CN-121599425-B - Multi-constraint intelligent environmental sanitation full-scene operation intelligent planning method

CN121599425BCN 121599425 BCN121599425 BCN 121599425BCN-121599425-B

Abstract

The application provides a multi-constraint intelligent environmental sanitation full-scene operation intelligent planning method which is applied to the technical field of data processing. The method comprises the steps of surrounding intelligent sanitation full-scene operation planning, firstly collecting multi-source data such as GIS geography, task demands and equipment states and constraint configuration parameters, generating a structured data set through standardized processing, then constructing ternary association nodes in a layered mode, aggregating characteristics of sliding time windows, generating a multi-constraint directed association graph, extracting constraint characteristic embedded vectors through collaborative optimization graph neural network modeling, importing the constraint characteristic embedded vectors into a layered planning engine, forming a planning basic model through dynamic weight adjustment, scene adaptation calibration and conflict processing, and finally combining real-time data dynamic deduction and multi-objective optimization to output a dynamic operation planning scheme and an execution rule of the adaptive full scene.

Inventors

CHENG ZHUO
FAN YANJUN
WANG CHAO

Assignees

苏州市伏泰信息科技股份有限公司

Dates

Publication Date: 20260508
Application Date: 20260129

Claims (6)

1. A multi-constraint intelligent environmental sanitation full-scene operation intelligent planning method is characterized by comprising the following steps: Acquiring environmental sanitation operation full-scene multi-source core data, wherein the environmental sanitation operation full-scene multi-source core data comprise operation area GIS geographic data, dynamic task demand lists, environmental sanitation equipment Internet of things state data, personnel skill configuration files, real-time environment meteorological data, constraint priority matrixes and dynamic weight coefficients; The method comprises the steps of carrying out regional meshing processing on GIS geographic data based on a geographic space mapping algorithm, analyzing equipment load state by combining equipment Internet of things data, quantitatively configuring matching degree by a personnel skill-task adaptation model, converting environmental meteorological data into operation difficulty coefficients, generating a multi-dimensional constraint aligned structured planning data set, carrying out road layering and grid refining dual processing on the operation regional GIS geographic data based on the geographic space mapping algorithm, dividing operation units and path boundaries, extracting core indexes comprising equipment endurance, operation efficiency and fault early warning by combining equipment Internet of things real-time data flow through a load state analysis model, converting the environmental meteorological data comprising wind speed, rainfall and temperature into operation difficulty coefficients by adopting a meteorological factor weighting algorithm, distinguishing different scene operation complexity, and finally generating a multi-dimensional constraint aligned structured data set with clear geographic boundaries, clear resource states, accurate task matching and reasonable quantification by adopting the personnel skill-task adaptation model; The method comprises the steps of layering a structured planning data set according to the operation urgency and the region complexity, constructing ternary association nodes comprising region-resource-task in each layer, aggregating real-time dynamic characteristics through a sliding time window to generate a full-scene-oriented multi-constraint directed association graph, processing the structured planning data set, layering according to the operation urgency and the region complexity, wherein each layer of graph comprises region nodes, resource nodes and task nodes, aggregating real-time road conditions, equipment states and weather change dynamic characteristics through a sliding time window, extracting regional boundaries of each layer, resource supply and task demand data to generate corresponding subsets, configuring the resources as supply nodes and task demands as target nodes by taking the region as core nodes, generating node attribute information by combining GIS (geographic coordinate system) geographic coordinate mapping and task type identification, and generating side association information according to the resource adaption degree and task execution time sequence relation; Modeling a multi-constraint directed association graph based on a collaborative optimization graph neural network, mining potential association relations among the constraints, generating constraint feature embedding vectors, performing structured acquisition and standardization processing on node attribute, side association strength and dynamic feature time sequence change data of the multi-constraint directed association graph based on constraint association mining requirements and feature extraction targets, generating graph structure basic data comprising constraint types, association weights and time sequence trends, designing modeling rules of the graph data based on sanitation operation constraint collaborative characteristics, defining area-resource-task association thresholds and dynamic feature attenuation coefficients, generating multi-constraint graph modeling specifications, setting a multi-level modeling mechanism for node feature enhancement, side association reinforcement and cross-view feature fusion by combining the collaborative optimization graph neural network architecture and feature embedding requirements, performing integrated execution on the graph structure basic data, the multi-constraint graph modeling specifications and the multi-level modeling mechanism, performing iterative training and optimizing feature mapping through the neural network, and generating constraint feature embedding vectors comprising constraint core features, association semantic information and time sequence dynamic attributes; the constraint feature embedded vector is imported into a hierarchical planning engine, a dynamic weight self-adaptive adjustment mechanism is adopted for optimization, a scene adaptation calibration module is imported, and an adjustment decision chain is generated through constraint conflict visualization at the same time, so that a planning basic model integrating geography, resources and task cooperative logic is formed; Dynamic deduction is carried out based on a planning basic model in combination with real-time road condition data flow and burst task trigger signals, and the operation efficiency, the resource utilization rate and the cost control index are balanced through a multi-objective optimization algorithm, so that an intelligent sanitation operation dynamic planning scheme and an execution rule which are adaptive to a full scene are generated.
2. The method of claim 1, wherein the constrained feature embedded vector is imported into a hierarchical planning engine, optimized using a dynamic weight adaptive adjustment mechanism, imported into a scene adaptation calibration module, and simultaneously an adjustment decision chain is generated through constraint conflict visualization to form a planning base model fusing geographic, resource, task collaborative logic, comprising: Based on constraint collaborative planning requirements and a full-scene adaptation target, building a core framework of a hierarchical planning engine, a dynamic weight adjusting module and a scene adaptation calibration unit, and performing hierarchical analysis and collaborative logic learning on geographic, resource and task constraint feature embedded vectors; designing planning logic based on a multi-constraint balance strategy, and defining the space adaptation dimension of geographic constraint, the supply matching dimension of resource constraint, the priority dimension of task constraint and the cooperative occupation ratio of the three to generate constraint cooperative planning rules; Setting a dynamic weight self-adaptive adjustment mechanism by combining the type of the operation area, the stock of environmental sanitation resources and the adaptive requirements of the emergency degree of the tasks, wherein the core urban area focuses on the visual geography and the resource supply constraint, the burst task scene focuses on the task priority and the resource scheduling constraint, and the suburban operation scene focuses on the geographic range and the cost control constraint; And integrating and executing the core architecture, the constraint collaborative planning rules and the dynamic weight adjustment mechanism, generating an adjustment decision chain through constraint conflict visualization, optimizing planning parameter configuration, and forming a planning basic model which is fused with geographic-resource-task collaborative logic and is suitable for the operation requirements of the whole scene.
3. The method of claim 2, wherein dynamically deducting based on a planning basic model in combination with real-time road condition data flow and burst task trigger signals, balancing operation efficiency, resource utilization rate and cost control indexes through a multi-objective optimization algorithm, generating an intelligent sanitation operation dynamic planning scheme and execution rules adapting to a full scene, and comprising: Dynamically deducting environmental sanitation operation scenes based on a planning basic model in combination with real-time road condition data flow and burst task trigger signals, constructing an efficiency-resource-cost three-dimensional balance model through a multi-objective optimization algorithm, and inputting constraint feature embedded vectors to simulate the adaptation degree of planning schemes in different scenes; When the real-time data characteristics in the dynamic deduction process are extracted, the road condition congestion index and the emergency coefficient of the sudden task are introduced; When analyzing the dynamic characteristic information of the operation scene, associating the resource supply state with the task execution time sequence, and balancing and optimizing a constraint dynamic change chain in a sliding time window; after the planning module outputs a preliminary planning scheme containing multi-dimensional adaptation indexes, the preliminary planning scheme is checked and adjusted by comparing with a historical operation optimal case library, and an intelligent sanitation full-scene operation dynamic planning scheme with scene dynamic adaptation and an execution rule are generated.
4. A multi-constraint intelligent environmental sanitation full scene operation intelligent planning system, characterized in that the system is configured to execute the multi-constraint intelligent environmental sanitation full scene operation intelligent planning method according to any one of claims 1 to 3 by executing executable instructions.
5. An electronic device, comprising: And a memory for storing executable instructions of the first processor; wherein the first processor is configured to execute the multi-constraint intelligent sanitation full scene operation intelligent planning method of any of claims 1-3 via execution of the executable instructions.
6. A computer readable storage medium having stored thereon a computer program, wherein the computer program when executed by a second processor implements the multi-constraint intelligent sanitation full scene operation intelligent planning method of any of claims 1 to 3.

Description

Multi-constraint intelligent environmental sanitation full-scene operation intelligent planning method Technical Field The invention relates to the technical field of data processing, in particular to a multi-constraint intelligent environmental sanitation full-scene operation intelligent planning method. Background Along with the acceleration of the urban process and the upgrading of the environmental sanitation demands, the traditional sanitation operation mode is difficult to adapt to complex and changeable urban treatment scenes, and intelligent sanitation is enabled by technologies such as the Internet of things, artificial intelligence, big data and the like, so that the intelligent sanitation operation mode becomes the core direction for improving the sanitation operation efficiency and optimizing the resource allocation. However, the existing intelligent sanitation operation planning technology still has a plurality of problems to be solved urgently, specifically as follows: the GIS geographic data, equipment Internet of things state data, personnel configuration files, environmental meteorological data and other multi-source information related to sanitation operation are stored in different systems in a scattered mode, and due to incompatibility of equipment protocols and non-unification of data standards, a data island is formed, so that multi-dimensional constraint conditions are difficult to align effectively, and comprehensive support cannot be provided for planning decisions. The existing planning scheme is designed aiming at a single fixed scene, lacks comprehensive consideration of dynamic factors such as operation urgency, regional complexity, burst tasks and the like, and is difficult to meet the requirements of full-scene operation due to the fact that the adjustment efficiency and the adaptation precision of the planning scheme are obviously reduced when the existing planning scheme is used for an irregular scene. In the planning process, potential association among geographic space constraint, resource load constraint and task demand constraint is not fully excavated, and planning is carried out only through simple rules or a single target algorithm, so that multiple targets such as operation efficiency, resource utilization rate, cost control and the like are difficult to cooperatively balance, and the situations of resource idling or task omission are easy to occur. The generalization capability of the existing algorithm model in a complex environment is insufficient, a scene adaptation calibration mechanism aiming at different urban areas and operation types is not established, and visualization and dynamic adjustment capability of constraint conflict are lacking, so that the floor execution effect of a planning scheme is poor, and the environment sanitation operation requirement of differentiation is difficult to adapt. It should be noted that the information disclosed in the above background section is only for enhancing understanding of the background of the present disclosure and thus may include information that does not constitute prior art known to those of ordinary skill in the art. Disclosure of Invention Other features and advantages of the application will be apparent from the following detailed description, or may be learned by the practice of the application. According to one aspect of the application, a multi-constraint intelligent environmental sanitation full-scene operation intelligent planning method is provided, which comprises the steps of obtaining environmental sanitation full-scene multi-source core data, wherein the multi-source core data comprise operation area GIS geographic data, dynamic task demand list, environmental sanitation equipment Internet of things state data, personnel skill configuration files, real-time environmental meteorological monitoring data, constraint priority matrix and dynamic weight coefficient; the method comprises the steps of carrying out regional gridding processing on GIS geographic data based on a geographic space mapping algorithm, analyzing the equipment load state by combining equipment Internet of things data, quantitatively configuring matching degree by a personnel skill-task adaptation model, converting environmental meteorological data into operation difficulty coefficients, generating a multi-dimensional constraint aligned structured planning data set, layering the structured planning data set according to operation urgency and regional complexity, constructing three-way association nodes comprising region-resource-task in each layer, aggregating real-time dynamic characteristics by a sliding time window, generating a multi-constraint directed association graph oriented to full scenes, modeling the multi-constraint directed association graph based on a collaborative optimization graph neural network, mining potential association relation among constraints, generating constraint characteristic embedded vectors, guiding