5, and 2.2 respectively. The stronger the business ability of the person, the larger the preset height difference threshold, and the wider the range of differences in working time that can be tolerated.

[0099] Furthermore, the "obtaining the business capability coefficient of the currently assigned meter reader" includes:

[0100] Obtain the seniority data, average completion efficiency data of historical meter reading tasks, and historical task familiarity data of all meter readers in the set of meter readers to be assigned, as well as the geographical area corresponding to the task topographic map. Calculate the average seniority, average completion efficiency, and average familiarity respectively. The corresponding geographical area is an area with the task cluster center as the center and a preset length as the radius.

[0101] For the currently assigned meter readers, calculate the ratio of seniority data to the average seniority as a seniority factor; calculate the ratio of average completion efficiency data to the average completion efficiency as an efficiency factor; and calculate the ratio of historical task familiarity data to the average familiarity as a familiarity factor.

[0102] The seniority factor, efficiency factor, and familiarity factor are weighted and summed according to preset weights to calculate the business competence coefficient of the currently assigned meter reader.

[0103] In this embodiment, the system first retrieves the seniority data, average completion efficiency data of historical meter reading tasks, and historical task familiarity data within the corresponding geographical area of ​​the task topographic map from the personnel management database. The average seniority, average completion efficiency, and average familiarity are then calculated. The corresponding geographical area is a region with the task cluster center as its center and a preset length as its radius. The preset length can be dynamically set based on the spatial density of tasks within the task cluster, the total number of tasks, and the maximum daily movement radius of the meter readers. Preferably, the preset length for task clusters in densely populated urban areas can be set to 3 kilometers, and the preset length for task clusters in remote rural areas can be set to 8 kilometers, ensuring reasonable regional coverage and operational accessibility.

[0104] For example, the length of service data of a meter reader in the set of meter readers to be assigned can be directly obtained from the personnel management database, such as 3 years. The ratio of the total number of historical meter reading tasks completed by the meter reader in the past 12 months to the total time spent is calculated as the average completion efficiency data of historical meter reading tasks. For example, if the meter reader completed 360 tasks in the past 12 months with a total time of 720 hours, then the average completion efficiency data = 360 / 720 = 0.5 tasks / hour. The ratio of the total number of tasks completed by the meter reader in the past 3 years to the number of tasks completed in the corresponding geographical area is calculated as the historical task familiarity data in the corresponding geographical area of ​​the task topography map. For example, if the meter reader completed a total of 1000 tasks in the past three years, of which 400 were completed in the geographical area corresponding to the task topography map, then the historical task familiarity data = 400 / 1000 = 0.4. Following the same method and calculation logic, obtain the seniority data, average completion efficiency data of historical meter reading tasks, and historical task familiarity data of all meter readers in the set of meter readers to be assigned, as well as the average value of the three types of data in the corresponding geographical area of ​​the task topographic map. Calculate the average seniority, average completion efficiency, and average familiarity for each type of data. For example, the average seniority is 4 years, the average completion efficiency is 0.4 meters / hour, and the average familiarity is 0.3.

[0105] Secondly, for the current meter readers awaiting assignment, the ratio of seniority data to average seniority is calculated as the seniority factor; the ratio of average completion efficiency data to average completion efficiency is calculated as the efficiency factor; and the ratio of historical task familiarity data to average familiarity is calculated as the familiarity factor. For example, if the current meter readers awaiting assignment have 3 years of seniority, an average completion efficiency of 0.5 meters / hour, and a historical task familiarity of 0.4, while the average seniority is 4 years, the average completion efficiency is 0.4 meters / hour, and the average familiarity is 0.3, then the seniority factor = 3 / 4 = 0.75, the efficiency factor = 0.5 / 0.4 = 1.25, and the familiarity factor = 0.4 / 0.3 ≈ 1.33. These factors are all relative values, reflecting the current meter readers' advantages or disadvantages relative to the average level.

[0106] Finally, the seniority factor, efficiency factor, and familiarity factor are weighted and summed according to preset weights to calculate the business competence coefficient of the meter reader to be assigned. The preset weights can be dynamically set according to the actual application scenario. Preferably, the sum of the weights of the seniority factor, efficiency factor, and familiarity factor is 1, because work efficiency directly affects the ability to complete tasks; therefore, the efficiency factor weight can be set to the highest. For example, the weights of the seniority factor, efficiency factor, and familiarity factor can be set to 0.3, 0.5, and 0.2, respectively.

[0107] For example, if the weights of the seniority factor, efficiency factor, and familiarity factor are 0.3, 0.5, and 0.2 respectively, and the seniority factor is 0.75, the efficiency factor is 1.25, and the familiarity factor is 1.33, then the business competence coefficient of the meter reader to be assigned is 0.75×0.3 + 1.25×0.5 + 1.33×0.2 = 1.116. The business competence coefficient comprehensively reflects the meter reader's proficiency, historical task completion efficiency, and familiarity with the geographical area corresponding to the task. When calculating the preset height difference threshold, the business competence coefficient is incorporated as a weighting factor into the weighted calculation formula, ensuring a precise match between the preset height difference threshold and the meter reader's actual business competence: meter readers with stronger business competence are assigned a higher preset height difference threshold and can handle task clusters with greater differences in work duration; conversely, those with weaker business competence are assigned to task clusters with more concentrated work durations. Ultimately, this achieves dynamic matching between task cluster division and personnel competence, ensuring balanced workload and execution efficiency.

[0108] In summary, compared to existing technologies, this application, based on the geographical location and the predicted operation time, uses an improved watershed algorithm to spatially cluster the set of meter reading tasks, forming several task clusters. Thus, the improved watershed algorithm divides scattered meter reading tasks into several spatially concentrated and workload-balanced task clusters, mapping the predicted operation time to terrain height. This ensures that the clustering results simultaneously satisfy spatial concentration and workload balance, achieving spatial concentration and workload-balanced clustering of the meter reading tasks. The resulting task clusters guarantee spatial proximity within each cluster and are adapted to the operational capabilities of meter readers, providing efficient basic units for subsequent accurate personnel allocation and multi-objective path planning, thereby improving the overall efficiency and rationality of meter reading task execution.

[0109] S40: For each task cluster, with the optimization objective of minimizing the total travel distance and the total predicted operation time, a path planning algorithm is used to perform initial path planning and generate the initial operation sequence for each task cluster.

[0110] In traditional meter reading route planning, task clusters are usually not optimized for multi-objective collaboration within the clusters. Simply considering travel distance or operation time can easily lead to path redundancy and task execution timeouts. Furthermore, the spatial distribution and operation time differences of tasks within the clusters still affect actual execution efficiency. The total travel distance is directly related to operation cost and commuting time, while the total predicted operation time determines personnel workload and task completion cycle. Both of these are core factors affecting meter reading efficiency.

[0111] To address the aforementioned issues, this application employs a path planning algorithm for each task cluster, with the optimization objective of minimizing the total travel distance and the total predicted operation time, to perform initial path planning and generate the initial operation sequence for each task cluster.

[0112] Specifically, step S40 in the method includes:

[0113] Obtain the geographical location information of all meter reading tasks within the task cluster and predict the operation time;

[0114] The first optimization objective is to minimize the total travel distance, and the second optimization objective is to minimize the total predicted operation time. A multi-objective optimization model is constructed based on the first and second optimization objectives.

[0115] The multi-objective optimization model is solved using a path planning algorithm to generate an initial job sequence covering all meter reading tasks within the task cluster.

[0116] In this embodiment, the geographical location information and predicted operation time of all meter reading tasks within the task cluster are first obtained. Specifically, for each generated task cluster, the geographical location information and predicted operation time of all meter reading tasks within the cluster are obtained to ensure that the data is consistent with the data used during clustering and partitioning, and to avoid unreasonable path planning due to data deviation.

[0117] Secondly, a multi-objective optimization model is constructed based on minimizing the total travel distance as the first optimization objective and minimizing the total predicted operation time as the second optimization objective. The total travel distance refers to the total mileage traveled by the meter reader from the starting point to the destination after completing all meter reading tasks according to the work sequence; the total predicted operation time refers to the sum of the predicted operation times of all meter reading tasks within the cluster. The constraints of the multi-objective optimization model include: each meter reading task is executed only once, the work sequence must cover all meter reading tasks within the cluster, and the total daily operation time of the meter reader does not exceed a preset upper limit. The objective function of the multi-objective optimization model can be expressed as: min(ω1 × total travel distance + ω2 × total predicted operation time), where ω1 and ω2 are initial weight coefficients, and ω1 + ω2 = 1. For example, the initial weight coefficients can be set to: ω1 = 0.6, ω2 = 0.4, prioritizing travel distance optimization.

[0118] Finally, a path planning algorithm is used to solve the multi-objective optimization model, generating an initial job sequence covering all meter reading tasks within the task cluster. For example, the path planning algorithm can be an improved genetic algorithm, simulated annealing, or ant colony optimization. For instance, by selecting an improved genetic algorithm, the job sequence is encoded as chromosomes, and the fitness function of the multi-objective optimization model is used as the objective function. Through iterative optimization operations such as selection, crossover, and mutation, the chromosome with the highest fitness is obtained, corresponding to the optimal initial job sequence. For example, a task cluster contains 5 meter reading tasks {meter reading task A, meter reading task B, meter reading task C, meter reading task D, and meter reading task E}. After solving the path planning algorithm, the generated initial job sequence is: meter reading task A → meter reading task C → meter reading task B → meter reading task E → meter reading task D. The total travel distance corresponding to the initial job sequence is 20 kilometers, and the total predicted job time is 3.5 hours, satisfying the multi-objective optimization requirement of minimizing the total travel distance and the total predicted job time.

[0119] In summary, compared to existing technologies, this application, for each task cluster, uses a path planning algorithm for initial path planning, with the optimization objectives of minimizing the total travel distance and the total predicted operation time, to generate the initial operation sequence for each task cluster. This achieves optimal planning of the execution order of meter reading tasks within the task cluster. Under the dual constraints of minimizing the total travel distance and the total predicted operation time, the work process of meter readers is clearly defined, improving the execution efficiency of meter reading tasks within the task cluster, reducing commuting costs, and ensuring that tasks are completed on schedule.

[0120] S50: Execute the initial job sequence, collect the actual job duration, calculate the deviation between the actual job duration and the predicted job duration, dynamically adjust the optimization parameters of the path planning algorithm, re-plan the path for subsequent unexecuted meter reading tasks, and generate an updated job sequence.

[0121] Traditional meter reading route planning is mostly static planning mode. After the initial operation sequence is generated, there is no dynamic adaptation mechanism. In actual operation, it is affected by uncertain factors such as traffic conditions, sudden abnormalities of metering devices, and on-site communication. The predicted operation time is prone to deviate from the actual operation time. If the initial planning parameters are used, it will lead to subsequent task execution timeouts, redundant travel distances, or decreased efficiency.

[0122] To address the aforementioned issues, this application executes the initial job sequence, collects the actual job duration, calculates the deviation between the actual job duration and the predicted job duration, dynamically adjusts the optimization parameters of the path planning algorithm, re-plans the path for subsequent unexecuted meter reading tasks, and generates an updated job sequence.

[0123] Specifically, step S50 in the method includes:

[0124] During the process of meter readers performing meter reading tasks according to the initial work sequence, the actual working time of each completed meter reading task is recorded in real time.

[0125] Calculate the deviation between the actual operation time and the corresponding predicted operation time, and calculate the average deviation of all completed tasks;

[0126] Based on the average deviation value, the weight parameters used to balance the two optimization objectives of total travel distance and total operation time in the path planning algorithm are dynamically adjusted.

[0127] Based on the revised weight parameters, an update job sequence is generated for all meter reading tasks that have not yet started.

[0128] In this embodiment, the actual working time of each completed meter reading task is recorded in real time during the process of meter reading personnel performing meter reading tasks according to the initial work sequence. Specifically, during the process of meter reading personnel performing meter reading tasks according to the initial work sequence, the actual working time of each completed meter reading task can be recorded in real time through meter reading terminals, such as mobile apps with positioning and timing functions, handheld meter reading devices, etc. The starting point of the actual working time is the time when the meter reader arrives at the location of the metering device corresponding to the task, and the ending point of the timing is the time when the meter reading data is successfully uploaded to the system, ensuring the accuracy of the timing.

[0129] For example, when meter readers perform meter reading tasks according to the initial work sequence: meter reading task A → meter reading task C → meter reading task B → meter reading task E → meter reading task D, they can use a mobile APP with positioning and timing functions to obtain the actual working time of completed meter reading task A as 25 minutes and completed task C as 28 minutes.

[0130] Secondly, calculate the deviation between the actual operation time and the corresponding predicted operation time, and calculate the average deviation of all completed tasks. Specifically, for each completed meter reading task, calculate the deviation between the actual operation time and the corresponding predicted operation time. The formula for calculating the deviation is: Deviation = (Actual operation time - Predicted operation time) / Predicted operation time × 100%. The deviation is a relative deviation with a positive or negative sign. A positive value indicates that the actual operation time exceeds the predicted operation time, that is, the actual execution time is longer, and a negative value indicates that the actual operation time is shorter than the predicted operation time, that is, the actual execution time is shorter.

[0131] For example, if the actual working time of meter reading task A is 25 minutes and the predicted working time is 20 minutes, and the actual working time of task C is 28 minutes and the predicted working time is 30 minutes, then the deviation value of meter reading task A = (25-20) / 20×100%=25%, indicating that the actual working time of meter reading task A exceeds the predicted working time. The deviation value of meter reading task C = (28-30) / 30×100%≈-6.7%, indicating that the actual working time of meter reading task A is shorter than the predicted working time. Then, the average deviation value of meter reading tasks A and C is calculated as (25%-6.7%) / 2 = 9.15%. If the average deviation value is positive, it means that the overall actual working time of the completed tasks is longer than the overall predicted working time. If the average deviation value is negative, it means that the overall actual working time of the completed tasks is shorter than the overall predicted working time.

[0132] Next, based on the average deviation value, the weight parameters used in the path planning algorithm to balance the two optimization objectives of total travel distance and total operation time are dynamically adjusted. Specifically, based on the calculated average deviation value, the weight parameters used in the path planning algorithm to balance the two optimization objectives of total travel distance and total operation time, namely ω1 and ω2 in step S40, are dynamically adjusted.

[0133] For example, the correction rule can be set as follows: If the average deviation value > a preset positive threshold (e.g., 10%, which can be dynamically configured according to the accuracy requirements of meter reading operations and task type), it indicates that the actual operation time generally exceeds expectations. The weight of the total predicted operation time ω2 needs to be increased, and the weight of the total travel distance ω1 needs to be decreased to avoid subsequent tasks exceeding the time limit due to insufficient duration. The correction magnitude is: ω2' = ω2 + |average deviation value| × 0.1, ω1' = 1 - ω2'. For example, if the original ω1 = 0.6, ω2 = 0.4, and the average deviation value is 9.15%, then ω2' = 0.4 + 9.15% × 0.1 = 0.41, ω1' = 1 - 0.41 = 0.59. If the average deviation value < a preset negative threshold (e.g., -10%, which can be dynamically configured according to the actual scenario), it indicates that the actual operation time is generally shorter than the predicted operation time. The weight of the total travel distance ω1 can be appropriately increased, and the weight of the total predicted operation time ω2 can be decreased to further optimize travel efficiency. If the deviation value is within the range of the preset positive threshold and the preset negative threshold, then keep the original ω1 and ω2 unchanged and no adjustment is needed.

[0134] Finally, based on the revised weight parameters, an update job sequence is generated for all meter reading tasks that have not yet started. Specifically, based on the revised weight parameters, a multi-objective optimization model is reconstructed, and the same algorithm as the initial path planning is used to solve all meter reading tasks that have not yet started, generating an update job sequence.

[0135] For example, for meter reading tasks B, E, and D that were not executed according to the initial task sequence, a multi-objective optimization model is reconstructed based on the corrected weight parameters ω1'=0.59 and ω2'=0.41. The updated task sequence generated after solving the problem using the path planning algorithm is: meter reading task B → meter reading task D → meter reading task E. The updated task sequence accurately adapts to the weight adjustment requirements. Under the premise of ensuring that the total task time is controllable, it further reduces the invalid commuting mileage, ensuring the execution efficiency and time controllability of subsequent unexecuted tasks, and realizing a closed loop of dynamic optimization.

[0136] In summary, compared to existing technologies, this application executes the initial job sequence, collects the actual job duration, calculates the deviation between the actual job duration and the predicted job duration, dynamically adjusts the optimization parameters of the path planning algorithm, and re-plans the paths for subsequent unexecuted meter reading tasks to generate an updated job sequence. Thus, by collecting the actual job duration and calculating its deviation from the predicted job duration, dynamically adjusting the optimization parameters of the path planning algorithm, and re-planning the paths for subsequent unexecuted meter reading tasks, real-time adaptation of path planning to the actual job scenario is achieved. This effectively avoids the limitations of static planning and ensures the controllability of subsequent task execution time and overall job efficiency.

[0137] In summary, the embodiments of this application have at least the following technical effects:

[0138] Compared to existing technologies, this application first obtains a set of meter reading tasks, extracts the geographical location and historical status data of the associated metering devices for each task, and uses this data as feature data for each task. In this way, the current meter reading tasks to be executed are selected, and the key features of each task are extracted, providing reliable data support for subsequent operation duration prediction and route planning.

[0139] Secondly, based on the aforementioned feature data, this application predicts the operation time of each meter reading task, thus obtaining the predicted operation time. This achieves accurate quantitative prediction of the operation time of each meter reading task, providing reliable data support for subsequent task clustering and multi-objective path planning based on the improved watershed algorithm.

[0140] Furthermore, based on the aforementioned geographical location and the predicted operation duration, this application uses an improved watershed algorithm to spatially cluster the set of meter reading tasks, forming several task clusters. Thus, the improved watershed algorithm divides scattered meter reading tasks into several spatially concentrated and workload-balanced task clusters, mapping the predicted operation duration to terrain height. This ensures that the clustering results simultaneously satisfy spatial concentration and workload balance, achieving spatial concentration and workload-balanced clustering of the meter reading tasks. The resulting task clusters guarantee spatial proximity within each cluster and are adapted to the operational capabilities of meter readers, providing efficient basic units for subsequent accurate personnel allocation and multi-objective path planning, thereby improving the overall efficiency and rationality of meter reading task execution.

[0141] Furthermore, for each task cluster, this application employs a path planning algorithm to perform initial path planning, aiming to minimize the total travel distance and the total predicted operation time, thereby generating the initial operation sequence for each task cluster. This achieves optimal planning of the execution order of meter reading tasks within the task cluster. Under the dual constraints of minimizing the total travel distance and the total predicted operation time, the work process of meter readers is clearly defined, improving the execution efficiency of meter reading tasks within the task cluster, reducing commuting costs, and ensuring that tasks are completed on schedule.

[0142] Finally, this application executes the initial job sequence, collects the actual job duration, calculates the deviation between the actual job duration and the predicted job duration, dynamically adjusts the optimization parameters of the path planning algorithm, and re-plans the paths for subsequent unexecuted meter reading tasks to generate an updated job sequence. In this way, by collecting the actual job duration and calculating its deviation from the predicted job duration, dynamically adjusting the optimization parameters of the path planning algorithm, and re-planning the paths for subsequent unexecuted meter reading tasks, real-time adaptation of path planning to the actual job scenario is achieved. This effectively avoids the limitations of static planning and ensures the controllability of the execution time of subsequent tasks and overall job efficiency.

[0143] Through the above technical solutions, this application integrates the entire process of meter reading task feature extraction, accurate prediction of operation time, multi-dimensional constraint clustering, and dynamic path optimization, achieving scientific and intelligent meter reading path planning. On the one hand, an improved watershed algorithm is used to cluster geographical locations and predicted operation time, ensuring that task clusters have spatial concentration and balanced operation load, effectively solving the problem of unbalanced traditional task allocation, reducing ineffective movement of meter readers, and improving the rationality of task allocation. On the other hand, a preset height difference threshold is dynamically set based on the meter reader's business ability coefficient to achieve accurate matching between tasks and personnel, avoiding low operation efficiency due to differences in personnel capabilities. Furthermore, the initial path planning with the dual objectives of minimizing total travel distance and total predicted operation time, combined with a closed-loop mechanism that dynamically adjusts and optimizes parameters based on actual operation deviations, not only reduces commuting redundancy costs and improves overall operation efficiency, but also achieves real-time adaptation of path planning to the actual on-site scenario, avoiding the limitations of static planning. This can meet the comprehensive needs of modern meter reading operations for efficiency improvement, cost control, and operation balance.

[0144] Example 2, as Figure 2 As shown, based on the same inventive concept as the meter reading path planning method based on multi-objective constraints provided in Embodiment 1, this embodiment of the invention also provides a meter reading path planning system based on multi-objective constraints, including:

[0145] Data acquisition module 11 is used to acquire the set of meter reading tasks, extract the geographical location and historical status data of the associated metering device for each meter reading task, and use them as feature data for each meter reading task.

[0146] The duration prediction module 12 is used to predict the operation duration of each meter reading task based on the feature data, and obtain the predicted operation duration.

[0147] Clustering module 13 is used to spatially cluster the set of meter reading tasks based on the geographical location and the predicted operation duration using an improved watershed algorithm, forming several task clusters.

[0148] The path planning module 14 is used to perform initial path planning for each task cluster with the optimization objective of minimizing the total travel distance and the total predicted operation time, and to generate the initial operation sequence for each task cluster.

[0149] The correction and update module 15 is used to execute the initial job sequence, collect the actual job duration, calculate the deviation between the actual job duration and the predicted job duration, dynamically adjust the optimization parameters of the path planning algorithm, re-plan the path for subsequent unexecuted meter reading tasks, and generate an updated job sequence.

[0150] Specifically, the data acquisition module 11 is used for:

[0151] Based on preset task time window rules and task priority classification rules, the meter reading tasks to be executed are dynamically selected from the meter reading task pool to form a set of meter reading tasks. The task time window rules are used to distinguish between rigid time window tasks and flexible time window tasks.

[0152] For each meter reading task in the set of meter reading tasks, the geographical location information of the meter reading task is parsed and extracted, wherein the geographical location information includes at least latitude and longitude coordinates;

[0153] For each meter reading task in the set of meter reading tasks, the historical status data of the metering device corresponding to the meter reading task is queried and extracted from the electricity metering device operation database. The historical status data includes at least fault records, operation and maintenance records and historical meter reading failure records.

[0154] The geographical location information corresponding to each meter reading task is combined with the historical status data to form the feature data of the meter reading task.

[0155] Specifically, the duration prediction module 12 is used for:

[0156] Retrieve the pre-trained job duration prediction model;

[0157] Input the characteristic data of the meter reading task into the operation duration prediction model;

[0158] The operation duration prediction model is based on the feature data. By analyzing the historical status data of the electricity metering device and combining it with the geographical location information of the meter reading task, it performs reasoning calculations and outputs the predicted operation duration of the meter reading task.

[0159] Specifically, the construction process of the "job duration prediction model" includes:

[0160] Historical meter reading tasks are obtained from the meter reading work order database and the electricity metering device operation database. The historical meter reading tasks include the geographical location information of the historical meter reading tasks, the historical status data of the associated metering devices, and the corresponding actual operation time.

[0161] After denoising and normalizing the collected geographic location information and historical status data of associated metering devices, they are combined into a sample feature dataset, and the corresponding actual working time is used as the sample label to form a sample label dataset.

[0162] A job duration prediction model was built based on machine learning algorithms.

[0163] The job duration prediction model is trained under supervision using the sample feature dataset and the sample label dataset until it is verified to converge, thus obtaining the trained job duration prediction model.

[0164] Specifically, the clustering module 13 is used for:

[0165] The geographical location information of each meter reading task is transformed into coordinates, and the latitude and longitude coordinates are mapped to coordinate points in a two-dimensional plane coordinate system. The predicted operation time of each meter reading task is converted into the terrain height value of the corresponding coordinate point to generate a task terrain map.

[0166] The current meter reader to be assigned is determined from the set of meter readers to be assigned, and the preset height difference threshold corresponding to the current meter reader to be assigned is calculated.

[0167] In the mission topographic map, the coordinate point with the smallest terrain height value is determined as the starting point of the simulated flooding process;

[0168] Starting from the starting point, the simulation process is carried out by gradually increasing the water level from the terrain height value at the starting point, and the terrain height value of the first coordinate point that the water level rises above is taken as the terrain height value of the first stage.

[0169] Calculate the height difference between the terrain height value at the starting point and the terrain height value in the first stage, and determine whether the height difference meets the preset height difference threshold corresponding to the current meter reader to be assigned. If so, continue to simulate the water level rise process.

[0170] If not, a watershed is generated at the terrain height value in the first stage, and the meter reading tasks corresponding to all coordinate points that have been submerged during this water level rise are grouped into a task cluster and assigned to the currently assigned meter reading personnel, thus completing one clustering.

[0171] Mark the coordinate points corresponding to the terrain height values ​​of the first stage as the new starting points, and remove the coordinate points corresponding to the assigned tasks from the task terrain map.

[0172] Continue simulating the rising water level until all meter reading tasks corresponding to all coordinate points on the task topographic map have been assigned, then end the simulation process.

[0173] Specifically, the process of determining the "preset height difference threshold corresponding to the currently assigned meter reader" includes:

[0174] Based on the terrain height values ​​of all coordinate points in the task terrain map, the standard deviation of the terrain height values ​​is calculated.

[0175] From the set of meter readers to be assigned, determine the current meter reader to be assigned according to preset rules, and obtain the business capability coefficient of the current meter reader to be assigned.

[0176] Based on the standard height difference threshold, the standard deviation of the terrain height value, and the business capability coefficient of the current meter reader to be assigned, the preset height difference threshold for this division process is obtained through weighted calculation.

[0177] Iterate through the set of meter readers to be assigned and calculate the preset height difference threshold for each meter reader in turn.

[0178] Furthermore, the "obtaining the business capability coefficient of the currently assigned meter reader" includes:

[0179] Obtain the seniority data, average completion efficiency data of historical meter reading tasks, and historical task familiarity data of all meter readers in the set of meter readers to be assigned, as well as the geographical area corresponding to the task topographic map. Calculate the average seniority, average completion efficiency, and average familiarity respectively. The corresponding geographical area is an area with the task cluster center as the center and a preset length as the radius.

[0180] For the currently assigned meter readers, calculate the ratio of seniority data to the average seniority as a seniority factor; calculate the ratio of average completion efficiency data to the average completion efficiency as an efficiency factor; and calculate the ratio of historical task familiarity data to the average familiarity as a familiarity factor.

[0181] The seniority factor, efficiency factor, and familiarity factor are weighted and summed according to preset weights to calculate the business competence coefficient of the currently assigned meter reader.

[0182] Specifically, the path planning module 14 is used for:

[0183] Obtain the geographical location information of all meter reading tasks within the task cluster and predict the operation time;

[0184] The first optimization objective is to minimize the total travel distance, and the second optimization objective is to minimize the total predicted operation time. A multi-objective optimization model is constructed based on the first and second optimization objectives.

[0185] The multi-objective optimization model is solved using a path planning algorithm to generate an initial job sequence covering all meter reading tasks within the task cluster.

[0186] Specifically, the correction and update module 15 is used for:

[0187] During the process of meter readers performing meter reading tasks according to the initial work sequence, the actual working time of each completed meter reading task is recorded in real time.

[0188] Calculate the deviation between the actual operation time and the corresponding predicted operation time, and calculate the average deviation of all completed tasks;

[0189] Based on the average deviation value, the weight parameters used to balance the two optimization objectives of total travel distance and total operation time in the path planning algorithm are dynamically adjusted.

[0190] Based on the revised weight parameters, an update job sequence is generated for all meter reading tasks that have not yet started.

[0191] In summary, the embodiments of this application have at least the following technical effects:

[0192] Compared to existing technologies, this application first acquires a set of meter reading tasks through a data acquisition module, extracting the geographical location and historical status data of the associated metering devices for each task as feature data, providing reliable data support for subsequent work duration prediction and path planning. Second, through a duration prediction module, the work duration of each meter reading task is predicted based on the feature data, achieving accurate quantitative prediction of the work duration for each task and providing reliable data support for subsequent task clustering and multi-objective path planning based on an improved watershed algorithm. Third, through a clustering module, the set of meter reading tasks is spatially clustered based on geographical location and predicted work duration using an improved watershed algorithm, forming several task clusters. These clusters ensure spatial proximity of tasks within the cluster and are adapted to the work capabilities of meter readers. Furthermore, through a path planning module, for each task cluster, with the optimization objective of minimizing the total travel distance and total predicted work duration, an initial path planning algorithm is used to generate an initial work sequence for each task cluster, achieving optimal planning of the execution order of meter reading tasks within the task cluster. Finally, by revising and updating the module, the initial job sequence is executed, the actual job duration is collected, the deviation between the actual and predicted job durations is calculated, and the optimization parameters of the path planning algorithm are dynamically adjusted. For subsequent unexecuted meter reading tasks, path planning is re-planned, generating an updated job sequence. This achieves real-time adaptation of path planning to the actual job scenario, effectively avoiding the limitations of static planning. In this way, commuting redundancy costs are reduced, overall operational efficiency is improved, and real-time adaptation of path planning to the actual field scenario is achieved, avoiding the limitations of static planning. This ultimately meets the comprehensive needs of modern meter reading operations for efficiency improvement, cost control, and job balance.

[0193] It should be noted that the descriptions of each embodiment in the above embodiments have different focuses. For parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0194] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0195] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0196] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0197] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0198] Although preferred embodiments of the invention have been described, those skilled in the art, once they have learned the basic inventive concept, can make other changes and modifications to these embodiments.

[0199] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of this invention and its equivalents, this invention also intends to include these modifications and variations.

Claims

1. A meter reading path planning method based on multi-objective constraints, characterized in that, The method comprises: obtaining a meter reading task set, extracting the geographical position and the historical state data of the associated metering device of each meter reading task as the feature data of each meter reading task; According to the feature data, the working time of each meter reading task is predicted to obtain the predicted working time; Based on the geographical position and the predicted working time, the improved watershed algorithm is used for spatial clustering division of the meter reading task set to form a plurality of task clusters; For each task cluster, the path planning algorithm is used for initial path planning to generate an initial working sequence of each task cluster, taking minimizing the total travel distance and the total predicted working time as the optimization objective; The initial working sequence is executed, the actual working time is collected, the deviation between the actual working time and the predicted working time is calculated, the optimization parameters of the path planning algorithm are dynamically adjusted, and the subsequent unexecuted meter reading tasks are re-planned to generate an updated working sequence; Wherein, based on the geographical position and the predicted working time, the improved watershed algorithm is used for spatial clustering division of the meter reading task set to form a plurality of task clusters, comprising: Coordinate transformation is performed on the geographical position information of each meter reading task, the latitude and longitude coordinates are mapped to coordinate points in a two-dimensional plane coordinate system, and the predicted working time of each meter reading task is converted into a terrain height value corresponding to the coordinate point to generate a task terrain map; A current meter reading personnel to be allocated is determined from a meter reading personnel set to be allocated, and a preset height difference threshold value corresponding to the current meter reading personnel to be allocated is calculated based on a standard height difference threshold value, a terrain height value standard deviation and a business ability coefficient of the current meter reading personnel to be allocated; In the task terrain map, the coordinate point with the minimum terrain height value is determined as the starting point of the simulated flooding process; From the starting point, the process of gradually rising water level from the terrain height value of the starting point is simulated, and the terrain height value of the first coordinate point submerged by the rising water level is taken as the first stage terrain height value; The height difference between the terrain height value of the starting point and the first stage terrain height value is calculated, and it is judged whether the height difference meets the preset height difference threshold value corresponding to the current meter reading personnel to be allocated. If yes, the water level rising process is continued; If not, a watershed is generated at the first stage terrain height value, and the meter reading tasks corresponding to all coordinate points submerged in this water level rising process are grouped into a task cluster and allocated to the current meter reading personnel to be allocated, completing a clustering; The coordinate point corresponding to the first stage terrain height value is marked as a new starting point, and the coordinate points corresponding to the allocated tasks are removed from the task terrain map; The water level rising process is continued until all meter reading tasks corresponding to the coordinate points in the task terrain map are allocated, and the simulation process is ended.

2. The multi-objective constraint based meter reading path planning method according to claim 1, wherein, Obtaining a meter reading task set, extracting the geographical position and the historical state data of the associated metering device of each meter reading task as the feature data of each meter reading task, comprising: Based on the preset task time window rule and the task priority classification rule, the current to-be-executed meter reading task is dynamically screened from the meter reading task pool to form a to-be-meter-reading task set, wherein the task time window rule is used to distinguish between rigid time window tasks and flexible time window tasks; For each meter reading task in the to-be-meter-reading task set, the geographic location information of the meter reading task is parsed and extracted, wherein the geographic location information at least includes longitude and latitude coordinates; For each meter reading task in the to-be-meter-reading task set, the historical state data of the metering device corresponding to the meter reading task is queried and extracted from the electric energy metering device operation database, wherein the historical state data at least includes fault records, operation and maintenance records, and historical meter reading failure records; The geographic location information and the historical state data corresponding to each meter reading task are combined to form the feature data of the meter reading task. 3.The multi-objective constraint based meter reading path planning method of claim 1, wherein, According to the feature data, the working time of each meter reading task is predicted to obtain the predicted working time, including: Calling a pre-trained working time prediction model; The feature data of the meter reading task is input into the working time prediction model; The working time prediction model predicts the working time of the meter reading task based on the feature data by analyzing the historical state data of the electric energy metering device and combining the geographic location information of the meter reading task.

4. The multi-objective constraint based meter reading path planning method of claim 3, wherein, The construction process of the working time prediction model includes: From the meter reading work order database and the electric energy metering device operation database, historical meter reading tasks are obtained, wherein the historical meter reading tasks include the geographic location information of the historical meter reading tasks, the historical state data of the associated metering device, and the corresponding actual working time; After denoising and normalization processing of the collected geographic location information and the historical state data of the associated metering device, the sample feature data set is combined, and the corresponding actual working time is taken as the sample label to form the sample label data set; Based on the machine learning algorithm, the working time prediction model is constructed; The sample feature data set and the sample label data set are used to supervise the training of the working time prediction model until the verification converges, and the trained working time prediction model is obtained. 5.The multi-objective constraint based meter reading path planning method of claim 1, wherein, The determination process of the preset height difference threshold corresponding to the current to-be-assigned meter reading personnel includes: From the to-be-assigned meter reading personnel set, the current to-be-assigned meter reading personnel is determined according to the preset rule, and the business ability coefficient of the current to-be-assigned meter reading personnel is obtained; Based on the standard height difference threshold, the standard deviation of the terrain height value, and the business ability coefficient of the current to-be-assigned meter reading personnel, the preset height difference threshold of the current division process is obtained through weighted calculation; The to-be-assigned meter reading personnel set is traversed, and the preset height difference threshold corresponding to each meter reading personnel is calculated in turn.

6. The multi-objective constraint based meter reading path planning method according to claim 5, wherein, The business ability coefficient of the current to-be-assigned meter reading personnel includes: obtain seniority data of all meter reading personnel in the set of meter reading personnel to be allocated, average completion efficiency data of historical meter reading tasks, and historical task familiarity data in a corresponding geographical area of the task terrain map, and calculate average seniority, average completion efficiency, and average familiarity, respectively, wherein the corresponding geographical area is an area with a task cluster center as the center and a preset length as the radius; for the current meter reading personnel to be allocated, calculate the ratio of the seniority data to the average seniority as a seniority factor, the ratio of the average completion efficiency data to the average completion efficiency as an efficiency factor, and the ratio of the historical task familiarity data to the average familiarity as a familiarity factor; weight and sum the seniority factor, the efficiency factor, and the familiarity factor according to a preset weight to calculate a business capability coefficient of the current meter reading personnel to be allocated. 7.The multi-objective constraint based meter reading path planning method of claim 1, wherein, For each task cluster, an initial path planning algorithm is used to perform initial path planning with the optimization objective of minimizing total travel distance and total predicted operation time to generate an initial operation sequence for each task cluster, including: obtaining geographical position information and predicted operation time of all meter reading tasks in the task cluster; building a multi-objective optimization model according to the first optimization objective and the second optimization objective, with the first optimization objective being to minimize total travel distance and the second optimization objective being to minimize total predicted operation time; solving the multi-objective optimization model using a path planning algorithm to generate an initial operation sequence covering all meter reading tasks in the task cluster. 8.The multi-objective constraint based meter reading path planning method of claim 1, wherein, performing the initial operation sequence, collecting actual operation time, calculating the deviation between the actual operation time and the predicted operation time, dynamically adjusting the optimization parameters of the path planning algorithm, and re-planning the path for subsequent unexecuted meter reading tasks to generate an updated operation sequence, including: recording the actual operation time of each completed meter reading task in real time during the execution of the meter reading tasks according to the initial operation sequence; calculating the deviation between the actual operation time and the corresponding predicted operation time, and calculating the average deviation value of all completed tasks; based on the average deviation value, dynamically correcting the weight parameter in the path planning algorithm for balancing the two optimization objectives of total travel distance and total operation time; based on the corrected weight parameter, generating an updated operation sequence for all meter reading tasks that have not yet started execution.

9. A meter reading route planning system based on multi-objective constraints, characterized by, A meter reading path planning method based on multi-objective constraints is used to perform any one of claims 1-8, including: a data acquisition module for obtaining a set of meter reading tasks, extracting geographical position and historical state data of associated metering devices of each meter reading task as feature data of each meter reading task; a time prediction module for predicting the operation time of each meter reading task based on the feature data to obtain predicted operation time; a clustering and division module for spatially clustering and dividing the set of meter reading tasks to be read based on the geographical position and the predicted operation time by using an improved watershed algorithm to form a plurality of task clusters. The path planning module is configured to perform initial path planning for each task cluster by using a path planning algorithm, so as to minimize total travel distance and total predicted operation time, and to generate an initial operation sequence for each task cluster. The correction and update module is configured to execute the initial operation sequence, collect actual operation time, calculate deviation between the actual operation time and the predicted operation time, dynamically adjust optimization parameters of the path planning algorithm, re-plan paths for subsequent unexecuted meter reading tasks, and generate an updated operation sequence.

Images