Cooperative processing method, device and system for polluted carbon at port

Abstract

The invention discloses a port polluted carbon cooperative processing method, device and system. The method comprises the following steps: collecting multi-modal data related to polluted carbon in a port, and carrying out abnormal value cleaning, time sequence alignment and spatial distribution mapping to obtain a standardized data set; establishing a first-layer dynamic game model according to the relationship between the port and the management platform, and establishing a second-layer dynamic game model according to the relationship between the subjects in the port; nodes are generated for pollution sources, emission reduction technologies and policies, edges are generated for the nodes according to the incidence relation between the nodes, and therefore a dynamic knowledge graph is constructed; defining a state space, an action space and a reward function for the reinforcement learning model; and determining a current emission reduction strategy by using the dynamic game model, evaluating a dynamic emission reduction effect of the current emission reduction strategy by using the reinforcement learning model, and determining a pollution source and a pollution propagation path by using the dynamic knowledge graph. The technical problem that in the related technology, the effect of cooperative treatment of sewage and carbon is poor is solved.

Description

Technical Field

[0001] The present application relates to the fields of environmental governance and artificial intelligence, and specifically, to a method, device and system for the coordinated treatment of port carbon pollution. Background Art

[0002] With the rapid development of global trade and port logistics, port pollutant emissions and carbon emissions have become more serious. Existing governance technologies have the following main problems: Difficulties in multi-agent collaboration: 1) Port pollution control involves multiple entities such as the government, port enterprises, logistics companies, and third-party governance agencies. Existing technologies lack an effective mechanism to coordinate the interests of all parties, resulting in poor policy implementation results; 2) Lack of dynamic adaptability: Traditional governance strategies are mostly based on static models, which are difficult to dynamically adjust according to real-time data and environmental changes, and cannot adapt to the complex and changeable operating environment of ports; 3) Insufficient knowledge expression and utilization: There is a lack of systematic integration of multi-source heterogeneous data, and it is impossible to effectively express information such as pollution sources, emission reduction technologies, and policy measures, which limits the scientific nature of accurate tracing and governance solutions; 4) Imperfect feedback mechanism: There is a lack of a closed-loop feedback mechanism, which makes it difficult to timely evaluate governance effects and optimize strategies, resulting in waste of resources and insufficient scientific decision-making.

[0003] In the existing technologies, dynamic game models, knowledge graphs, and enhanced feedback mechanisms have certain applications in different fields. However, in the field of coordinated governance of port carbon pollution, there is no technical solution to organically combine the three to form a complete intelligent governance framework. Summary of the invention

[0004] The embodiments of the present application provide a method, device and system for the collaborative treatment of port carbon pollution. The present solution provides a collaborative governance solution for port carbon pollution based on multi-level dynamic game, knowledge graph technology and enhanced feedback mechanism, which is used to achieve dynamic optimization, precise tracing and real-time adjustment of collaborative governance of port pollutants (such as PM2.5, PM10) and carbon emissions, so as to at least solve the technical problem of poor effect of collaborative governance of carbon pollution in related technologies.

[0005] According to one aspect of an embodiment of the present application, a method for collaborative processing of carbon pollution in ports is provided, comprising: collecting multimodal data related to carbon pollution in ports, and performing outlier cleaning, time series alignment, and spatial distribution mapping to obtain a standardized data set; establishing a first-layer dynamic game model based on the interactive relationship between the port and a management platform obtained from the standardized data set, and establishing a second-layer dynamic game model based on the interactive relationship between various subjects in the port obtained from the standardized data set, wherein the management platform is used to manage the port; generating nodes for pollution sources, emission reduction technologies, and policy measures determined from the standardized data set, and generating edges for the nodes based on the association relationship between the nodes, thereby constructing a dynamic knowledge graph, wherein the attribute values ​​of the nodes in the dynamic knowledge graph are updated in real time according to the status of the pollution sources in the port, and the weight values ​​of the edges are updated in real time according to the changes in the propagation path; defining a state space, an action space, and a reward function for a reinforcement learning model; in the process of collaborative processing of carbon pollution in ports, the first-layer dynamic game model and the second-layer dynamic game model are used to determine the current emission reduction strategy, the reinforcement learning model is used to evaluate the dynamic emission reduction effect of the current emission reduction strategy, and the dynamic knowledge graph is used to determine the pollution sources and pollution propagation paths.

[0006] Optionally, a first-layer dynamic game model is established based on the interactive relationship between the port and the management platform obtained from the standardized data set, including: constructing the payment function of the port: U port (t) = R profit (t) - C tech (s port , t) + S subsidy (s gov , t) - P penalty (s gov ,t) , R profit (t) represents the economic benefits of the port, C tech (s port ,t) represents the investment and operating costs of emission reduction technologies, S subsidy (s gov ,t) Indicates subsidies provided by the government. P penalty (s gov ,t) It means the fine for exceeding the emission limit; Construct the payment function of the management platform government: gov (t) = B env (t) - S subsidy (s gov , t) + P penalty (s gov ,t)− C admin( t) , B env (t) represents the environmental benefits of reduced pollutant and carbon emissions, C admin (t) Represents the administrative cost of policy implementation.

[0007] Optionally, a second-layer dynamic game model is established based on the interactive relationship between various entities in the port obtained from the standardized data set, including: constructing a profit function for the entities in the port: π i (s i ,t)=π i base(t) - γ × E pollution (t) , π i (s i ,t) Representation subject i of income, E pollution (t) represents the impact value of pollution diffusion, γ Represents the weight coefficient of pollution diffusion. The game players in the port include port enterprises, logistics companies and third-party governance agencies.

[0008] Optionally, the attribute values ​​of nodes in the dynamic knowledge graph are updated in real time according to the following formula: Attr node (t + 1) = Attr node (t) + δ × ΔAttr(t) , Attr node (t) Indicates that the node is at time t , Attr node (t+1) Indicates that the node is at timet+1 , δ is the update step size, ΔAttr(t) Indicates the attribute change value; The weight values ​​of edges in the dynamic knowledge graph are updated in real time according to the following formula: w ij (t + 1) = w ij (t) + η × Δw ij (t) , w ij (t) Indicates the edge at time t The weight value when w ij (t+1) Indicates the edge at time t+1 The weight value when η is the weight adjustment coefficient, Δw ij (t) Indicates the edge weight change value.

[0009] Optionally, defining a state space, an action space, and a reward function for a reinforcement learning model includes: defining a state space including the following state variables: pollutant emissions, carbon emissions, equipment operating status, and policy parameters; defining an action space including the following decision actions: selection of emission reduction technology, adjustment of policy incentives, and adjustment of equipment operating parameters; defining a reward function: R(s, a) = w 1 × ΔE pollution (s, a) + w 2× ΔE carbon (s, a) - w 3 ×C total (s, a) , ΔE pollution (s, a) Indicates the current status s and actions a The amount of pollutant reduction achieved, ΔE carbon (s, a) Indicates the current status s and actions a The amount of carbon emissions reduction achieved, C total (s, a) Indicates the current status of execution s and actions a The total cost consumed, w1, w2, w3 Represents the corresponding weight coefficient; definition Q Value update formula: , Q(s, a) Indicates status s and actions a The value of α Indicates control Q The learning rate of the value update speed, γ represents the discount factor that determines the importance of future rewards, λ represents the penalty factor for pollution spread, max a′ Q(s’, a′) The optimal strategy for the future Q Value, indicating that in the state s′ Choose the best action a′ The maximum benefit that can be obtained.

[0010] Optionally, the optimal path for pollution propagation is calculated based on the shortest path algorithm modified by the pollution propagation effect : min , w ij To indicate the pollutant at the node i and nodes j The pollution transmission weight of the transmission intensity between E ij (t) To represent pollutants from nodes i Propagate to nodes j Diffusion effect parameters.

[0011] According to another aspect of the embodiment of the present application, a collaborative processing device for port carbon pollution is also provided, including: a collection unit, which is used to collect multimodal data related to carbon pollution in the port, and perform outlier cleaning, time series alignment and spatial distribution mapping to obtain a standardized data set; a game model creation unit, which is used to establish a first-level dynamic game model based on the interactive relationship between the port and the management platform obtained from the standardized data set, and to establish a second-level dynamic game model based on the interactive relationship between various subjects in the port obtained from the standardized data set. The management platform is used to manage the port; a knowledge graph construction unit, which is used to construct a knowledge graph for the pollution sources, emission reduction technologies and policies determined from the standardized data set; The system generates nodes based on the policies and measures, and generates edges for the nodes according to the association between the nodes, so as to construct a dynamic knowledge graph. The attribute values ​​of the nodes in the dynamic knowledge graph are updated in real time according to the status of the pollution sources in the port, and the weight values ​​of the edges are updated in real time according to the changes in the propagation path; the reinforcement learning model construction unit is used to define the state space, action space and reward function for the reinforcement learning model; the collaborative processing unit is used to determine the current emission reduction strategy using the first-layer dynamic game model and the second-layer dynamic game model in the process of collaborative processing of port pollution carbon, use the reinforcement learning model to evaluate the dynamic emission reduction effect of the current emission reduction strategy, and use the dynamic knowledge graph to determine the pollution source and pollution propagation path.

[0012] According to another aspect of the embodiment of the present application, a collaborative processing system for port carbon pollution is also provided, including: a collection device for collecting multimodal data related to carbon pollution in the port, and performing outlier cleaning, time series alignment and spatial distribution mapping to obtain a standardized data set; a multi-level dynamic game model, including a first-level dynamic game model established based on the interactive relationship between the port and the management platform obtained from the standardized data set and a second-level dynamic game model established based on the interactive relationship between various subjects in the port obtained from the standardized data set, the management platform is used to manage the port; a dynamic knowledge graph, the nodes in the dynamic knowledge graph are used to represent the interaction between the port and the management platform obtained from the standardized data set. The pollution sources, emission reduction technologies and policy measures determined in the standardized data set, the edges in the dynamic knowledge graph are used to represent the association relationship between nodes, the attribute values ​​of the nodes in the dynamic knowledge graph are updated in real time according to the status of the pollution sources in the port, and the weight values ​​of the edges are updated in real time according to the changes in the propagation path; the reinforcement learning model includes a predefined state space, action space and reward function; in the process of collaborative treatment of port carbon pollution, the first-layer dynamic game model and the second-layer dynamic game model are used to determine the current emission reduction strategy, the reinforcement learning model is used to evaluate the dynamic emission reduction effect of the current emission reduction strategy, and the dynamic knowledge graph is used to determine the pollution sources and pollution propagation paths.

[0013] According to another aspect of an embodiment of the present application, a storage medium is further provided, which includes a stored program, and the above method is executed when the program is run.

[0014] According to another aspect of an embodiment of the present application, an electronic device is further provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the above method through the computer program.

[0015] According to one aspect of the present application, a computer program product or a computer program is provided, the computer program product or the computer program comprising computer instructions, the computer instructions being stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device performs the steps of any one of the embodiments of the above method.

[0016] In an embodiment of the present application, multimodal data related to carbon pollution in the port are collected, and outlier cleaning, time series alignment and spatial distribution mapping are performed to obtain a standardized data set; a first-layer dynamic game model is established based on the interactive relationship between the port and the management platform obtained from the standardized data set, and a second-layer dynamic game model is established based on the interactive relationship between various subjects in the port obtained from the standardized data set, and the management platform is used to manage the port; nodes are generated for the pollution sources, emission reduction technologies and policy measures determined from the standardized data set, and edges are generated for the nodes according to the association relationship between the nodes, so as to construct a dynamic knowledge graph, and the attribute values ​​of the nodes in the dynamic knowledge graph are updated in real time according to the status of the pollution sources in the port, and the weight values ​​of the edges are updated in real time according to the changes in the propagation path; the state space, action space and reward function are defined for the reinforcement learning model; in the process of collaborative processing of carbon pollution in the port, the first-layer dynamic game model and the second-layer dynamic game model are used to determine the current emission reduction strategy, the reinforcement learning model is used to evaluate the dynamic emission reduction effect of the current emission reduction strategy, and the dynamic knowledge graph is used to determine the pollution sources and pollution propagation paths, thereby solving the technical problem of poor effect of carbon pollution collaborative governance in related technologies. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings: Figure 1 is a flow chart of an optional coordinated treatment method of port carbon pollution according to an embodiment of the present application; Figure 2 is a schematic diagram of an optional coordinated treatment scheme for port carbon pollution according to an embodiment of the present application; Figure 3 is a schematic diagram of an optional coordinated treatment scheme for port carbon pollution according to an embodiment of the present application; Figure 4is a schematic diagram of an optional coordinated treatment device for port carbon pollution according to an embodiment of the present application; Figure 5 It is a structural block diagram of a terminal according to an embodiment of the present application. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the drawings here can be arranged and designed in various different configurations.

[0019] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0020] First, some nouns or terms that appear in the description of the embodiments of the present application are subject to the following interpretation: According to one aspect of the embodiment of the present application, an embodiment of a coordinated treatment system for port carbon pollution is provided, including: The collection equipment is used to collect multimodal data related to carbon pollution in the port, and perform outlier cleaning, time series alignment and spatial distribution mapping to obtain a standardized data set; a multi-level dynamic game model (simulating the interactive relationship between the port and the government, and between multiple subjects within the port to achieve balanced optimization of emission reduction costs and benefits), including a first-level dynamic game model established based on the interactive relationship between the port and the management platform obtained from the standardized data set and a second-level dynamic game model established based on the interactive relationship between various subjects in the port obtained from the standardized data set, wherein the management platform is used to manage the port; a dynamic knowledge graph (i.e., a knowledge graph model: integrating multi-source heterogeneous data, Build an association network between pollution sources and emission reduction technologies, policy measures, and achieve full coverage of pollution links and accurate traceability). The nodes in the dynamic knowledge graph are used to represent the pollution sources, emission reduction technologies, and policy measures determined from the standardized data set. The edges in the dynamic knowledge graph are used to represent the association relationship between nodes. The attribute values ​​of the nodes in the dynamic knowledge graph are updated in real time according to the status of the pollution sources in the port, and the weight values ​​of the edges are updated in real time according to the changes in the propagation path. Reinforcement learning model (provides a reinforcement feedback mechanism, based on the reinforcement learning algorithm, designs dynamic feedback and strategy optimization methods to achieve continuous improvement in governance effects), including pre-defined state space, action space, and reward function.

[0021] In the process of collaborative treatment of port carbon pollution, the first-layer dynamic game model and the second-layer dynamic game model are used to determine the current emission reduction strategy, the reinforcement learning model is used to evaluate the dynamic emission reduction effect of the current emission reduction strategy, and the dynamic knowledge graph is used to determine the pollution source and pollution propagation path.

[0022] According to another aspect of the embodiments of the present application, a method embodiment of a method for collaboratively treating port carbon pollution is also provided.

[0023] Figure 1 is a flow chart of an optional coordinated treatment method for port carbon pollution according to an embodiment of the present application, such as Figure 1 As shown, the method may include the following steps: Step S102, collect multimodal data related to carbon pollution in the port, and perform outlier cleaning, time series alignment and spatial distribution mapping to obtain a standardized data set. Time series alignment includes: time standardization, unified timestamp format and time zone; interpolation processing, interpolation of missing time points to ensure data continuity; resampling, adjusting the sampling frequency to make the time points of different data sources consistent. Spatial distribution mapping: coordinate unification, converting the spatial coordinates of different data sources to the same reference system; spatial interpolation, interpolation of missing areas to ensure spatial continuity; gridding processing, mapping the data to a unified grid for subsequent analysis.

[0024] Step S104, a first-layer dynamic game model is established based on the interactive relationship between the port and the management platform obtained from the standardized data set, and a second-layer dynamic game model is established based on the interactive relationship between various entities in the port obtained from the standardized data set. The management platform is used to manage the port and can be the government or a third-party manager designated by the government.

[0025] Step S106 generates nodes for the pollution sources, emission reduction technologies and policy measures determined from the standardized data set, and generates edges for the nodes based on the associations between the nodes, thereby constructing a dynamic knowledge graph. The attribute values ​​of the nodes in the dynamic knowledge graph are updated in real time according to the status of the pollution sources in the port, and the weight values ​​of the edges are updated in real time according to the changes in the propagation path.

[0026] Step S108, defining a state space, an action space, and a reward function for the reinforcement learning model.

[0027] Step S110, in the process of collaborative treatment of port carbon pollution, the first-layer dynamic game model and the second-layer dynamic game model are used to determine the current emission reduction strategy (such as the selected emission reduction technology or emission reduction technology combination, the formulation of policy optimization plans, etc.), the reinforcement learning model is used to evaluate the dynamic emission reduction effect of the current emission reduction strategy, and the dynamic knowledge graph is used to determine the pollution source and pollution propagation path.

[0028] This application proposes an innovative coordinated governance scheme for port pollution and carbon emissions: 1) Dynamic coordinated governance: Through a multi-level dynamic game model, dynamically simulate the cooperation and conflict relationship between multiple subjects, and adjust the emission reduction strategy in real time; 2) Comprehensive pollution source tracing: Using knowledge graph technology, comprehensively cover different types of pollution sources in ports, support accurate tracing and causal analysis; 3) Intelligent feedback optimization: Through strengthening the feedback mechanism, dynamically evaluate the emission reduction effect, optimize policy incentives and technical configuration. This can solve the technical problem of poor effect of coordinated governance of pollution and carbon emissions in related technologies.

[0029] As an optional embodiment, the following is combined with the specific implementation method. Figure 2 , Figure 3 Take this as an example to further describe the technical solution of this application: Step 1: Multimodal data collection and preprocessing 1) Data collection Real-time data collection: collect data such as pollutant concentration, equipment operating status, traffic flow, etc. in the port area through fixed surveillance cameras, drones and sensor networks; auxiliary data sources: including GPS positioning of port equipment, IMU (inertial measurement unit) data, and environmental characteristics of lidar and infrared detection; historical data integration: import historical emissions, policy implementation effects and emission reduction technology application records.

[0030] 2) Data preprocessing Outlier cleaning: remove abnormal device data and invalid sensor readings; time synchronization and spatial alignment: perform time series alignment and spatial distribution mapping of multi-source data to generate a standardized data set.

[0031] Step 2: Construction of multi-level dynamic game model Aiming at the complex collaborative relationship among multiple entities in the port, a two-layer dynamic game model is designed: 1) First-level game: collaborative game between ports and governments Game players: port enterprises and government departments.

[0032] Payment function innovation: U port (t) = R profit (t) - C tech (s port , t) + S subsidy (s gov , t) - P penalty (s gov ,t) , R profit (t) : Economic benefits of port enterprises; C tech (s port ,t) : Investment and operating costs of emission reduction technologies; S subsidy (s gov ,t) : Subsidies provided by the government; P penalty (s gov ,t) : Fines for exceeding emission standards.

[0033] The government's payment function is: U gov (t) = B env (t) - S subsidy (s gov , t) + P penalty (s gov , t) - C admin( t) , B env (t) : Environmental benefits from reduced pollutants and carbon emissions; C admin (t) : Administrative costs of policy implementation.

[0034] Support dynamic weight adjustment, specifically the payment function parameters can be dynamically adjusted according to pollution traceability data: R profit( t + 1) = R profit (t) + α × ΔR, B env (t + 1) = B env (t) + β × ΔB。

[0035] 2) Second-level game: multi-agent game within the port Game players: port enterprises, logistics companies, and third-party governance agencies.

[0036] In the optimization of evolutionary game strategy, the pollution diffusion model is introduced to modify the profit function: π i (s i ,t)=π i base(t) - γ × E pollution (t) , π i (s i ,t) :main body i of income; E pollution (t) : Impact value of pollution diffusion; γ : Weight coefficient of pollution diffusion.

[0037] Step 3: Construction and application of dynamic knowledge graph model 1) Dynamic construction of knowledge graph Node definition: including pollution source nodes (equipment, vehicles, ships), emission reduction technology nodes, policy measure nodes; relationship definition: cause and effect relationship (emissions and pollution sources), dependency relationship (technology and policy), etc.

[0038] 2) Dynamic update mechanism Node attributes are dynamically updated: Attr node (t + 1) = Attr node(t) + δ × ΔAttr(t), The parameters for dynamic update of node attributes are explained as follows: Attr node (t) : Node attribute value, describing a node in the knowledge graph (such as pollution source node, emission reduction technology node, policy node) at a certain time t For example, the attribute value of the pollution source node may be the pollution emission, equipment utilization rate, etc. ΔAttr(t) : Attribute change, calculated from real-time data (such as sensors or historical records), indicating the change of node attributes over time t For example, a sudden increase in the pollution emissions of a certain device ,ΔAttr(t) That is, it is a positive value; δ : Update step size, used to adjust the update speed of node attributes, usually set according to actual conditions, larger δ Indicates a faster response to real-time data, while a smaller δ is used to smooth the response and prevent excessive fluctuations.

[0039] Dynamic adjustment of edge weights: w ij (t + 1) = w ij (t) + η × Δw ij (t) , where the parameters for dynamic adjustment of edge weights are explained as follows: w ij (t) : Edge weight, representing the node i and nodes j The strength of the relationship between them. For example, the weight between a pollution source and an emission reduction technology node may reflect the strength of the effect of the technology on pollution source control. Δw ij (t) : The change in edge weight, which is calculated in real time from pollution data or governance effects, and represents the change in the relationship between two nodes. For example, if a certain emission reduction technology has a stronger governance effect on a certain pollution source, it will be a positive value; η : Adjustment coefficient, which controls the sensitivity of edge weight update. It is similar to the learning rate. A larger value indicates a faster update of relationship changes, which is suitable for scenarios with fast response; a smaller value is suitable for scenarios with slowly changing relationships.

[0040] 3) Pollution source tracing and transmission path analysis Optimal path calculation: Shortest path algorithm based on correction of pollution propagation effect: , where the parameters for pollution source tracing and transmission path analysis are explained as follows: w ij : Pollution propagation weight, indicating the pollution spread at the nodei and nodes j The intensity of transmission between them is calculated in combination with the impact of pollution diffusion; E ij (t) : Pollution propagation effect, reflecting the pollution from the node i Propagate to nodes j The specific diffusion effects are usually estimated from physical models (e.g., diffusion equations) or sensor data; : Optimal path parameter, which indicates the optimal tracing path of pollution propagation calculated through knowledge graph analysis.

[0041] Step 4: Design and implement a strengthened feedback mechanism 1) Reinforcement Learning Model State space S: defines pollution state, equipment state and policy parameters. It includes various state information in the port system. The main parameters are: E pollution : Pollutant emission levels; E carbon : Carbon emissions; Tech status : Equipment operating status; Policy param : Policy incentive parameters.

[0042] These statuses together describe the overall status of port pollution control.

[0043] Action space A: includes emission reduction technology selection and policy adjustments, including possible strategies or operations in the process of port pollution control. The main parameters are: Tech apply : selection and application of specific emission reduction technologies; Device adjust : Equipment operating parameters adjustment; Policy update :Optimization and adjustment of policy incentives.

[0044] 2) Reward function design R(s, a) = w 1 × ΔE pollution (s, a) + w 2× ΔE carbon (s, a) - w 3 ×C total (s, a) in: ΔE pollution (s, a) : Pollutant emission reduction, describing the current status s and actionsa Contribution to the reduction of pollutant emissions, the larger the value, the better the emission reduction effect; ΔE carbon (s, a) : Carbon emission reduction, describing the current status s and actions a Contribution to carbon emission reduction, the larger the value, the better the carbon emission reduction effect; C total (s, a) : Total cost, including the investment cost of emission reduction technology, equipment maintenance cost and policy implementation cost. The smaller the value, the lower the cost. w1, w2, w3 : Weight coefficient, used to balance the relationship between pollutant emission reduction, carbon emission reduction and cost. The setting of weight depends on the specific application scenario and governance objectives.

[0045] 3) Q value update , in: Q(s, a) :state s and actions a The value of s Next action a long-term benefits; α : learning rate, control Q The speed of value update, a larger α indicates fast learning, a smaller α indicates stable learning; γ : Discount factor, which determines the importance of future rewards. Values ​​closer to 1 indicate more focus on long-term benefits, and values ​​closer to 0 indicate more focus on current benefits. λ : Pollution propagation penalty factor, an additional penalty parameter introduced to reduce Q The value indicates the negative impact of pollutant diffusion on the algorithm. The larger the value, the stronger the constraint on pollution diffusion. max a′ Q(s’, a′) : The optimal strategy for the future Q Value, indicating that in the state s′ Choose the best action a′ The maximum benefit that can be obtained, E pollution (s) Indicates the current status s The cumulative pollution level brought by itself.

[0046] Step 5: Result storage and display Data storage: storage of carbon footprint inventories, pollution source tracing paths, and decision optimization records.

[0047] Graphical output: Displays carbon footprint distribution, equipment ranking, and optimization suggestions.

[0048] The present application also provides a preferred embodiment, which further describes the technical solution of the present application in detail in combination with specific scenarios: Application background: A coastal port has serious pollution problems. The main pollution sources come from: diesel trucks in the port, ship exhaust, yard dust, etc.

[0049] Governance challenges: high governance costs, uncertain policy implementation results, and lack of collaboration among different entities.

[0050] The goal is to use this invented system to synergistically optimize pollutant emissions and carbon reduction, develop scientific pollution control strategies, and achieve a balance between economic and environmental benefits.

[0051] Input, output and application process 1) Input data Pollution emission data: atmospheric pollutant concentrations collected by real-time sensors (PM2.5, SO2, NOx, etc.), emission data from ships and vehicles (such as exhaust emissions), and background pollution data provided by environmental monitoring stations; equipment and facility data: operating status data of port trucks, cranes, ships and other equipment (GPS, IMU, etc.), and dust emissions from storage yards and logistics warehouses; policy and historical data: historical emissions and governance effect data, current government policies (such as subsidy intensity, emission standards, and penalty rules); economic and technical data: emission reduction technology costs, equipment upgrade costs, and carbon trading market prices.

[0052] 2) System application process Phase 1: Data collection and preprocessing Real-time data collection: Sensors monitor the concentration of pollutants such as PM2.5 and NOx. Equipment operating status (such as ship berthing time and port truck operation trajectory) is collected through IoT devices.

[0053] Data cleaning: Remove abnormal data (such as extreme values ​​caused by sensor failure).

[0054] Data integration: Integrate pollution emission data, equipment operation data, policy data, etc. into a unified format to facilitate subsequent model use.

[0055] Phase 2: Dynamic Knowledge Graph Construction Node generation: Generate pollution source nodes, such as diesel trucks, ships, and storage yards; generate emission reduction technology nodes, such as exhaust treatment devices, low-sulfur fuel technology, and spraying dust suppressants; generate policy nodes, such as emission standards and government subsidies.

[0056] Relationship construction: construct the dependency relationship between pollution sources and technology nodes (such as low-sulfur fuel and ship exhaust); construct causal relationships (such as NOx emissions from diesel trucks and deteriorating air quality); construct the temporal relationship between policies and technologies (such as the delay of subsidies on the effect of technology implementation).

[0057] Dynamic update: real-time adjustment of node attributes (such as pollution source emissions, technology emission reduction effects); adjustment of edge weights (such as the impact of policy implementation on pollution control).

[0058] Phase 3: Multi-level dynamic game modeling The first level of game (port and government): port enterprises choose emission reduction technologies, and the government sets subsidies or fines; simulate the responses of port enterprises under different policies to find the balance between economic and environmental benefits.

[0059] The second level of game (multiple entities within the port): port enterprises, logistics companies, and third-party governance agencies cooperate or compete in emission reduction; dynamically evolve strategies to optimize internal resource allocation.

[0060] Phase 4: Strengthening feedback mechanism optimization Reinforcement Learning Model: Define the state space: pollutant emission levels, equipment operating status, policy parameters.

[0061] Define the action space: technology selection, equipment adjustment, policy optimization.

[0062] Design reward function: R = w 1 ⋅ Emission reduction +w 2 ⋅ Carbon Emission Reduction −w 3 ⋅ treat Management costs.

[0063] Real-time updates Q value, dynamically optimize emission reduction strategies.

[0064] Phase 5: Result storage and visualization Storage: record pollution source traceability, carbon footprint inventory, and optimization strategy history; Visualization: Generate pollution distribution maps, carbon footprint heat maps, equipment emission rankings, etc.

[0065] 3) Output results Pollution source tracing path: Identify the main pollution sources (such as diesel trucks contribute 60% of total NOx emissions); provide pollution transmission paths (such as dust spreading from the storage yard to residential areas).

[0066] Optimization strategy: recommend the best combination of emission reduction technologies (such as upgrading exhaust treatment devices for port trucks and using low-sulfur fuel for ships); formulate policy optimization plans (such as increasing subsidies by 10% to significantly improve the rate of technology application).

[0067] Economic and environmental benefit analysis: economic costs are reduced (e.g., emission reduction costs are reduced by 15% through subsidies and carbon trading); environmental benefits are improved (e.g., PM2.5 concentration is reduced by 20% and carbon emissions are reduced by 30%).

[0068] Dynamic monitoring and feedback: adjust strategies in real time to ensure that pollutant emissions always meet standards.

[0069] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0070] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0071] According to another aspect of an embodiment of the present application, a coordinated treatment device for port carbon pollution is provided for implementing the coordinated treatment method for port carbon pollution described above. Figure 4 is a schematic diagram of an optional coordinated treatment device for port pollution carbon according to an embodiment of the present application, such as Figure 4 As shown, the device may include: The collection unit 41 is used to collect multimodal data related to carbon pollution in the port, and perform outlier cleaning, time series alignment and spatial distribution mapping to obtain a standardized data set; A game model creation unit 43 is used to establish a first-level dynamic game model according to the interactive relationship between the port and the management platform obtained from the standardized data set, and to establish a second-level dynamic game model according to the interactive relationship between various entities in the port obtained from the standardized data set, wherein the management platform is used to manage the port; The knowledge graph construction unit 45 is used to generate nodes for the pollution sources, emission reduction technologies and policy measures determined from the standardized data set, and to generate edges for the nodes according to the association relationship between the nodes, so as to construct a dynamic knowledge graph, wherein the attribute values ​​of the nodes in the dynamic knowledge graph are updated in real time according to the status of the pollution sources in the port, and the weight values ​​of the edges are updated in real time according to the changes in the propagation path; A reinforcement learning model building unit 47, used to define a state space, an action space and a reward function for the reinforcement learning model; The collaborative processing unit 49 is used to determine the current emission reduction strategy using the first-layer dynamic game model and the second-layer dynamic game model during the collaborative processing of port carbon pollution, evaluate the dynamic emission reduction effect of the current emission reduction strategy using the reinforcement learning model, and determine the pollution source and pollution propagation path using the dynamic knowledge graph.

[0072] Optionally, the game model creation unit is further used to: construct a payment function of the port: U port (t) = R profit (t) - C tech (s port , t) + S subsidy (s gov , t) - P penalty (s gov ,t) , in, R profit (t) represents the economic benefits of the port, C tech (s port ,t) represents the investment and operating costs of emission reduction technologies, S subsidy (s gov ,t) Indicates subsidies provided by the government. P penalty (s gov ,t) It means the fine for exceeding the emission limit; Construct the payment function of the management platform government: gov (t) = B env (t) - S subsidy (s gov , t) + Ppenalty (s gov ,t)− C admin( t) , in, B env (t) represents the environmental benefits of reduced pollutant and carbon emissions, C admin (t) Represents the administrative cost of policy implementation.

[0073] Optionally, the game model creation unit is further used to: construct a profit function for the subject in the port: π i (s i ,t)=π i base(t) - γ × E pollution (t) , in, π i (s i ,t) Representation subject i of income, E pollution (t) represents the impact value of pollution diffusion, γ Represents the weight coefficient of pollution diffusion. The game players in the port include port enterprises, logistics companies and third-party governance agencies.

[0074] Optionally, the attribute values ​​of nodes in the dynamic knowledge graph are updated in real time according to the following formula: Attr node (t + 1) = Attr node (t) + δ × ΔAttr(t) ,in, Attr node (t) Indicates that the node is at time t , Attr node (t+1) Indicates that the node is at time t+1 , δ is the update step size, ΔAttr(t) Indicates the attribute change value; The weight values ​​of edges in the dynamic knowledge graph are updated in real time according to the following formula: w ij (t + 1) = w ij(t) + η × Δw ij (t) ,in, w ij (t) Indicates the edge at time t The weight value when w ij (t+1) Indicates the edge at time t+1 The weight value when η is the weight adjustment coefficient, Δw ij (t) Indicates the edge weight change value.

[0075] Optionally, the reinforcement learning model building unit is further used to: define a state space including the following state variables: pollutant emissions, carbon emissions, equipment operating status, policy parameters; define an action space including the following decision actions: selection of emission reduction technology, adjustment of policy incentives, adjustment of equipment operating parameters; define a reward function: R(s, a) = w 1 × Δ E pollution (s, a) + w 2× ΔE carbon (s, a) - w 3 ×C total (s, a) ,in, ΔE pollution (s, a) Indicates the current status s and actions a The amount of pollutant reduction achieved, ΔE carbon (s, a) Indicates the current status s and actions a The amount of carbon emissions reduction achieved, C total (s, a) Indicates the current status of execution s and actions a The total cost consumed, w1, w2, w3 Represents the corresponding weight coefficient; definition Q Value update formula: , in, Q(s, a) Indicates status s and actions a The value of α Indicates control Q The learning rate of the value update speed, γ represents the discount factor that determines the importance of future rewards,λ represents the penalty factor for pollution spread, max a′ Q(s’, a′) The optimal strategy for the future Q Value, indicating that in the state s′ Choose the best action a′ The maximum benefit that can be obtained.

[0076] Optionally, the optimal path for pollution propagation is calculated based on the shortest path algorithm modified by the pollution propagation effect : min , in, w ij To indicate the pollutant at the node i and nodes j The pollution transmission weight of the transmission intensity between E ij (t) To represent pollutants from nodes i Propagate to nodes j Diffusion effect parameters.

[0077] According to another aspect of an embodiment of the present application, a server or terminal for implementing the above-mentioned collaborative treatment method of port carbon pollution is also provided.

[0078] Figure 5 is a structural block diagram of a terminal according to an embodiment of the present application, such as Figure 5 As shown, the terminal may include: one or more (only one is shown in the figure) processors 501, a memory 503, and a transmission device 505, such as Figure 5 As shown, the terminal may also include an input and output device 507.

[0079] Among them, the memory 503 can be used to store software programs and modules, such as the program instructions / modules corresponding to the coordinated processing method and device of port pollution carbon in the embodiment of the present application. The processor 501 executes various functional applications and data processing by running the software programs and modules stored in the memory 503, that is, realizing the above-mentioned coordinated processing method of port pollution carbon. The memory 503 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 503 may further include a memory remotely arranged relative to the processor 501, and these remote memories can be connected to the terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0080] The above-mentioned transmission device 505 is used to receive or send data via a network, and can also be used for data transmission between a processor and a memory. Specific examples of the above-mentioned network may include a wired network and a wireless network. In one example, the transmission device 505 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices and routers via a network cable so as to communicate with the Internet or a local area network. In one example, the transmission device 505 is a radio frequency (Radio Frequency, RF) module, which is used to communicate with the Internet wirelessly.

[0081] Specifically, the memory 503 is used to store application programs.

[0082] The processor 501 may call the application stored in the memory 503 through the transmission device 505 to perform the following steps: Collect multimodal data related to carbon pollution in the port, and perform outlier cleaning, time series alignment and spatial distribution mapping to obtain a standardized data set; establish a first-layer dynamic game model based on the interactive relationship between the port and the management platform obtained from the standardized data set, and establish a second-layer dynamic game model based on the interactive relationship between various subjects in the port obtained from the standardized data set, wherein the management platform is used to manage the port; generate nodes for the pollution sources, emission reduction technologies and policy measures determined from the standardized data set, and generate edges for the nodes according to the association relationship between the nodes, so as to construct a dynamic knowledge graph, wherein the attribute values ​​of the nodes in the dynamic knowledge graph are updated in real time according to the status of the pollution sources in the port, and the weight values ​​of the edges are updated in real time according to the changes in the propagation path; define the state space, action space and reward function for the reinforcement learning model; in the process of collaboratively processing the carbon pollution in the port, use the first-layer dynamic game model and the second-layer dynamic game model to determine the current emission reduction strategy, use the reinforcement learning model to evaluate the dynamic emission reduction effect of the current emission reduction strategy, and use the dynamic knowledge graph to determine the pollution sources and pollution propagation paths.

[0083] Optionally, the specific examples in this embodiment may refer to the examples described in the above embodiments, and this embodiment will not be described in detail here.

[0084] It can be understood by those skilled in the art that Figure 5 The structure shown is for illustration only, and the terminal may be a smart phone (such as an Android phone, an iOS phone, etc.), a tablet computer, a PDA, a mobile Internet device (Mobile Internet Devices, MID), a PAD, or other terminal devices. Figure 5 It does not limit the structure of the above electronic device. For example, the terminal may also includeFigure 5 More or fewer components (such as network interfaces, display devices, etc.) shown in, or with Figure 5 Different configurations are shown.

[0085] A person of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing the hardware related to the terminal device through a program, and the program can be stored in a computer-readable storage medium, and the storage medium may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0086] The embodiment of the present application further provides a storage medium. Optionally, in this embodiment, the storage medium can be used to execute the program code of the coordinated treatment method of port carbon pollution.

[0087] Optionally, in this embodiment, the storage medium may be located on at least one network device among a plurality of network devices in the network shown in the above embodiment.

[0088] Optionally, in this embodiment, the storage medium is configured to store program codes for executing the following steps: Collect multimodal data related to carbon pollution in the port, and perform outlier cleaning, time series alignment and spatial distribution mapping to obtain a standardized data set; establish a first-layer dynamic game model based on the interactive relationship between the port and the management platform obtained from the standardized data set, and establish a second-layer dynamic game model based on the interactive relationship between various subjects in the port obtained from the standardized data set, wherein the management platform is used to manage the port; generate nodes for the pollution sources, emission reduction technologies and policy measures determined from the standardized data set, and generate edges for the nodes according to the association relationship between the nodes, so as to construct a dynamic knowledge graph, wherein the attribute values ​​of the nodes in the dynamic knowledge graph are updated in real time according to the status of the pollution sources in the port, and the weight values ​​of the edges are updated in real time according to the changes in the propagation path; define the state space, action space and reward function for the reinforcement learning model; in the process of collaboratively processing the carbon pollution in the port, use the first-layer dynamic game model and the second-layer dynamic game model to determine the current emission reduction strategy, use the reinforcement learning model to evaluate the dynamic emission reduction effect of the current emission reduction strategy, and use the dynamic knowledge graph to determine the pollution sources and pollution propagation paths.

[0089] Optionally, the specific examples in this embodiment may refer to the examples described in the above embodiments, and this embodiment will not be described in detail here.

[0090] Optionally, in this embodiment, the above-mentioned storage medium may include but is not limited to: a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk, and other media that can store program codes.

[0091] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0092] If the integrated units in the above embodiments are implemented in the form of software functional units and sold or used as independent products, they can be stored in the above computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for enabling one or more computer devices (which may be personal computers, servers or network devices, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application.

[0093] In the above embodiments of the present application, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0094] In the several embodiments provided in the present application, it should be understood that the disclosed client can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0095] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0096] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0097] The above is only a preferred implementation of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A collaborative treatment method for port carbon pollution, characterized in that: include: Collect multimodal data related to carbon pollution in the port, clean up outliers, align time series, and map spatial distribution to obtain a standardized data set; A first-layer dynamic game model is established based on the interactive relationship between the port and the management platform obtained from the standardized data set, and a second-layer dynamic game model is established based on the interactive relationship between various entities in the port obtained from the standardized data set, wherein the management platform is used to manage the port; Generate nodes for the pollution sources, emission reduction technologies and policy measures determined from the standardized data set, and generate edges for the nodes according to the association relationship between the nodes, so as to construct a dynamic knowledge graph, wherein the attribute values ​​of the nodes in the dynamic knowledge graph are updated in real time according to the status of the pollution sources in the port, and the weight values ​​of the edges are updated in real time according to the changes in the propagation path; Define the state space, action space, and reward function for the reinforcement learning model; In the process of collaborative treatment of port carbon pollution, the first-layer dynamic game model and the second-layer dynamic game model are used to determine the current emission reduction strategy, the reinforcement learning model is used to evaluate the dynamic emission reduction effect of the current emission reduction strategy, and the dynamic knowledge graph is used to determine the pollution source and pollution propagation path.

2. The method according to claim 1, characterized in that The first-level dynamic game model is established based on the interactive relationship between the port and the management platform obtained from the standardized data set, including: Construct the payment function of the port: U port (t) = R profit (t)−C tech (s port ,t)+S subsidy (s gov ,t)−P penalty (s gov ,t) , in, R profit (t) represents the economic benefits of the port, C tech (s port ,t) represents the investment and operating costs of emission reduction technologies, S subsidy (s gov ,t) Indicates subsidies provided by the government. P penalty (s gov ,t) It means the fine for exceeding the emission limit; Construct the payment function of the management platform government: U gov (t) = B env (t)−S subsidy (s gov ,t)+P penalty (s gov ,t)− C admin( t) , in, B env (t) represents the environmental benefits of reduced pollutant and carbon emissions, C admin (t) Represents the administrative cost of policy implementation.

3. The method according to claim 1, characterized in that The second-level dynamic game model is established based on the interactive relationship between various entities in the port obtained from the standardized data set, including: Construct a payoff function for the entities in the port: π i (s i ,t)=π i base(t)−γ×E pollution (t) , in, π i (s i ,t) Representing the subject i of income, E pollution (t) represents the impact value of pollution diffusion, γ It represents the weight coefficient of pollution diffusion. The game players in the port include port enterprises, logistics companies and third-party governance agencies.

4. The method according to claim 1, characterized in that: The attribute values ​​of the nodes in the dynamic knowledge graph are updated in real time according to the following formula: Attr node (t+1)=Attr node (t)+δ×ΔAttr(t) ,in, Attr node (t) Indicates that the node is at time t , Attr node (t+1) Indicates that the node is at time t+1 , δ is the update step size, ΔAttr(t) Indicates the attribute change value; The weight values ​​of the edges in the dynamic knowledge graph are updated in real time according to the following formula: w ij (t+1)=w ij (t)+η×Δw ij (t) ,in, w ij (t) Indicates the edge at time t The weight value when w ij (t+1) Indicates the edge at time t+1 The weight value when η is the weight adjustment coefficient, Δw ij (t) Indicates the edge weight change value.

5. The method according to claim 1, characterized in that Define the state space, action space, and reward function for the reinforcement learning model, including: Define a state space including the following state variables: pollutant emissions, carbon emissions, equipment operating status, and policy parameters; Define an action space that includes the following decision actions: selection of abatement technology, adjustment of policy incentives, and adjustment of equipment operating parameters; Define the reward function: R(s,a)=w 1 ×ΔE pollution (s,a)+w 2× ΔE carbon (s,a)−w 3 ×C total (s,a) ,in, ΔE pollution (s,a) Indicates the current status s and actions a The amount of pollutant reduction achieved, ΔE carbon (s,a) Indicates the current status s and actions a The amount of carbon emissions reduction achieved, C total (s,a) Indicates the current status of execution s and actions a The total cost consumed, w1, w2, w3 represents the corresponding weight coefficient; definition Q Value update formula: ， in, Q(s,a) Indicates status s and actions a The value of α Indicates control Q The learning rate of the value update speed, γ represents the discount factor that determines the importance of future rewards, λ represents the penalty factor for pollution spread, max a′ Q(s',a') represents the optimal emission reduction strategy in the future Q Value, indicating that in the state s′ Choose the best action a′ The maximum benefit that can be obtained, E pollution (s) Indicates the current status s The cumulative pollution level brought by itself.

6. The method according to claim 4, characterized in that The shortest path algorithm based on the correction of pollution propagation effect is used to calculate the optimal path for pollution propagation. : 𝑚𝑖𝑛 in, w ij To indicate the pollutant at the node i and nodes j The pollution transmission weight of the transmission intensity between E ij (t) To represent pollutants from nodes i Propagate to nodes j Diffusion effect parameters.

7. A coordinated treatment device for port carbon pollution, characterized in that: include: The collection unit is used to collect multimodal data related to carbon pollution in the port, and perform outlier cleaning, time series alignment and spatial distribution mapping to obtain a standardized data set; A game model creation unit, used to establish a first-layer dynamic game model according to the interactive relationship between the port and the management platform obtained from the standardized data set, and to establish a second-layer dynamic game model according to the interactive relationship between various entities in the port obtained from the standardized data set, wherein the management platform is used to manage the port; A knowledge graph construction unit, used to generate nodes for pollution sources, emission reduction technologies and policy measures determined from the standardized data set, and to generate edges for the nodes according to the association relationship between the nodes, so as to construct a dynamic knowledge graph, wherein the attribute values ​​of the nodes in the dynamic knowledge graph are updated in real time according to the status of the pollution sources in the port, and the weight values ​​of the edges are updated in real time according to the changes in the propagation path; Reinforcement learning model building unit, used to define the state space, action space and reward function for the reinforcement learning model; The collaborative processing unit is used to determine the current emission reduction strategy using the first-layer dynamic game model and the second-layer dynamic game model during the collaborative processing of port carbon pollution, evaluate the dynamic emission reduction effect of the current emission reduction strategy using the reinforcement learning model, and determine the pollution source and pollution propagation path using the dynamic knowledge graph.

8. A coordinated treatment system for port carbon pollution, characterized in that: include: Collection equipment is used to collect multimodal data related to carbon pollution in the port, and perform outlier cleaning, time series alignment, and spatial distribution mapping to obtain a standardized data set; A multi-level dynamic game model, comprising a first-level dynamic game model established based on the interactive relationship between the port and the management platform obtained from the standardized data set and a second-level dynamic game model established based on the interactive relationship between various entities in the port obtained from the standardized data set, wherein the management platform is used to manage the port; A dynamic knowledge graph, wherein the nodes in the dynamic knowledge graph are used to represent the pollution sources, emission reduction technologies and policy measures determined from the standardized data set, and the edges in the dynamic knowledge graph are used to represent the association relationship between the nodes, wherein the attribute values ​​of the nodes in the dynamic knowledge graph are updated in real time according to the status of the pollution sources in the port, and the weight values ​​of the edges are updated in real time according to the changes in the propagation path; Reinforcement learning model, including predefined state space, action space and reward function; Among them, in the process of collaborative treatment of port carbon pollution, the first-layer dynamic game model and the second-layer dynamic game model are used to determine the current emission reduction strategy, the reinforcement learning model is used to evaluate the dynamic emission reduction effect of the current emission reduction strategy, and the dynamic knowledge graph is used to determine the pollution source and pollution propagation path.

9. A computer-readable storage medium, characterized in that: The storage medium includes a stored program, wherein the program executes the method described in any one of claims 1 to 6 when executed.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: The processor executes the method described in any one of claims 1 to 6 through the computer program.

Images