Power public opinion data-driven work deployment adaptive optimization method and system

Abstract

The invention relates to an electric power public opinion data driven work deployment adaptive optimization method and system, and the method comprises the following steps: S1, obtaining multi-source data related to electric power public opinions, filtering noise data, forming a standardized data set, and constructing a unified public opinion data pool; s2, constructing a public opinion knowledge graph, and obtaining a comprehensive public opinion risk index; s3, constructing a multi-target service deployment adaptive optimization model; s4, according to the multi-target service deployment adaptive optimization model, obtaining an optimal work deployment scheme based on a scheme solving and generating method of an adaptive optimization algorithm; and S5, according to the optimal work deployment scheme, pushing the optimal work deployment scheme to a decision maker through a visual billboard, tracking the whole execution process of the scheme, and feeding back an execution result and newly added public opinion data to the step S1. Through real-time public opinion analysis, high-risk events are automatically identified before event expansion, active intervention is realized, and a traditional passive complaint waiting mode is changed.

Description

Technical Field

[0001] This invention relates to job deployment optimization, and more particularly to a job deployment adaptive optimization method and system driven by power public opinion data. Background Technology

[0002] Public opinion in the power industry is highly sensitive and has a wide-ranging impact. Topics such as policy adjustments and power outages can easily trigger public outcry. With the continuous improvement of social informatization, the public increasingly expresses their opinions, demands, and emotions regarding power services through news media, social media platforms, online forums, customer service hotlines, and other channels, resulting in massive amounts of power-related public opinion data. This data directly reflects public perceptions and potential risks regarding power supply, service quality, grid security, and electricity pricing policies.

[0003] Currently, the work deployment of power companies' business departments mainly relies on the following methods: planned work: tasks are arranged according to the annual, quarterly or monthly routine plans; event-driven: work orders are passively dispatched for processing after a fault occurs or a user complaint is received; experience-based decision-making: work priorities and resource allocation are determined based on the personal experience of managers.

[0004] However, the existing working methods have the following prominent problems: Disconnect between public opinion monitoring and business deployment: Most public opinion systems only provide early warning functions and fail to form effective linkages with business departments such as operations and maintenance and customer service; Insufficient adaptability of resource scheduling: Traditional fixed processes are unable to cope with changes in the urgency of public opinion, failing to identify potential mass incidents or service risks in advance, and also unable to dynamically adjust resources according to the public opinion situation during the implementation of the plan, resulting in delayed response and missed opportunities for handling; One-sided analytical perspective: Lack of integrated analysis of multi-source public opinion data makes it difficult to comprehensively and quantitatively assess the overall impact of events; Low execution efficiency: Resource allocation is not entirely reasonable, with some departments overloaded while other departments have idle resources, resulting in low overall operational efficiency. Summary of the Invention

[0005] To address the aforementioned issues, the present invention aims to provide a data-driven adaptive optimization method and system for power public opinion deployment. Through real-time public opinion analysis, it automatically identifies high-risk events before the situation escalates, enabling proactive intervention and changing the traditional passive waiting mode for complaints.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: An adaptive optimization method for job deployment driven by power public opinion data includes the following steps: S1: Acquire multi-source data related to public opinion on electricity, filter out noisy data based on an integrated model, form a standardized dataset, and build a unified public opinion data pool; S2: Based on the public opinion data pool, construct a public opinion knowledge graph to obtain a comprehensive public opinion risk index; S3: By combining the staffing, response time, and processing historical data of each department, construct an adaptive optimization model for multi-objective business deployment; S4: Based on the multi-objective business deployment adaptive optimization model, and the solution solving and generation method based on the adaptive optimization algorithm, obtain the optimal work deployment scheme; S5: Based on the optimal work deployment plan, push it to the decision-maker through a visual dashboard, track the entire process of plan execution, and feed back the execution results and new public opinion data to step S1.

[0007] Furthermore, the construction of the public opinion knowledge graph is as follows: Define the knowledge framework of the graph, clarifying entity types, including events, people / organizations, geographical locations, public opinion objects, sentiment tendencies, and solutions, as well as their interrelationships; based on the defined knowledge framework, utilize natural language processing technology to extract specific entities, attributes, and relationship instances from public opinion data, and eliminate ambiguity through entity linking and other technologies, merging heterogeneous data into unified knowledge units; store the merged knowledge in a graph database to generate a visual graph, intuitively presenting the relationships, geographical distribution, and evolution paths of public opinion events; utilize the constructed knowledge graph to support in-depth business applications, including performing complex relational queries, analyzing the diffusion path of public opinion impact, and quickly locating the root cause of events through semantic reasoning, as well as recommending responsible business departments and emergency resources, providing direct and accurate decision-making basis for adaptive optimization models.

[0008] Furthermore, the comprehensive public opinion risk index R:

[0009] Where R is the comprehensive public opinion risk index; S is the sentiment index; H is the popularity value; V is the speed of dissemination; A is the scope of influence; B is the business relevance; Ws, Wh, Wv, Wa, and Wb are the weight coefficients corresponding to the above five dimensions, and satisfy Ws+Wh+Wv+Wa+Wb=1.

[0010] Furthermore, an adaptive optimization model is deployed for multi-objective business operations, as detailed below: Decision variables: Define binary decision variables X ij , where X ij =1 indicates that the public opinion event i is assigned to the business department j for handling; X ij =0 indicates no allocation; let R be... i C is the comprehensive risk index for public opinion event i; j Estimated resource costs for department j to handle unit tasks; T ij Estimate the response time for department j to handle event i; S ijThe expected customer satisfaction gain after department j handles event i; The multi-objective optimization function is as follows. The model aims to find a set of decision variables X={X_i} that can simultaneously optimize the following four objectives. ij Minimize overall public opinion risk; ; This function calculates the sum of the risk indices of all unassigned events; by minimizing this value, it ensures that the system prioritizes allocating resources to high-risk events, preventing them from being overlooked. Minimize total resource cost: ; This function calculates the total resource cost generated by all allocation schemes; minimizing the total response time improves processing efficiency and quickly quells public opinion. ; Maximize overall customer satisfaction: Improve service quality and enhance user experience; .

[0011] Furthermore, based on the adaptive optimization algorithm, a solution-solving and generation method is used to obtain the optimal work deployment scheme, as follows: In the algorithm initialization phase, semantic association information provided by the power public opinion knowledge graph is used for population guidance generation; the initial work deployment scheme prioritizes the principles of geographical proximity and departmental professional matching: public opinion event clusters in the same geographical area are preferentially assigned to the nearest service team, and specific types of events are accurately matched with departments with corresponding professional capabilities based on the event-solution association relationship established in the knowledge graph; the quality of the initial solution group is significantly improved by introducing domain knowledge, effectively accelerating the algorithm convergence process; the dynamic multi-objective evolutionary optimization adopts the improved NSGA-II as the core optimization engine; a complete feedback control loop is established, and key business indicators after the scheme is executed are continuously tracked through an online effect evaluator, including the negative sentiment decay rate, the positive feedback growth rate, and the task closed-loop cycle.

[0012] Furthermore, the improved NSGA-II is based on a non-dominated sorting genetic algorithm. It automatically adjusts the weight coefficients of risk minimization, cost control, response efficiency, and satisfaction improvement objectives according to the real-time monitored public opinion risk level. When the system identifies a high-risk situation, it automatically increases the weight priority of the risk minimization objective to ensure the generation of emergency-oriented response plans. When designing cross operations, it prioritizes closely related task groups in the knowledge graph as cross units, preserving existing excellent collaborative processing patterns. It establishes a historical evaluation archive of department-task pairing effects, assigns higher mutation probabilities to persistently inefficient pairing combinations, and proactively breaks through performance bottlenecks.

[0013] Furthermore, the visualization dashboard adopts a multi-view linkage visualization dashboard, as detailed below: (1) Parallel coordinate graph view: In this view, each scheme is represented as a broken line that crosses multiple vertical coordinate axes, and the coordinate axes represent the core optimization objectives; decision-makers can intuitively compare the performance of different schemes on each objective; (2) Resource scheduling Gantt chart and map view: After selecting a solution, the dashboard displays the task sequence and load of each department through a Gantt chart; at the same time, on the digital map, the scheduling path of personnel and vehicle resources and the geographical distribution of public opinion events are displayed through the overlay of dynamic flow lines and heat maps, realizing the solution deduction in the spatiotemporal dimension. (3) Scheme comparison panel: The quantitative indicators of the preferred schemes are clearly listed in the form of a table, and the decision-maker can manually adjust the weights. The system will recalculate in real time and highlight the schemes that meet the new preferences.

[0014] A power public opinion data-driven adaptive optimization system for work deployment includes a processor, a memory, and a computer program stored in the memory. When the processor executes the computer program, it specifically performs the steps of the power public opinion data-driven adaptive optimization method for work deployment as described above.

[0015] The present invention has the following beneficial effects: 1. This invention uses real-time public opinion analysis to automatically identify high-risk events before the situation escalates, enabling proactive intervention and changing the traditional passive waiting-for-complaints model; 2. Based on a quantitative public opinion risk index and a multi-objective optimization model, this invention scientifically generates the optimal deployment plan, replacing subjective decision-making that relies on personal experience. This makes resource allocation more precise and constructs a closed loop of "perception-decision-execution-feedback". It can adjust strategies in real time according to the public opinion situation and the execution effect, and realize adaptive scheduling of resources. 3. This invention integrates the capabilities of departments such as operation and maintenance and customer service through a unified model, achieving optimal matching of cross-departmental tasks and resources, and solving the problems of uneven resource availability and insufficient coordination. Attached Figure Description

[0016] Figure 1 This is a flowchart of the method of the present invention; Figure 2 This is a detailed step diagram illustrating the in-depth perception and quantitative assessment of public opinion in one embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of a multi-objective adaptive optimization model in one embodiment of the present invention. Detailed Implementation

[0017] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments: refer to Figure 1 In this embodiment, a data-driven adaptive optimization method for power public opinion deployment is provided, which includes the following steps: S1: Acquire multi-source data related to public opinion on electricity, filter out noisy data based on an integrated model, form a standardized dataset, and build a unified public opinion data pool; S2: Based on the public opinion data pool, construct a public opinion knowledge graph, obtain a comprehensive public opinion risk index, and quantify the impact of public opinion on power companies; S3: By combining the staffing, response time, and processing historical data of each department, construct an adaptive optimization model for multi-objective business deployment; S4: Based on the multi-objective business deployment adaptive optimization model, and the solution solving and generation method based on the adaptive optimization algorithm, obtain the optimal work deployment scheme; S5: Based on the optimal work deployment plan, push it to the decision-maker through a visual dashboard, track the entire process of plan execution, and feed back the execution results and new public opinion data to step S1.

[0018] In this embodiment, the construction of the public opinion knowledge graph is as follows: The knowledge framework of the graph is defined, clarifying entity types, including events, people / organizations, geographical locations, public opinion objects, sentiment tendencies, and solutions, as well as their interrelationships (e.g., "event - occurred at -> geographical location", "event - needs -> solution"). Based on the defined knowledge framework, natural language processing technology is used to extract specific entities, attributes, and relationship instances from the public opinion data. Ambiguity is eliminated through entity linking and other technologies, and heterogeneous data is merged into a unified knowledge unit. The merged knowledge is stored in a graph database (e.g., Neo4j) to generate a visual graph that intuitively presents the relationships, geographical distribution, and evolution paths of public opinion events. The constructed knowledge graph supports in-depth business applications, including performing complex relational queries, analyzing the diffusion path of public opinion impacts, and quickly locating the root cause of events through semantic reasoning, associating and recommending responsible business departments and emergency resources, providing direct and accurate decision-making basis for adaptive optimization models.

[0019] In this embodiment, the comprehensive public opinion risk index R is:

[0020] Where R is the comprehensive public opinion risk index, with a value range of (0, +∞). The larger the value, the higher the risk and the higher the priority of response. S is the sentiment index; H is the popularity value; V is the speed of dissemination; A is the scope of influence; B is the business relevance; Ws, Wh, Wv, Wa, and Wb are the weight coefficients corresponding to the above five dimensions, and satisfy Ws + Wh + Wv + Wa + Wb = 1. The weights can be obtained through expert experience, analytic hierarchy process, or training with historical data. Sentiment index is obtained by analyzing the sentiment polarity (positive, negative, neutral) and intensity of public opinion texts; Key information extraction: using named entity recognition technology, entities related to the power business (such as "power outage", "electricity fee", "substation", "charging pile", etc.) are extracted, and their associated geographical location, time and responsible department are identified; Topic clustering and event identification: using topic models to cluster public opinion, automatically identifying hot topics and emergencies (such as "regional power outage protests", "discussion on new energy policies", etc.), and presenting the evolution path of public opinion; Construct a public opinion knowledge graph that links entities, events, emotions, and geographic location information to create a power-related public opinion knowledge graph that intuitively presents the relationships and evolution paths of public opinion events.

[0021] In this embodiment, an adaptive optimization model is deployed for multi-objective services, as detailed below: Decision variables: Define binary decision variables X ij , where X ij =1 indicates that the public opinion event i is assigned to the business department j for handling; X ij =0 indicates no allocation; let R be... i C is the comprehensive risk index for public opinion event i; j Estimated resource costs (including human, material, and financial resources) for department j to handle unit tasks; T ij Estimate the response time for department j to handle event i (related to department efficiency, skill matching, and geographical distance); S ij The expected customer satisfaction gain after handling event i for department j (based on historical handling performance data); P j This represents the maximum number of tasks that department j can process within the current period (resource capacity constraint). The multi-objective optimization function is as follows. The model aims to find a set of decision variables X={X_i} that can simultaneously optimize the following four objectives. ij Minimize overall public opinion risk (prioritize high-risk events and prevent escalation). ; This function calculates the sum of the risk indices of all unassigned events; by minimizing this value, it ensures that the system prioritizes allocating resources to high-risk events, preventing them from being overlooked. Minimize total resource cost: Conserve resources as much as possible while meeting response requirements. ; This function calculates the total resource cost generated by all allocation schemes; minimizing the total response time improves processing efficiency and quickly quells public opinion. ; Maximize overall customer satisfaction: Improve service quality and enhance user experience; ; This function calculates the sum of the expected satisfaction gains from all allocation schemes; as it is a maximization objective, it directs resources toward the task-department combinations that can most effectively improve satisfaction.

[0022] In this embodiment, the optimal work deployment scheme is obtained based on the adaptive optimization algorithm for scheme solving and generation, as follows: During the algorithm initialization phase, semantic association information provided by the power public opinion knowledge graph is used to guide population generation; the initial work deployment plan prioritizes the principles of geographical proximity and departmental professional matching: public opinion event clusters in the same geographical area are preferentially assigned to the nearest service team, and specific types of events are accurately matched with departments with corresponding professional capabilities based on the event-solution association relationship established in the knowledge graph. This method significantly improves the quality of the initial solution group by introducing domain knowledge, effectively accelerating the algorithm's convergence process. Dynamic multi-objective evolutionary optimization uses an improved NSGA-II as the core optimization engine; Establish a complete feedback control loop and continuously track key business indicators after the implementation of the plan through an online performance evaluator, including the rate of negative emotion decay, the growth rate of positive feedback, and the task closure cycle.

[0023] In this embodiment, the improved NSGA-II is based on a non-dominated sorting genetic algorithm. It automatically adjusts the weight coefficients of risk minimization, cost control, response efficiency, and satisfaction improvement objectives according to the real-time monitored public opinion risk level. When the system identifies a high-risk situation, it automatically increases the weight priority of the risk minimization objective to ensure the generation of emergency-oriented response plans. When designing cross operations, it prioritizes closely related task groups in the knowledge graph as cross units, retaining existing excellent collaborative processing patterns. It establishes a historical evaluation archive of department-task pairing effects, assigns higher mutation probabilities to persistently inefficient pairing combinations, and proactively breaks through performance bottlenecks.

[0024] In this embodiment, the visualization dashboard adopts a multi-view linkage visualization dashboard, as detailed below: (1) Parallel coordinate graph view: In this view, each solution is represented as a broken line that crosses multiple vertical coordinate axes. The coordinate axes represent the core optimization objectives (such as: overall risk reduction rate, estimated total cost, average response time, and expected satisfaction). Decision-makers can intuitively compare the performance of different solutions on each objective and quickly identify options with different focuses, such as low-cost solutions and "high-response" solutions. (2) Resource scheduling Gantt chart and map view: After selecting a solution, the dashboard displays the task sequence and load status of each department (such as power distribution and customer service) through a Gantt chart; at the same time, on the digital map, the scheduling path of personnel and vehicle resources and the geographical distribution of public opinion events are displayed through the overlay of dynamic flow lines and heat maps, realizing the temporal and spatial dimension of solution deduction. (3) Scheme comparison panel: The quantitative indicators of 2-3 preferred schemes are clearly listed in a table and the decision-maker can manually adjust the weights (such as temporarily increasing the importance of "cost"). The system will recalculate in real time and highlight the schemes that meet the new preferences.

[0025] A power sector public opinion data-driven adaptive optimization system for job deployment includes a processor, a memory, and a computer program stored in the memory. When the processor executes the computer program, it specifically performs the steps of a power sector public opinion data-driven adaptive optimization method for job deployment as described above. In this embodiment, after the decision-maker confirms the final plan through the Kanban board, the system ensures accurate execution and full-process monitoring of the plan through the following technical processes: (1) Instruction encapsulation and distribution: The system encapsulates the selected scheme into a standardized JSON instruction set, and through the RESTful API interface, calls the work order system of different business departments (such as power distribution production management system PMS and customer service system CRM) to automatically create and distribute work orders.

[0026] (2) Real-time execution status feedback: While each business system is processing work orders, their status changes (such as "accepted", "in progress", "completed") are fed back to this system in real time through message middleware (such as Kafka / RocketMQ). The system dynamically updates the execution status of all tasks on the Kanban board through color changes and progress bars.

[0027] In this embodiment, the effectiveness of the solution execution is quantified and fed back by the system to form a closed-loop learning process: (1) Effect data collection: The system automatically associates completed work orders with new rounds of public opinion data to calculate key effect indicators, including: actual task closure time, actual resource consumption cost, and the corresponding negative sentiment decline rate and positive feedback growth rate of public opinion.

[0028] (2) Model parameter calibration: Compare the actual performance data with the model predictions. Using the difference, dynamically fine-tune the key parameters in the model (such as the weight coefficients Ws and Wh in step two, and the estimated cost Cj and processing time Tij in step three) through online learning algorithms (such as stochastic gradient descent) to make the model's predictions more and more accurate.

[0029] (3) Knowledge base iteration: Regularly (e.g., monthly) automatically analyze all feedback data within a cycle, mine emerging hot words and event patterns that have not been effectively identified, and after review by the administrator, automatically update the keyword library and risk threshold of public opinion monitoring to achieve continuous evolution of the system's perception capabilities.

[0030] 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.

[0031] 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 processor, 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 and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0032] 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.

[0033] 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.

[0034] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A data-driven adaptive optimization method for work deployment in the field of power public opinion, characterized in that, Includes the following steps: S1: Acquire multi-source data related to public opinion on electricity, filter out noisy data based on an integrated model, form a standardized dataset, and build a unified public opinion data pool; S2: Based on the public opinion data pool, construct a public opinion knowledge graph to obtain a comprehensive public opinion risk index; S3: By combining the staffing, response time, and processing historical data of each department, construct an adaptive optimization model for multi-objective business deployment; S4: Based on the multi-objective business deployment adaptive optimization model, and the solution solving and generation method based on the adaptive optimization algorithm, obtain the optimal work deployment scheme; S5: Based on the optimal work deployment plan, push it to the decision-maker through a visual dashboard, track the entire process of plan execution, and feed back the execution results and new public opinion data to step S1.

2. The adaptive optimization method for work deployment driven by power public opinion data according to claim 1, characterized in that, The construction of the public opinion knowledge graph is as follows: Define the knowledge framework of the graph, clarifying entity types, including events, people / organizations, geographical locations, public opinion objects, sentiment tendencies, and solutions, as well as their interrelationships; based on the defined knowledge framework of the graph, use natural language processing technology to extract specific entities, attributes, and relationship instances from public opinion data, eliminate ambiguity through entity linking and other technologies, and merge heterogeneous data into unified knowledge units; store the merged knowledge in a graph database to generate a visual graph that intuitively presents the relationships, geographical distribution, and evolution paths of public opinion events.

3. The adaptive optimization method for work deployment driven by power public opinion data according to claim 2, characterized in that, The comprehensive public opinion risk index R: ； Where R is the comprehensive public opinion risk index; S is the sentiment index; H is the popularity value; V is the speed of dissemination; A is the scope of influence; B is the business relevance; Ws, Wh, Wv, Wa, and Wb are the weight coefficients corresponding to the above five dimensions, and satisfy Ws+Wh+Wv+Wa+Wb=1.

4. The adaptive optimization method for work deployment driven by power public opinion data according to claim 1, characterized in that, The multi-objective business deployment adaptive optimization model is as follows: Decision variables: Define binary decision variables X ij , where X ij =1 indicates that the public opinion event i is assigned to the business department j for handling; X ij =0 indicates no allocation; let R be... i C is the comprehensive risk index for public opinion event i; j Estimated resource costs for department j to handle unit tasks; T ij Estimate the response time for department j to handle event i; S ij The expected customer satisfaction gain after department j handles incident i; The multi-objective optimization function is as follows. The model aims to find a set of decision variables X={X_i} that can simultaneously optimize the following four objectives. ij Minimize overall public opinion risk; ； This function calculates the sum of the risk indices of all unassigned events; by minimizing this value, it ensures that the system prioritizes allocating resources to high-risk events, preventing them from being overlooked. Minimize total resource cost: ; This function calculates the total resource cost generated by all allocation schemes; minimizing the total response time improves processing efficiency and quickly quells public opinion. ； Maximize overall customer satisfaction: Improve service quality and enhance user experience; 。 5. The adaptive optimization method for work deployment driven by power public opinion data according to claim 1, characterized in that, The proposed method for solving and generating solutions based on adaptive optimization algorithms obtains the optimal work deployment plan as follows: In the algorithm initialization phase, semantic association information provided by the power public opinion knowledge graph is used for population guidance and generation; the initial work deployment plan prioritizes the principles of geographical proximity and departmental professional matching: public opinion event clusters in the same geographical area are preferentially assigned to the nearest service team, and specific types of events are precisely matched with departments with corresponding professional capabilities based on the event-solution association relationship established in the knowledge graph; By introducing domain knowledge, the quality of the initial solution group is significantly improved, which effectively accelerates the algorithm's convergence process. The dynamic multi-objective evolutionary optimization adopts the improved NSGA-II as the core optimization engine. Establish a complete feedback control loop and continuously track key business indicators after the implementation of the plan through an online performance evaluator, including the rate of negative emotion decay, the growth rate of positive feedback, and the task closure cycle.

6. The adaptive optimization method for work deployment driven by power public opinion data according to claim 5, characterized in that, The improved NSGA-II is based on a non-dominated sorting genetic algorithm. It automatically adjusts the weight coefficients of risk minimization, cost control, response efficiency, and satisfaction improvement objectives according to the real-time monitored public opinion risk level. When the system identifies a high-risk situation, it automatically increases the weight priority of the risk minimization objective to ensure the generation of emergency-oriented response plans. When designing cross-operations, it prioritizes closely related task groups in the knowledge graph as cross-operation units, preserving existing excellent collaborative processing patterns. It establishes a historical evaluation archive of department-task pairing effects, assigning higher mutation probabilities to persistently inefficient pairing combinations to proactively break performance bottlenecks.

7. The adaptive optimization method for work deployment driven by power public opinion data according to claim 1, characterized in that, The visualization dashboard uses a multi-view linkage approach, as detailed below: (1) Parallel coordinate graph view: In this view, each scheme is represented as a broken line that crosses multiple vertical coordinate axes, and the coordinate axes represent the core optimization objectives; decision-makers can intuitively compare the performance of different schemes on each objective; (2) Resource scheduling Gantt chart and map view: After selecting a solution, the dashboard displays the task sequence and load of each department through a Gantt chart; at the same time, on the digital map, the scheduling path of personnel and vehicle resources and the geographical distribution of public opinion events are displayed through the overlay of dynamic flow lines and heat maps, realizing the solution deduction in the spatiotemporal dimension. (3) Scheme comparison panel: The quantitative indicators of the preferred schemes are clearly listed in the form of a table, and the decision-maker can manually adjust the weights. The system will recalculate in real time and highlight the schemes that meet the new preferences.

8. A data-driven adaptive optimization system for power sector public opinion deployment, characterized in that, It includes a processor, a memory, and a computer program stored in the memory. When the processor executes the computer program, it specifically performs the steps in the adaptive optimization method for work deployment driven by power public opinion data as described in any one of claims 1-7.

Images