Electricity metering work order intelligent dispatching method and system
Patent Information
- Application Number
- CN202611211485.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-11
- Publication Date
- 2026-10-27
AI Technical Summary
[0004]为了克服现有技术的上述缺陷，本发明的实施例提供电力装表工单智能调度方法，以解决现有工单调度中指标权重分配受主观经验偏差影响、异构数据物理量纲不统一干扰评价结果以及局部贪婪派单导致资源错配与整体效益偏低的问题
本发明通过构建包含目标层、准则层和方案层的层次分析体系，并结合判断矩阵的一致性校验机制，能够将工单紧急度、区域距离和人员业务水平等多维度的调度影响指标转化为定量的权重参数，在计算权重的过程中，通过一致性比率的判定和极值元素的修正计算，减少业务经验设定标度时产生的主观偏差，使评价模型更贴合实际业务需求，提高各项指标权重分配的客观性。
Smart Images
Figure CN122988602A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of power system operation and maintenance service technology, and more specifically, to a method and system for intelligent dispatching of power meter installation work orders. Background Technology
[0002] In the daily operation of power systems, work order scheduling for low-voltage meter installation and connection services is a crucial link in ensuring customer power experience and improving operation and maintenance efficiency. When allocating work orders, the scheduling system typically needs to comprehensively consider multiple influencing factors, such as the urgency of the work order, the spatial distance between the work site and maintenance personnel, the professional skills of the personnel, and the workload.
[0003] Existing work order scheduling methods have at least the following technical problems in practical applications: (1) When dealing with multi-dimensional scheduling evaluation indicators, the existing system often relies directly on manual business experience to set the weight parameters of each indicator. It lacks a scientific judgment matrix establishment process and an effective consistency verification mechanism, which can easily lead to logical subjective bias and result in the evaluation model's weight setting being not objective enough. (2) Existing data processing procedures usually lack differentiated processing of cost-type indicators and benefit-type indicators, and have not established a sound dimensionless conversion mechanism, which makes it easy for the absolute size of the values to interfere with the comprehensive evaluation results. The system has difficulty in accurately integrating multi-source heterogeneous data into a single standard evaluation basis. (3) Conventional scheduling systems usually dispatch orders in chronological order or in a local greedy allocation mode based solely on proximity. This satisfies the local optimal conditions for a single assignment, but does not take into account the overall benefit maximization of a single scheduling batch. In actual business scenarios, the number of pending work orders is usually different from the number of available field personnel. Existing local dispatch strategies are difficult to establish an effective global complete matching topology model, which can easily lead to resource mismatches where high-priority work orders are assigned to remote personnel or high-skilled personnel to handle regular work orders, resulting in low overall dispatch efficiency. Summary of the Invention
[0004] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide an intelligent scheduling method for power meter installation work orders, which solves the problems in existing work order scheduling such as the influence of subjective experience bias on the allocation of indicator weights, the interference of inconsistent physical dimensions of heterogeneous data on evaluation results, and the resource misallocation and low overall efficiency caused by local greedy work order assignment.
[0005] To achieve the above objectives, the present invention provides the following technical solution: The intelligent scheduling method for power meter installation work orders includes the following steps: Obtain the set of work orders to be assigned and the set of available maintenance personnel, and extract the original evaluation data for each work order and each available maintenance personnel respectively; Based on the set of work orders to be assigned and the set of available maintenance personnel, a first set of indicators is set up and the final weight vector of each indicator is solved. Based on the set of work orders to be assigned, the set of available maintenance personnel, and the original evaluation data, a set of work order-personnel matching pairs is generated. The final weight vectors of each indicator in the first indicator set are combined to traverse each matching pair and calculate the comprehensive matching score one by one to generate a global matching score matrix. Based on the global matching score matrix, a weighted bipartite graph topology model is constructed and virtual nodes are set for symmetry compensation. The weighted bipartite graph topology model is then solved to obtain the maximum weight matching set. The maximum weight matching set is parsed, invalid matching items involving virtual nodes are removed, and work order allocation instructions are generated and issued.
[0006] In a preferred embodiment, the original evaluation data for each work order includes work order urgency parameters and location coordinate data; the original evaluation data for each available maintenance personnel includes real-time location data, daily carrying capacity margin parameters, business performance score, and historical interaction score.
[0007] In a preferred embodiment, setting the first set of indicators and solving for the final weight vector of each indicator includes: Set a first set of indicators at the criteria level, and construct a positive and negative judgment matrix based on the first set of indicators; The positive and negative judgment matrix is solved and verified to obtain the final weight vector of each indicator in the first indicator set.
[0008] In a preferred embodiment, solving and verifying the positive-reverse judgment matrix includes: Solve the positive-negative judgment matrix and output the initial weight vector of each indicator in the first indicator set; Perform a consistency check on the positive-negative judgment matrix to obtain the consistency ratio. Set a tolerance threshold, compare the consistency ratio with the tolerance threshold, and obtain the final weight vector based on the comparison result.
[0009] In a preferred embodiment, generating the global matching score matrix includes: Based on the original evaluation data, construct the original feature vector of each matching pair in the work order-person matching pair set; The original feature vectors of each matching pair are dimensionless to obtain a normalized score vector. The normalized score vector of each matching pair is multiplied by the final weight vector of each indicator in the first indicator set to obtain the comprehensive matching score of each matching pair. The comprehensive matching scores of each matching pair are stored in memory according to the row and column correspondence to generate a global matching score matrix.
[0010] In a preferred embodiment, the dimensionless processing of the original feature vectors of each matching pair includes: Determine the polarity of each indicator in each original feature vector, and classify them into benefit-type indicators and cost-type indicators; Obtain the maximum and minimum original values among the original evaluation values of the k-th evaluation indicator for each matching pair. If the maximum and minimum original values are the same, then the normalized score of the k-th evaluation indicator is uniformly assigned to a constant 1. If the maximum and minimum original values are not the same, the positive range transformation formula is used to calculate the benefit-type indicators, and the negative range transformation formula is used to calculate the cost-type indicators.
[0011] In a preferred embodiment, the step of constructing a weighted bipartite graph topology model and setting virtual nodes for symmetry compensation includes: Based on the global matching score matrix, a weighted bipartite graph topology model is constructed with the work order projects to be assigned and the available maintenance personnel as vertices; Obtain the total number of left and right vertices of the weighted bipartite graph topology model, compare the total number of vertices on both sides, set virtual nodes according to the comparison results, and obtain the compensated set of vertices on both sides. Construct an extended weight matrix based on the compensated vertex set.
[0012] In a preferred embodiment, solving the weighted bipartite graph topology model includes: Based on the extended weight matrix, the vertex labels of each vertex in the weighted bipartite graph topology model are initialized, and the initial vertex label values of each left and right node in the extended weight matrix are set respectively. Extract the equality subgraph based on the initial top label value, and solve for the maximum weight matching set in the equality subgraph.
[0013] In a preferred embodiment, solving for the maximum weight matching set in the equality subgraph includes: In the currently generated equality subgraph, find an augmenting path for the left node; if an augmenting path is found for the current left node, record the matching pair and add it to the maximum weight matching set; If no augmenting path is found for the current left node, traverse all nodes traversed by the unmatched edge and the matched edge alternately to obtain the search path tree. Calculate the top label adjustment based on the search path tree and update the top labels of the nodes in the search path tree. Re-extract the equal subgraph and continue searching for augmenting paths until all left nodes in the expanded weight matrix have found matching right nodes, thus obtaining the maximum weight matching set.
[0014] This invention provides an intelligent dispatching system for power meter installation work orders, comprising: The data acquisition module is used to obtain the set of work orders to be assigned and the set of available maintenance personnel, and to extract the raw evaluation data of each work order and each available maintenance personnel. The architecture analysis module sets a first set of indicators based on the set of work orders to be assigned and the set of available maintenance personnel, and solves the final weight vector of each indicator. The matrix generation module generates a set of work order-personnel matching pairs based on the set of work orders to be assigned, the set of available maintenance personnel, and the original evaluation data. It then traverses each matching pair by combining the final weight vector of each indicator in the first indicator set to calculate the comprehensive matching score one by one, and generates a global matching score matrix. The matching solution module constructs a weighted bipartite graph topology model based on the global matching score matrix and sets virtual nodes for symmetry compensation processing. It then solves the weighted bipartite graph topology model to obtain the maximum weight matching set. The instruction dispatch module is used to parse the maximum weight matching set, remove invalid matching items involving virtual nodes, and generate and issue work order allocation instructions.
[0015] The technical effects and advantages of the intelligent dispatching method for power meter installation work orders of this invention are as follows: This invention constructs a hierarchical analysis system comprising a target layer, a criterion layer, and a solution layer, and combines it with a consistency verification mechanism for the judgment matrix. This system can transform multi-dimensional scheduling impact indicators such as work order urgency, regional distance, and personnel business skills into quantitative weight parameters. In the process of calculating weights, the system reduces subjective biases caused by setting scales based on business experience through the determination of consistency ratios and the correction calculation of extreme value elements. This makes the evaluation model more in line with actual business needs and improves the objectivity of the weight allocation of various indicators.
[0016] This invention performs range transformation on the original evaluation values of various indicators, distinguishes between benefit-type indicators and cost-type indicators, and performs forward or reverse dimensionless calculations respectively. It unifies the dimensions of multi-source heterogeneous data, eliminates the interference of different physical dimensions and absolute numerical values on the evaluation results, and enables multi-dimensional business data to be directly multiplied with the weight vector. This transforms complex multi-objective scheduling elements into a single comprehensive matching score, providing a standardized data foundation for subsequent global optimization.
[0017] This invention constructs a weighted bipartite graph topology model based on a global matching score matrix. When the number of work orders to be assigned and the number of available maintenance personnel are unequal, virtual nodes are added to symmetrically compensate the topology structure. The maximum weight matching set of the weighted bipartite graph topology model is solved by combining the KM algorithm. The work order assignment process is transformed into a complete matching calculation in mathematical graph theory. Work order assignment instructions are output from the perspective of maximizing overall business benefits. This overcomes the resource mismatch problem caused by conventional sequential assignment or local greedy assignment, and improves the overall assignment efficiency and personnel resource utilization rate. Attached Figure Description
[0018] Figure 1 This is a schematic diagram of the intelligent scheduling method for power meter installation work orders provided in an embodiment of the present invention; Figure 2 This is a block diagram of the intelligent dispatching system for power meter installation work orders provided in an embodiment of the present invention. Detailed Implementation
[0019] 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 of ordinary skill in the art without creative effort are within the scope of protection of the present invention.
[0020] Example 1, Figure 1 The present invention provides an intelligent scheduling method for power meter installation work orders, comprising the following steps: S1: Obtain the set of work orders to be assigned and the set of available maintenance personnel, and extract the original evaluation data for each work order and each available maintenance personnel.
[0021] S11, Obtain the set of work orders to be assigned, and extract the original evaluation data of each work order item to be assigned; The system reads the set of work orders to be assigned from the business middleware server, including all low-voltage meter installation and connection work orders to be assigned. Specifically:
[0022] In the formula, This is a set of work orders to be assigned. The total number of work orders to be assigned; Let i be the i-th work order item to be assigned, and .
[0023] The original evaluation data for each work order item in the set of work orders to be assigned is extracted. This original evaluation data includes the work order urgency parameter and location coordinate data, where the location coordinate data includes the geographical longitude and latitude values of the work order's on-site operation location. As for how to use structured query language to extract specified field information from a relational database, those skilled in the art can write corresponding query scripts based on actual business logic to accomplish this. The backend data interaction and retrieval mechanism is a well-known technology in this field and will not be elaborated upon here.
[0024] Furthermore, the specific steps for extracting the urgency parameter of a work order are as follows: Obtain the work order type identifier, customer power outage duration record, and customer electricity consumption level label for each work order project to be assigned, and map them to the corresponding basic score; The emergency score for the i-th work order item to be assigned is obtained by weighted summation of the basic scores. Obtain the system's preset emergency level comparison table, compare the emergency score of the i-th work order item with the emergency level comparison table to obtain the work order emergency parameter. The work order emergency parameter is used to characterize the priority order of each meter installation and power connection task.
[0025] S12, Obtain the set of available maintenance personnel and extract the original evaluation data for each available maintenance personnel; The status query command is sent to each field mobile terminal through the communication network. The status flag information of each field mobile terminal indicates the working status of the corresponding maintenance personnel. Obtain the status flag information of each field mobile terminal, remove terminal nodes that are on leave, offline, in sleep mode, or in their scheduled rest period, and include personnel whose status flag shows online and who have order-accepting permission into the available scheduling pool to form a set of available maintenance personnel, specifically:
[0026] In the formula, A pool of available maintenance personnel; This represents the total number of maintenance personnel currently available for scheduling. For the j-th available maintenance personnel, and .
[0027] Based on the set of available maintenance personnel, the raw evaluation data for each available maintenance personnel is calculated, including real-time location data, daily capacity margin parameters, business performance score, and historical interaction score. This objectively records the current operational status of field personnel from four dimensions: spatial distribution, skill matching, workload, and service quality. Real-time location data is collected in real time by the satellite positioning and communication components built into the field mobile terminal and reported to the business middleware server, including the first... The latitude and longitude coordinates of available maintenance personnel at the current timestamp; The daily carrying capacity margin parameter reflects the first The remaining work order acceptance capacity during the shift of available maintenance personnel is calculated by subtracting the number of work orders from the preset daily quota dispatch threshold in the system. The number of work orders that available maintenance personnel have signed and are currently executing in the system on a given day is obtained; where the preset daily quota for work order dispatch is based on the first... The calculation is set by dividing the total daily statutory working hours of the area where the available maintenance personnel are located by the standard working hours of a single meter installation and connection work order. The skills rating is used to characterize the individual's ability to handle complex electrical connection faults, and to obtain the first... The system uses the professional skill assessment level code and the number of years of service on the network for each available maintenance personnel. Based on a pre-set skill scoring table, the professional skill assessment level code is converted into a basic skill score. The number of years of service on the network is multiplied by a pre-set experience coefficient to obtain an additional experience score. Finally, the basic skill score and the additional experience score are added together to obtain the score for the [number missing]. A score of the professional competence of available maintenance personnel.
[0028] Historical interaction scores are calculated by the first The average customer satisfaction score is obtained from all work orders completed by available maintenance personnel within a set time period in the past. The time period can be set according to business assessment requirements, such as the past thirty days or the past six months. The customer satisfaction score is extracted from the customer service feedback record table stored in the system.
[0029] S2, based on the set of work orders to be assigned and the set of available maintenance personnel, set a first set of indicators and solve the final weight vector of each indicator.
[0030] S21. Establish a hierarchical analysis system of scheme layer, criterion layer and target layer. Set the first set of indicators in the criterion layer and construct a positive and negative judgment matrix based on the first set of indicators. The set of work orders to be assigned and the set of available maintenance personnel are mapped to a hierarchical analysis system consisting of a target layer, a criterion layer, and a solution layer. In the hierarchical analysis system, the target layer is used to calculate the comprehensive matching score of the combination of work orders to be assigned and field maintenance personnel; the criterion layer sets the first set of indicators that affect the work order dispatch decision; and the solution layer is used to generate all possible matching pairs consisting of the set of work orders to be dispatched and the set of available maintenance personnel.
[0031] Furthermore, the first set of indicators specifically includes:
[0032] in For the first set of indicators, Assess the urgency of the work order; The regional distance reflects the straight-line spatial distance between the current real-time location of the maintenance personnel and the work site of the work order. It is calculated based on the location coordinate data of the i-th work order project to be assigned and the real-time location data of the j-th available maintenance personnel. For the professional skills of personnel; This represents the daily carrying capacity margin. The score is based on historical interaction evaluations.
[0033] Obtain the system's preset criterion importance scoring table, and construct a positive and negative judgment matrix based on the criterion importance scoring table using the analytic hierarchy process. The criterion importance scoring table is set by the system based on historical scheduling data statistics and business expert experience, reflecting the relative importance between each pair of indicators in the first indicator set in the criterion layer. The specific positive and negative judgment matrix is as follows:
[0034] In the formula, This is a matrix for determining positive and negative reciprocities. Let u be the importance scale value of the u-th indicator relative to the v-th indicator. and All values are within the range of Positive integers within; The importance scale value of the u-th indicator relative to the v-th indicator satisfies the following relationship:
[0035] when hour, The larger the importance scale value, the higher the weight of the u-th indicator relative to the v-th indicator in the order dispatch decision.
[0036] S22, Solve and verify the positive and negative judgment matrix to obtain the final weight vector of each indicator in the first indicator set; S221, Solve for the positive and negative judgment matrix and output the initial weight vector of each indicator in the first indicator set; Find the largest eigenvalue and eigenvector of the positive-reciprocal judgment matrix, such that the positive-reciprocal judgment matrix, the largest eigenvalue, and their corresponding eigenvector satisfy the following relationship:
[0037] In the formula, The largest eigenvalue of the positive-negative judgment matrix A; The eigenvector corresponding to the largest eigenvalue; The eigenvector corresponding to the largest eigenvalue is normalized so that the sum of all elements in the vector is 1. The initial weight vector of each indicator in the first indicator set is then output as follows:
[0038] In the formula, This is the initial weight vector; Let be the weight value corresponding to the k-th indicator, where k is a positive integer in the range [1, 5]. For the specific numerical calculation and solution algorithm of matrix eigenvalues and eigenvectors, those skilled in the art can use the power method or the sum-product method to write a program to implement it. The underlying linear algebra calculation process is a well-known technology in this field and will not be described in detail here.
[0039] S222, perform a consistency check on the positive and negative judgment matrix to obtain the consistency ratio, set a tolerance threshold, compare the consistency ratio with the tolerance threshold, and obtain the final weight vector based on the comparison result; Calculate the consistency index based on the largest eigenvalue:
[0040] In the formula, As a consistency indicator; The order of the positive and negative judgment matrix is given. In this embodiment, since the number of criteria layer indicators is 5, therefore... The value is 5.
[0041] Calculate the consistency ratio using consistency metrics:
[0042] In the formula, Consistency ratio; The average random consistency index, The value is obtained by consulting the standard random consistency index table when determining the matrix order. When it is 5, The constant value is 1.12.
[0043] The tolerance threshold is set to 0.1. The consistency ratio is compared with the tolerance threshold, and the final weight vector is obtained based on the comparison result, as follows: like If the positive and negative judgment matrices are consistent, the current initial weight vector is confirmed as the final weight vector and output. like If the positive-negative judgment matrix fails the consistency check, calculate the consistency contribution of each row element in the positive-negative judgment matrix, locate the row with the largest contribution, and find the extreme value element with the largest deviation from the theoretical ratio in that row; obtain the preset correction rule, and according to the preset correction rule, move the extreme value element closer to the value 1 and update the corresponding symmetric element of the extreme value element synchronously; regenerate the positive-negative judgment matrix and return to execute the eigenvalue solving and verification steps until the consistency ratio meets the requirements; update the initial weight vector and output it as the final weight vector. Specifically, the correction rule is as follows: determine the relationship between the extreme value element and the constant 1; if the extreme value element is greater than 1, subtract the preset step size; if the extreme value element is less than 1, add the preset step size; regenerate the positive and negative judgment matrix and return to execute the eigenvalue solving and verification steps.
[0044] This step constructs a hierarchical analysis system comprising a target layer, a criterion layer, and a solution layer. Combined with a consistency verification mechanism for the judgment matrix, it can transform multi-dimensional scheduling impact indicators such as work order urgency, regional distance, and personnel business skills into quantitative weight parameters. In the process of calculating weights, the determination of consistency ratios and the correction calculation of extreme value elements reduce subjective biases generated when setting scales based on business experience, making the system more aligned with actual business needs and improving the objectivity of weight allocation for various indicators.
[0045] S3. Based on the set of work orders to be assigned, the set of available maintenance personnel, and the original evaluation data, generate a set of work order-personnel matching pairs. Combine the final weight vectors of each indicator in the first indicator set, traverse each matching pair one by one to calculate the comprehensive matching score and generate a global matching score matrix.
[0046] S31, Generate a set of work order-person matching pairs and construct the original feature vector of each matching pair; A set of work order-person matching pairs is generated based on the set of work orders to be assigned and the set of available maintenance personnel. This set contains all possible work order-person matching pairs. The matching pair formed by the i-th work order to be assigned and the j-th available maintenance personnel is denoted as […]. ; Obtain the original evaluation data of the i-th work order to be assigned and the j-th available maintenance personnel, and construct a matching pair. The original feature vector is as follows:
[0047] In the formula, For matching pairs The original feature vector; For matching pairs The original evaluation value of the k-th evaluation indicator, where k is the range of values within... Positive integers within; This serves as the base score for the urgency of the work order; The spatial physical distance between the work order site and the current location of the maintenance personnel can be calculated by extracting the latitude and longitude coordinates of both and using the semi-sine formula to write a distance measurement algorithm. The principle of spherical distance calculation is a well-known technology in this field and will not be elaborated here. Quantify the professional skills of operations and maintenance personnel into scores; This represents the daily carrying capacity margin for maintenance personnel. This is a score for the historical interaction evaluations of operations and maintenance personnel.
[0048] S32, the original feature vectors of each matching pair are dimensionless to obtain the normalized score vector; Because the physical meanings of the original feature vectors of different matching pairs are different, their dimensions (such as kilometers and fractions) cannot be directly added or subtracted, and their numerical values contribute differently to the quality of the matching. The range transformation method is used to unify the dimensions of multi-source heterogeneous data, eliminate evaluation bias caused by the absolute magnitude of the indicator values, and map the values to a fixed... Within the range; The polarity of each original feature vector is determined based on the direction of its numerical contribution. If the value of a certain original feature vector is larger and the matching benefit is better, it is set as a benefit-type indicator. If the value is smaller and the matching benefit is better, it is set as a cost-type indicator. In this embodiment, the urgency of the work order, the professional level of the personnel, the daily carrying capacity margin, and the historical interaction evaluation score are set as benefit-type indicators; the spatial physical distance is set as a cost-type indicator. Obtain the maximum and minimum original values of the k-th evaluation index for each matching pair. If the maximum and minimum original values are the same, it indicates that all matching pairs perform completely identically on this index. Directly assign the normalized score of this index to a constant 1 to avoid triggering a division-by-zero calculation error. If the maximum and minimum original values are not the same, then the benefit-type indicators and cost-type indicators are calculated separately. The benefit-type indicators are calculated using the positive range transformation formula, specifically:
[0049] The cost index is calculated using the inverse range transformation formula, specifically:
[0050] In the formula, Original evaluation values The normalized score obtained after dimensionless processing; and The first The maximum and minimum raw values of the indicator among all current matching pairs.
[0051] Match pairs The original feature vector The result is transformed into a dimensionless normalized score vector, specifically:
[0052] In the formula, For matching pairs The normalized score vector, Original evaluation values The normalized score obtained after dimensionless processing.
[0053] S33, perform an inner product operation between the normalized score vector of each matching pair and the final weight vector of each indicator in the first indicator set to obtain the comprehensive matching score of each matching pair, and summarize to generate a global matching score matrix. The matching pair is obtained by multiplying each element of the normalized score vector by its corresponding final weight vector and then summing the results. Overall matching score:
[0054] in, The comprehensive matching score is used to assign the i-th work order to the j-th available maintenance personnel. The higher the score, the better the comprehensive business benefits of the dispatching scheme. This transforms the complex multi-objective evaluation problem into a single scalar optimization problem.
[0055] Iterate through all work order-person matching pairs in the work order-person matching pair set, and for the i-th work order item to be assigned, start from... to The j-th available maintenance personnel from to Perform nested loops to traverse and calculate the comprehensive matching score for all work order-person matching pairs. Then, combine all the calculated scores. Stored in memory according to the row-to-column correspondence, generating a scale of Global matching score matrix:
[0056] Where M is the global matching score matrix, The comprehensive matching score for assigning the i-th work order to the j-th available maintenance personnel.
[0057] This step involves performing range transformation on the original evaluation values of various indicators, distinguishing between benefit-type and cost-type indicators, and performing forward or reverse dimensionless calculations respectively. This unifies the dimensions of multi-source heterogeneous data, eliminates the interference of different physical dimensions and absolute numerical values on the evaluation results, and enables multi-dimensional business data to be directly multiplied with the weight vector, thereby converting complex multi-objective scheduling elements into a single comprehensive matching score.
[0058] S4. Based on the global matching score matrix, a weighted bipartite graph topology model is constructed and virtual nodes are set for symmetry compensation. The weighted bipartite graph topology model is solved to obtain the maximum weight matching set.
[0059] S41. Based on the global matching score matrix, a weighted bipartite graph topology model is constructed with the work order projects to be assigned and the available maintenance personnel as vertices. The asymmetric vertex set is then symmetrically compensated to obtain an extended weight matrix. The weighted bipartite graph topology model transforms the many-to-many assignment relationship between work order items and available maintenance personnel into a mathematical graph topology structure. Specifically, the set of left vertices in the weighted bipartite graph topology model is defined as the set of work order items to be assigned, and the total number of left vertices is... ; Define the set of right-hand vertices in the weighted bipartite graph topology as the set of available maintenance personnel, and the total number of right-hand vertices is . There is a connection edge between any work order vertex on the left and any available maintenance personnel vertex on the right. The weight of this connection edge is assigned to the corresponding comprehensive matching score value in the global matching score matrix.
[0060] Since the number of pending work orders often differs from the number of on-duty personnel in actual field operations, it is necessary to perform symmetry compensation processing on the asymmetric vertex set. This involves obtaining the total number of left and right vertices of the weighted bipartite graph topology model, comparing the total number of vertices on both sides, and setting virtual nodes based on the comparison results. Specifically: when When this occurs, it indicates that the number of work orders to be assigned exceeds the number of available maintenance personnel. Therefore, an additional vertex is added to the right-hand vertex set of the weighted bipartite graph topology model. One virtual personnel node; when When this indicates that the number of available maintenance personnel exceeds the number of work orders to be assigned, an additional vertex is added to the set of the left vertices of the weighted bipartite graph topology model. One virtual work order node; After symmetry compensation, the compensated set of vertices on both sides is obtained. The total number of vertices in both sides of the compensated set is uniformly N. .
[0061] Scale is constructed based on the compensated set of two vertices. To expand the weighted matrix, for the i-th work order item to be assigned to the j-th available maintenance personnel, the weight value is... Keep the original comprehensive matching score value. For any edge associated with a virtual node (virtual work order node or virtual personnel node), assign its weight value to 0. That is, the virtual node does not generate any actual business benefits. Prioritize matching the entity node with high weight. The extended weight matrix is as follows:
[0062] in, To expand the weight matrix, Let be the weight value of the edge connecting the i-th left vertex and the j-th right vertex.
[0063] S42, Initialize the vertex labels of each vertex in the weighted bipartite graph topology model based on the extended weight matrix; Set the top index array of the work orders to be assigned on the left as follows: The right side can be used by maintenance personnel to set the top label array as follows: ; The initial top value is set based on the initial expected state of the business assignment, that is, each work order item to be assigned is expected to be assigned to the maintenance personnel with the highest comprehensive matching score, specifically: For the left-hand nodes in the expanded weight matrix, set the initial vertex index of each left-hand node to the maximum weight value among all edges connected to it:
[0064] In the formula, Let i be the initial index value of the i-th node in the index array of the work order items to be assigned; Let be the weight of the edge connecting the i-th left vertex and the j-th right vertex; The total number of vertices. and ; For the right-hand nodes in the expanded weight matrix, set the initial top-value of each right-hand node to 0:
[0065] In the formula, This is the initial top label value of the j-th node in the top label array of the work order items to be assigned.
[0066] S43, extract the equality subgraph based on the initial top label value, and solve the maximum weight matching set in the equality subgraph; Traverse all vertices and connecting edges in the weighted bipartite graph topology model, and extract equal subgraphs based on the initial vertex labels and label constraint equations of each vertex. The label constraint equations are as follows:
[0067] In the formula, Let be the initial index value of the i-th node in the index array of the work order items to be assigned. Let j be the initial index value of the j-th node in the index array of the work order items to be assigned; Let be the weight value of the edge connecting the i-th left vertex and the j-th right vertex.
[0068] Based on the extracted equality subgraph, the Hungarian algorithm is used to find augmenting paths for the left-hand nodes. The maximum weight matching set is then determined based on the search results. Specifically: If an augmenting path is found in the current left node, record this matching pair and update it to the maximum weight matching set; If a left-hand node cannot find an augmenting path in the current equality subgraph, it indicates that conflict-free resource allocation cannot be completed under the current highest expectation. Alternately traverse all nodes traversed by unmatched and matched edges to obtain a search path tree. Let S be the set of left-hand nodes involved in the search path tree, and T be the set of right-hand nodes involved. Obtain the access status identifier of each node. Based on the access status identifier, determine the nodes contained in the left-hand and right-hand node sets. Calculate the vertex adjustment amount based on the search path tree to expand the edge set of the equality subgraph, specifically:
[0069] In the formula, The top-level adjustment reflects the minimum expected threshold that needs to be lowered to introduce a new matching edge; S and T are the left and right node sets, respectively.
[0070] Update the node top labels in the search path tree based on the top label adjustment amount. For each left node in the left node set, subtract the top label adjustment amount from its top label value. For each right node in the right node set, subtract the top label adjustment amount from its top label value. After updating the top labels, re-extract the equality subgraph and continue searching for augmenting paths; Continue executing the above iterative process of finding augmenting paths, calculating top-level adjustments, and updating top-level values until all left-side nodes in the expanded weight matrix have found matching right-side nodes.
[0071] This step, through the top-index adjustment process, relaxes the matching conditions by calculating the minimum expected gap when a high-expectation matching scheme encounters a competitive conflict. This allows suboptimal connections to be added to the equal subgraph, thereby breaking the deadlock and ultimately forming the maximum weight matching set. This set ensures that the sum of the comprehensive matching scores of all assigned pairs mathematically reaches the theoretical maximum.
[0072] S5. Parse the maximum weight matching set, remove invalid matching items involving virtual nodes, and generate and issue work order allocation instructions.
[0073] S51, remove invalid matching items involving virtual nodes from the maximum weight matching set to obtain the effective allocation set; Since virtual work orders and virtual personnel nodes are introduced in step S4 to meet the matrix dimension requirements of the KM algorithm, the allocation relationship involving virtual nodes needs to be removed when parsing the maximum weight matching set in order to extract the actual work order allocation instructions.
[0074] Let the set of maximum weight matchings be denoted as This includes N pairs of matching relationships; Traversing the maximum weight matching set For each matching pair in the table, the matching pair consisting of the i-th left-side work order node and the j-th left-side personnel node is denoted as... Obtain the total number of actual work order nodes to be assigned and the total number of available maintenance personnel nodes, and extract valid matching pairs that meet the entity constraints to form a valid assignment set:
[0075] In the formula, The effective allocation set; m is the total number of actual work orders to be allocated; This represents the total number of currently available maintenance personnel. when or If the match is invalid, it indicates that the matching relationship has assigned the work order to a virtual person or the available maintenance personnel to a virtual work order. Such matching does not generate actual business actions and is therefore invalid. The system will discard the matching relationship directly.
[0076] Calculate the actual total matching benefit for this scheduling task:
[0077] In the formula, This is the sum of the overall matching scores for all actually effective matching relationships; This represents the comprehensive matching score extracted from the global matching score matrix, corresponding to the i-th work order item to be assigned and the j-th available maintenance personnel. This value can be saved to the system log for subsequent evaluation of the overall assignment quality of the current scheduling batch.
[0078] S52 generates and issues work order allocation instructions based on the valid allocation set; Based on the valid allocation set, a business execution message is generated for each valid matching pair. The specific business fields of the corresponding work order are extracted from the associated database. The work order number, customer repair description, customer contact information, on-site geographical coordinates and the time limit for completion are encapsulated into a data packet in JSON (JavaScript Object Notation) format to obtain the encapsulated business execution message. Establish a long-term communication connection with the corresponding maintenance personnel's mobile terminal in the field (e.g., a two-way communication channel based on the WebSocket or MQTT protocol), and send the encapsulated business execution message to the designated mobile terminal in the field. The business middleware server synchronously updates the resource status table in the database based on the distribution results. The system changes the status field of successfully allocated work orders from pending allocation to executing and records the dispatch timestamp. The system updates the multi-dimensional business attribute data of the corresponding maintenance personnel according to preset rules. Specifically, it parses the business type code of the work order and performs a matching search in the system's preset work order workload conversion table to obtain the standard consumption hours or piece-rate score corresponding to this type of work order as the workload value. Then, it deducts the workload value from the personnel's daily carrying capacity margin to ensure the accuracy of personnel load data in the next round of scheduling decisions.
[0079] The field mobile terminal receives a service execution message, triggering prompts on the local operation interface. The terminal then invokes its internal geographic information system (GIS) component, using the real-time coordinates obtained from current satellite positioning and the work order's location coordinates from the message, to initiate a route planning request to the map service engine. Regarding how to utilize network interfaces to call third-party map services for route planning and navigation guidance, those skilled in the art can implement this by writing code based on the documentation of the relevant development packages. The underlying path search algorithm and map rendering mechanism are well-known technologies in the field and will not be elaborated upon here.
[0080] After maintenance personnel arrive on-site and complete the meter installation and power connection work, the field mobile terminal collects on-site photos and completion confirmation information and uploads them to the business middleware server. Upon receiving the completion data, the server triggers the business settlement process and sends a service evaluation request link to the customer via the associated SMS gateway. The customer's satisfaction score is written into the customer service feedback record table, which is used to dynamically update the maintenance personnel's historical interaction evaluation scores in subsequent dispatches, thereby establishing a closed-loop system operation mechanism for dispatching, execution, and evaluation.
[0081] Example 2, Figure 2A smart scheduling system for power meter installation work orders is presented, deployed on a business middleware server, and establishes bidirectional communication connections with multiple field mobile terminals. The system includes a data acquisition module, an architecture analysis module, a matrix generation module, a matching solution module, and an instruction dispatch module, specifically: The data acquisition module is used to obtain the set of work orders to be assigned and the set of available maintenance personnel, and to extract the raw evaluation data of each work order and each available maintenance personnel. The architecture analysis module sets a first set of indicators based on the set of work orders to be assigned and the set of available maintenance personnel, and solves the final weight vector of each indicator. The matrix generation module generates a set of work order-personnel matching pairs based on the set of work orders to be assigned, the set of available maintenance personnel, and the original evaluation data. It then traverses each matching pair by combining the final weight vector of each indicator in the first indicator set to calculate the comprehensive matching score one by one, and generates a global matching score matrix. The matching solution module constructs a weighted bipartite graph topology model based on the global matching score matrix and sets virtual nodes for symmetry compensation processing. It then solves the weighted bipartite graph topology model to obtain the maximum weight matching set. The instruction dispatch module is used to parse the maximum weight matching set, remove invalid matching items involving virtual nodes, and generate and issue work order allocation instructions.
[0082] Furthermore, the business middleware server provides computing power support and data storage services; the field mobile terminal is configured with each maintenance personnel and is used to receive dispatch instructions and report real-time location and status data.
[0083] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.
[0084] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.
[0085] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.
[0086] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.
[0087] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.
[0088] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.
Claims
1. A method for intelligent scheduling of power meter installation work orders, characterized in that, Includes the following steps: Obtain the set of work orders to be assigned and the set of available maintenance personnel, and extract the original evaluation data for each work order and each available maintenance personnel respectively; Based on the set of work orders to be assigned and the set of available maintenance personnel, a first set of indicators is set up and the final weight vector of each indicator is solved. Based on the set of work orders to be assigned, the set of available maintenance personnel, and the original evaluation data, a set of work order-personnel matching pairs is generated. The final weight vectors of each indicator in the first indicator set are combined to traverse each matching pair and calculate the comprehensive matching score one by one to generate a global matching score matrix. Based on the global matching score matrix, a weighted bipartite graph topology model is constructed and virtual nodes are set for symmetry compensation. The weighted bipartite graph topology model is then solved to obtain the maximum weight matching set. The maximum weight matching set is parsed, invalid matching items involving virtual nodes are removed, and work order allocation instructions are generated and issued.
2. The intelligent scheduling method for power meter installation work orders according to claim 1, characterized in that, The original evaluation data for each work order includes work order urgency parameters and location coordinate data; the original evaluation data for each available maintenance personnel includes real-time location data, daily carrying capacity margin parameters, business performance score, and historical interaction score.
3. The intelligent scheduling method for power meter installation work orders according to claim 1, characterized in that, The process of setting up the first set of indicators and solving for the final weight vector of each indicator includes: Set a first set of indicators at the criteria level, and construct a positive and negative judgment matrix based on the first set of indicators; The positive and negative judgment matrix is solved and verified to obtain the final weight vector of each indicator in the first indicator set.
4. The intelligent scheduling method for power meter installation work orders according to claim 3, characterized in that, The process of solving and verifying the positive-reverse judgment matrix includes: Solve the positive-negative judgment matrix and output the initial weight vector of each indicator in the first indicator set; Perform a consistency check on the positive-negative judgment matrix to obtain the consistency ratio. Set a tolerance threshold, compare the consistency ratio with the tolerance threshold, and obtain the final weight vector based on the comparison result.
5. The intelligent scheduling method for power meter installation work orders according to claim 4, characterized in that, The generation of the global matching score matrix includes: Based on the original evaluation data, construct the original feature vector of each matching pair in the work order-person matching pair set; The original feature vectors of each matching pair are dimensionless to obtain a normalized score vector. The normalized score vector of each matching pair is multiplied by the final weight vector of each indicator in the first indicator set to obtain the comprehensive matching score of each matching pair. The comprehensive matching scores of each matching pair are stored in memory according to the row and column correspondence to generate a global matching score matrix.
6. The intelligent scheduling method for power meter installation work orders according to claim 5, characterized in that, The dimensionless processing of the original feature vectors of each matching pair includes: Determine the polarity of each indicator in each original feature vector, and classify them into benefit-type indicators and cost-type indicators; Obtain the maximum and minimum original values among the original evaluation values of the k-th evaluation indicator for each matching pair. If the maximum and minimum original values are the same, then the normalized score of the k-th evaluation indicator is uniformly assigned to a constant 1. If the maximum and minimum original values are not the same, the positive range transformation formula is used to calculate the benefit-type indicators, and the negative range transformation formula is used to calculate the cost-type indicators.
7. The intelligent scheduling method for power meter installation work orders according to claim 6, characterized in that, The construction of the weighted bipartite graph topology model and the setting of virtual nodes for symmetry compensation processing include: Based on the global matching score matrix, a weighted bipartite graph topology model is constructed with the work order projects to be assigned and the available maintenance personnel as vertices; Obtain the total number of left and right vertices of the weighted bipartite graph topology model, compare the total number of vertices on both sides, set virtual nodes according to the comparison results, and obtain the compensated set of vertices on both sides. Construct an extended weight matrix based on the compensated vertex set.
8. The intelligent scheduling method for power meter installation work orders according to claim 7, characterized in that, Solving the weighted bipartite graph topology model includes: Based on the extended weight matrix, the vertex labels of each vertex in the weighted bipartite graph topology model are initialized, and the initial vertex label values of each left and right node in the extended weight matrix are set respectively. Extract the equality subgraph based on the initial top label value, and solve for the maximum weight matching set in the equality subgraph.
9. The intelligent dispatching method for power meter installation work orders according to claim 8, characterized in that, The step of solving for the maximum weight matching set in the equality subgraph includes: In the currently generated equality subgraph, find an augmenting path for the left node; if an augmenting path is found for the current left node, record the matching pair and add it to the maximum weight matching set; If no augmenting path is found for the current left node, traverse all nodes traversed by the unmatched edge and the matched edge alternately to obtain the search path tree. Calculate the top label adjustment based on the search path tree and update the top labels of the nodes in the search path tree. Re-extract the equal subgraph and continue searching for augmenting paths until all left nodes in the expanded weight matrix have found matching right nodes, thus obtaining the maximum weight matching set.
10. A system using the intelligent scheduling method for power meter installation work orders as described in any one of claims 1-9, characterized in that, include: The data acquisition module is used to obtain the set of work orders to be assigned and the set of available maintenance personnel, and to extract the raw evaluation data of each work order and each available maintenance personnel. The architecture analysis module sets a first set of indicators based on the set of work orders to be assigned and the set of available maintenance personnel, and solves the final weight vector of each indicator. The matrix generation module generates a set of work order-personnel matching pairs based on the set of work orders to be assigned, the set of available maintenance personnel, and the original evaluation data. It then traverses each matching pair by combining the final weight vector of each indicator in the first indicator set to calculate the comprehensive matching score one by one, and generates a global matching score matrix. The matching solution module constructs a weighted bipartite graph topology model based on the global matching score matrix and sets virtual nodes for symmetry compensation processing. It then solves the weighted bipartite graph topology model to obtain the maximum weight matching set. The instruction dispatch module is used to parse the maximum weight matching set, remove invalid matching items involving virtual nodes, and generate and issue work order allocation instructions.