Multi-agent collaborative data exchange dynamic routing optimization method and system

Abstract

The invention discloses a multi-agent collaborative data exchange dynamic routing optimization method and system, and relates to the technical field of network communication, and the method comprises the steps: obtaining the real-time state information of each node in a network, and outputting a node state data set; constructing a historical state sequence of each node, calculating a load change trend, and outputting a routing adjustment signal when the load change trend reaches a load critical value; calculating a collaborative weight according to the network contribution degree of each agent, establishing a task allocation mapping relationship among the agents based on the collaborative weight, and outputting a collaborative scheduling result; determining a candidate path set, performing index evaluation on paths in the candidate path set, and outputting an optimal routing path; and executing data transmission, obtaining a quality difference between an actual transmission effect and an expected effect, and performing parameter correction. According to the invention, active optimization and multi-target balance of network routing are realized, and routing efficiency and system stability in a dynamic network environment are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of network communication, in particular to a multi-agent collaborative data exchange dynamic routing optimization method and system. BACKGROUND

[0002] The application of multi-agent systems in network routing optimization is becoming increasingly mature, and existing research mainly focuses on distributed routing algorithms, load balancing mechanisms and collaborative control strategies. Current multi-agent routing systems generally use reinforcement learning-based methods to achieve path selection and load distribution through Q-learning, deep neural networks and other technologies. At the same time, the software-defined network (SDN) architecture provides a new technical framework for centralized routing control, making it possible to obtain global network state and path planning. At the algorithm level, swarm intelligence algorithms such as ant colony optimization and particle swarm optimization are widely used in multi-objective routing optimization problems, which lay a theoretical foundation for collaborative decision-making among agents.

[0003] However, existing multi-agent routing technologies still have significant shortcomings in dynamic network environments. First, most existing solutions use a passive response mechanism that only initiates routing adjustments after network congestion or failure occurs, resulting in system response lag and inability to effectively prevent network performance degradation. Second, the collaborative mechanism between agents lacks an effective contribution evaluation system, often using simple polling or random allocation strategies, which fails to fully leverage the network resource advantages of each agent and results in low collaborative efficiency. Third, existing hierarchical architectures often lack organic connections between levels, with global planning and local execution disjointed, making it difficult to achieve true multi-level collaborative optimization. In addition, traditional routing algorithms focus on single-objective optimization, such as shortest path or minimum latency, lacking comprehensive consideration of transmission quality, energy consumption control and load balancing, and unable to meet the multi-dimensional performance requirements in complex network environments. SUMMARY

[0004] In view of the above problems, the present application is proposed.

[0005] Therefore, the present application provides a multi-agent collaborative data exchange dynamic routing optimization method and system, which can solve the problems mentioned in the background art.

[0006] To solve the above technical problems, the present application provides the following technical solutions: In a first aspect, the application provides a multi-agent cooperative data exchange dynamic routing optimization method, comprising: acquiring real-time state information of each node in the network, and outputting a node state data set; constructing a historical state sequence of each node based on the node state data set, calculating a load change trend of each node through sequence analysis, and outputting a routing adjustment signal when the load change trend reaches a load critical value; in response to the routing adjustment signal, calculating a cooperation weight according to the network contribution degree of each agent, establishing a task allocation mapping relationship between agents based on the cooperation weight, and outputting a cooperative scheduling result; determining a candidate path set through a hierarchical processing architecture according to the cooperative scheduling result, evaluating the paths in the candidate path set according to indicators, and outputting an optimal routing path; performing data transmission based on the optimal routing path, acquiring a quality difference between actual transmission effect and expected effect, and performing parameter correction on routing decision according to the quality difference.

[0007] Preferably, the historical state sequence of each node is constructed based on the node state data set, and the load change trend of each node is calculated through sequence analysis, comprising: extracting state data of each node within a preset time period, and arranging the state data in chronological order to form the historical state sequence; performing change analysis on the historical state sequence to identify the regular characteristics of load change; calculating a load prediction value at a target time based on the regular characteristics; comparing the load prediction value with the load critical value, and generating the routing adjustment signal when the load prediction value exceeds the load critical value.

[0008] Preferably, the cooperation weight of each agent is calculated according to the network contribution degree of each agent, comprising: counting the number of path resources provided by each agent to the network to determine a path contribution value; calculating the corresponding relationship between the network load shared by each agent and the system average load to determine a load balancing contribution value; evaluating the improvement amplitude of system effect after each agent joins the cooperation to determine an effect optimization contribution value; and comprehensively operating the path contribution value, the load balancing contribution value and the effect optimization contribution value to obtain the cooperation weight of each agent.

[0009] Preferably, the candidate path set is determined through a hierarchical processing architecture, comprising: establishing a path basic framework based on the network topology structure at a global planning layer to form an initial path set; updating the initial path set by each agent exchanging local information based on the node state data set at a local coordination layer; selecting paths meeting the quality requirements from the updated path set through a quality index screening mechanism at an execution layer to constitute the candidate path set.

[0010] Preferably, the index evaluation on the paths in the candidate path set comprises: obtaining the transmission delay index, the energy consumption index and the load balancing index of each candidate path; determining the calculation coefficients of each index according to the network application requirements; comprehensively calculating each index value and the corresponding calculation coefficient to obtain the comprehensive quality score of each candidate path; and selecting the path with the highest comprehensive quality score as the optimal routing path.

[0011] Preferably, the parameter correction on the routing decision according to the quality difference comprises: calculating the deviation of the actual transmission delay from the expected transmission delay to obtain a first quality deviation; calculating the deviation of the actual energy consumption from the expected energy consumption to obtain a second quality deviation; combining the first quality deviation with a first correction factor and combining the second quality deviation with a second correction factor to generate a parameter adjustment value; and applying the parameter adjustment value to the reference value of the corresponding parameter in the routing decision process.

[0012] Preferably, the length of the preset time period is set according to the network state fluctuation frequency, the length of the preset time period is reduced when the network state fluctuation frequency is higher than a fluctuation threshold, and the length of the preset time period is increased when the network state fluctuation frequency is lower than the fluctuation threshold.

[0013] In a second aspect, the application further provides a multi-agent cooperative data exchange dynamic routing optimization system, comprising: a state acquisition module configured to acquire real-time state information of each node in a network and output a node state data set; an analysis and early warning module configured to construct a historical state sequence of each node based on the node state data set, calculate a load change trend of each node through sequence analysis, and output a routing adjustment signal when the load change trend reaches a load critical value; a cooperative calculation module configured to respond to the routing adjustment signal, calculate a cooperative weight according to the network contribution degree of each agent, establish a task allocation mapping relationship between agents based on the cooperative weight, and output a cooperative scheduling result; a hierarchical decision module configured to determine a candidate path set through a hierarchical processing architecture according to the cooperative scheduling result, perform index evaluation on the paths in the candidate path set, and output an optimal routing path; and a parameter correction module configured to perform data transmission based on the optimal routing path, acquire a quality difference between an actual transmission effect and an expected effect, and perform parameter correction on a routing decision according to the quality difference.

[0014] In a third aspect, the present application also provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the following steps when executing the computer program: obtaining real-time state information of each node in a network, and outputting a node state data set; constructing a historical state sequence of each node based on the node state data set, calculating a load change trend of each node through sequence analysis, and outputting a routing adjustment signal when the load change trend reaches a load critical value; in response to the routing adjustment signal, calculating a collaboration weight according to a network contribution degree of each agent, establishing a task allocation mapping relationship between agents based on the collaboration weight, and outputting a collaborative scheduling result; determining a candidate path set through a hierarchical processing architecture according to the collaborative scheduling result, performing index evaluation on paths in the candidate path set, and outputting an optimal routing path; performing data transmission based on the optimal routing path, obtaining a quality difference between an actual transmission effect and an expected effect, and performing parameter correction on a routing decision according to the quality difference.

[0015] In a fourth aspect, the present application also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program is executed by a processor to implement the following steps: obtaining real-time state information of each node in a network, and outputting a node state data set; constructing a historical state sequence of each node based on the node state data set, calculating a load change trend of each node through sequence analysis, and outputting a routing adjustment signal when the load change trend reaches a load critical value; in response to the routing adjustment signal, calculating a collaboration weight according to a network contribution degree of each agent, establishing a task allocation mapping relationship between agents based on the collaboration weight, and outputting a collaborative scheduling result; determining a candidate path set through a hierarchical processing architecture according to the collaborative scheduling result, performing index evaluation on paths in the candidate path set, and outputting an optimal routing path; performing data transmission based on the optimal routing path, obtaining a quality difference between an actual transmission effect and an expected effect, and performing parameter correction on a routing decision according to the quality difference.

[0016] The present application has the following beneficial effects: the present application provides a multi-agent collaborative data exchange dynamic routing optimization method and system, when the historical state sequence of each node in the network is constructed, the load change trend can be calculated through sequence analysis, and then a routing adjustment signal can be output before load abnormality occurs, so that the routing adjustment can be changed from passive response to active prevention, and the response can be timely even under the condition of rapid fluctuation of network load. When the network contribution degree of each agent is quantified, the collaboration weight can be calculated and the task allocation mapping relationship can be established, so as to avoid the low efficiency problem of traditional polling or random allocation, and then the rational allocation of agent resources can be realized, and the collaboration efficiency between agents can be greatly improved.

[0017] When the global planning layer, the local coordination layer and the execution layer in the hierarchical processing architecture work together, unified decision can be realized through information transmission between layers, so as to solve the problem that global planning and local execution are disconnected in the prior art, and then make the path selection have both global optimization vision and local execution efficiency. When the actual transmission effect and the expected effect have quality difference, the routing decision parameters can be adjusted through the parameter correction mechanism, so as to form a complete feedback optimization loop, and then continuous monitoring and improvement can be realized, and comprehensive optimization can be realized in multiple dimensions such as transmission delay, energy consumption and load balancing.

[0018] In summary, the present application not only realizes predictive analysis of network state, but also establishes an efficient coordination mechanism between agents, which guarantees the global nature of routing decision and ensures the local adaptability of the execution process. BRIEF DESCRIPTION OF DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related art, the drawings needed to be used in the description of the embodiments of the present application or the related art will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other related drawings can be obtained by those skilled in the art without creative labor.

[0020] Figure 1 is a whole flowchart of a multi-agent cooperative data exchange dynamic routing optimization method involved in the present application; Figure 2 is an application environment diagram of a multi-agent cooperative data exchange dynamic routing optimization method involved in the present application; Figure 3 is a whole structure schematic diagram of a multi-agent cooperative data exchange dynamic routing optimization system involved in the present application; Figure 4 is a computer device diagram of a multi-agent cooperative data exchange dynamic routing optimization method involved in the present application. DETAILED DESCRIPTION

[0021] In order to make the purpose, technical scheme and advantages of the present application more clear, the present application will be further described in detail below in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, and are not used to limit the present application.

[0022] The agent in the present application refers to a software program module deployed on a network node, and one agent is deployed on each network node participating in routing optimization. The agent has autonomous decision-making and cooperative communication capabilities, can perceive the state information of the node, exchange information with other agents, and execute a cooperative routing algorithm. The agent establishes a connection with other agents through a preset communication protocol to form a distributed cooperative network. Therefore, the network contribution degree of the agent in the present application is essentially the network contribution degree of the network node on which the agent is deployed, and the task allocation between agents is essentially the task allocation between network nodes.

[0023] The data exchange dynamic routing optimization technology based on multi-agent cooperation is widely applied in modern network communication. For example, it can be applied to intelligent transportation systems, industrial internet, data center networks, edge computing networks and various fields. When there are a large number of data exchange requirements or dynamic load changes in the network, the multi-agent cooperative routing system can achieve efficient path planning and load distribution in a complex network environment.

[0024] In related technologies, the traditional network routing system selects a path based on a static routing table or a simple load balancing algorithm, and outputs a relatively fixed routing scheme. However, in a dynamic network environment, the network load, node state and topology structure change frequently, and it is a relatively difficult problem to enable the routing system to have predictive analysis and cooperative optimization capabilities.

[0025] In view of the problem that the related technologies cannot effectively predict network state changes and achieve efficient cooperation between agents, a multi-agent cooperative data exchange dynamic routing optimization method is proposed. The historical state sequence is constructed according to the real-time state information of each node, and the load change trend is calculated. The cooperative weight and task allocation mapping relationship are established based on the network contribution degree of each agent. The candidate path set is determined through a hierarchical processing architecture and index evaluation, and the routing decision is parameter corrected according to the quality difference between the actual transmission effect and the expected effect.

[0026] The multi-agent cooperative data exchange dynamic routing optimization method provided by the embodiments of the present application can be applied to, for example, Figure 2The application environment shown. Among them, the network node 102 is connected with the control center 104 through the communication link. The distributed database can store the network state data and routing information required by the control center 104 to process. The distributed database can be integrated on the control center 104, or deployed on the cloud or other distributed servers. Each network node obtains real-time state information, outputs a node state data set; based on the node state data set, the historical state sequence of each node is constructed, the load change trend of each node is calculated through sequence analysis, and the routing adjustment signal is output when the load change trend reaches the load critical value; in response to the routing adjustment signal, the collaborative weight is calculated according to the network contribution of each intelligent agent, the task allocation mapping relationship between intelligent agents is established based on the collaborative weight, and the collaborative scheduling result is output; according to the collaborative scheduling result, the candidate path set is determined through the hierarchical processing architecture, the paths in the candidate path set are evaluated, and the optimal routing path is output; based on the optimal routing path, the data transmission is executed, the quality difference between the actual transmission effect and the expected effect is obtained, and the routing decision is parameter modified according to the quality difference.

[0027] Among them, the network node 102 can be but not limited to various routers, switches, base stations, edge computing devices and Internet of Things gateway devices, and the Internet of Things gateway device can be an industrial gateway, a smart home gateway, a vehicle Internet of Things gateway, a sensor gateway, etc. The edge computing device can be an edge server, an MEC device, a fog computing node, etc. The base station device can be a 5G base station, a 4G base station, a Wi-Fi access point, a LoRa gateway, etc. The control center 104 can be an independent network control server, or a distributed control system composed of multiple control servers, or a cloud control platform providing software defined network service.

[0028] In an exemplary embodiment, as Figure 2 shown, a multi-agent collaborative data exchange dynamic routing optimization method is provided, which is applied to the application environment in Figure 1 for example, including the following steps 202 to 210. Among them: Step 202: Obtain the real-time state information of each node in the network, and output the node state data set.

[0029] Among them, the real-time state information includes the current load rate of the node, the link bandwidth usage, the processor utilization, the cache occupancy, the number of network connections, and the data packet transmission delay, etc. The node state data set provides basic data support for subsequent load prediction and routing optimization.

[0030] The network node can be a router, a switch, a server, a base station or other network devices with data forwarding capability. Each node has a state monitoring function and can collect its own running parameters and network performance indicators in real time.

[0031] Exemplarily, the running state data of the nodes are periodically collected through the state monitoring agents deployed on the nodes. The monitoring agent can be a lightweight software module that samples the node state at a preset time interval (such as 1 second, 5 seconds or 10 seconds). The collected state information is standardized to form a state data record in a unified format.

[0032] The storage format of the node state data set can be represented as: {node identification node_id, timestamp, load rate load_rate, bandwidth usage bandwidth_usage, processor utilization cpu_usage, cache occupancy memory_usage, connection count connection_count, transmission delay delay}. Among them, the node identification is used to uniquely identify a single node in the network, the timestamp records the specific time of data collection, and each performance indicator represents the current running state of the node in numerical form.

[0033] In actual implementation, the collection of state information can be achieved through the Simple Network Management Protocol (SNMP), the API interface of network equipment or a dedicated monitoring protocol. For different types of network equipment, the collection method may differ, but ultimately the collected raw data needs to be converted into a standardized node state data format to ensure data consistency and processability.

[0034] The collected node state data is organized in time sequence to form a data set containing the state information of all network nodes. This data set not only contains the state snapshot at the current time, but also retains the historical state records within a certain time window, providing necessary data basis for subsequent trend analysis and prediction calculation.

[0035] Step 204: Based on the node state data set, the historical state sequence of each node is constructed, the load change trend of each node is calculated through sequence analysis, and the routing adjustment signal is output when the load change trend reaches the load critical value.

[0036] Among them, the historical state sequence is a combination of node state data recorded in chronological order, containing the load rate, bandwidth usage and other state information of the node at consecutive time points. The load change trend is determined by calculating the change rate and direction of the load rate in the historical state sequence, reflecting the growth or decline trend of the node load. The load critical value is a load threshold preset according to the processing capacity of the node, when the predicted load exceeds this threshold, the routing adjustment mechanism is started.

[0037] Exemplarily, the state records of the same node at different time instants are extracted from the node state data set, and are arranged in ascending order of time stamp to form a historical state sequence of the node. By calculating the difference between the load rates of adjacent time points in the sequence, load change rate data are obtained, and the change direction and speed of the load are analyzed. Based on the change rate data and historical load values, a load prediction value at the target time instant is calculated and compared with the load threshold value.

[0038] Further, in step 204, the historical state sequence of each node is constructed based on the node state data set, and the load change trend of each node is calculated through sequence analysis, including steps A1 to A4: Step A1, extract the state data of each node within a preset time period, and arrange the historical state sequence in chronological order.

[0039] The preset time period is a fixed time window for trend analysis, for example, the last 60 seconds, 120 seconds or 300 seconds. The length of the time period affects the amount of data analyzed and the accuracy of the prediction, and needs to be selected according to the stability of the network environment.

[0040] Exemplarily, all state records of a single node within a preset time period are selected from the node state data set. The selected records are arranged in ascending order of time stamp from early to late to ensure the time continuity of the data. The arranged data constitute the historical state sequence of the node, each data point in the sequence containing a time stamp and a corresponding load rate value. When there is missing data in the historical state sequence, the missing data points are supplemented by linear interpolation method, and the load rate value of the missing time point is calculated according to the time interval ratio using the two valid data points before and after the missing time point.

[0041] The data organization form of the historical state sequence is that the node identifier corresponds to a time sequence array, and each element in the array contains two fields of time stamp and load rate. By traversing all records in the node state data set, the records are grouped according to the node identifier, and an independent historical state sequence is constructed for each node.

[0042] Step A2, change analysis is performed on the historical state sequence to identify the regular characteristics of load change.

[0043] The change analysis is realized by calculating the change characteristics of the load rate in the time dimension, including the judgment of the change direction and the calculation of the change rate. The regular characteristics include whether the load shows an upward trend, a downward trend or remains stable, and the speed of the change.

[0044] Exemplarily, the difference between the load rates of each adjacent two time points in the historical state sequence is calculated to obtain a load change rate sequence. The sum of the values in the change rate sequence is calculated. If the sum is positive, it indicates that the load overall shows an upward trend; if the sum is negative, it indicates that the load overall shows a downward trend; and if the sum is close to zero, it indicates that the load remains relatively stable.

[0045] Further, the linear trend of the load change is determined by performing linear fitting on the historical state sequence. The linear fitting calculates the slope and intercept of the fitting straight line by the least square method, wherein the slope value reflects the rate of the load change, the positive slope indicates the upward trend, the negative slope indicates the downward trend, and the absolute value of the slope reflects the speed of the change.

[0046] Step A3, calculating the load prediction value of the target time point based on the regularity feature.

[0047] The target time point is usually set as the time point after the next data collection period or a plurality of periods. The load prediction value is calculated by continuing the change regularity identified in step A2 to the target time point.

[0048] Exemplarily, the linear fitting parameters obtained in step A2 are used to calculate the load prediction value of the target time point by substituting the time value of the target time point into the fitting straight line equation. The calculation method is: multiplying the target time point by the slope of the fitting straight line, and then adding the intercept value of the fitting straight line to obtain the predicted load rate value. The linear fitting needs to contain at least 3 data points in the historical state sequence. When the data points are less than 3, the load rate of the latest time point is used as the prediction value.

[0049] To enhance the reliability of the prediction, the results of multiple prediction methods are comprehensively judged. In addition to the linear fitting, a simple prediction value based on the load change rate of the latest time point is also calculated. The simple prediction value is obtained by calculating the average value of the load change rates of the latest 3 time points, and then adding the average value to the current load rate. The linear fitting prediction value and the simple prediction value are weighted and averaged according to the weight ratio of 7:3 to obtain the final load prediction value.

[0050] Step A4, comparing the load prediction value with the load critical value. When the load prediction value exceeds the load critical value, a routing adjustment signal is generated.

[0051] The load critical value is determined according to the maximum processing capacity and the quality of service of the node, and is usually set to 70% to 90% of the maximum load capacity of the node. When the predicted load exceeds the critical value, it indicates that the node will face the problem of insufficient processing capacity, and part of the data flow needs to be distributed to other nodes through routing adjustment.

[0052] Exemplarily, the load prediction value calculated in step A3 is compared with the load threshold value. If the load prediction value is greater than the load threshold value, a routing adjustment signal containing the node identifier, the current load, the predicted load, the threshold value and the exceeding degree is created. The exceeding degree is represented by calculating the difference between the predicted load and the threshold value as a percentage of the threshold value.

[0053] The routing adjustment signal contains the signal type identifier, the node identifier where the threshold is exceeded, the current load rate of the node, the predicted load rate, the preset threshold value, the exceeding degree percentage and the timestamp of signal generation. The signal will be passed to the collaborative calculation module as a trigger condition for starting the agent collaborative scheduling.

[0054] It should be noted that the length of the preset time period is set according to the network state fluctuation frequency. The network state fluctuation frequency is quantified by calculating the standard deviation of the node state data within a fixed time period. A large standard deviation value indicates that the network state changes frequently, and a small standard deviation value indicates that the network state is relatively stable.

[0055] Specifically, when the network state fluctuation frequency is higher than the preset fluctuation threshold, the length of the preset time period is set to a smaller value, for example, from the default 120 seconds to 60 seconds, to improve the response ability to rapid changes in network state. When the network state fluctuation frequency is lower than the preset fluctuation threshold, the length of the preset time period is set to a larger value, for example, from the default 120 seconds to 240 seconds, to obtain more sufficient historical data to support prediction analysis. The preset fluctuation threshold is set to 1.5 times the standard deviation under normal network operation state, and the adjustment of the time period length ensures that the effective data points contained are not less than 5 and not more than 20.

[0056] By adjusting the length of the preset time period according to the network state fluctuation frequency, the analysis time window is matched with the network dynamic characteristics, ensuring the accuracy of the prediction while considering the system response speed, thereby improving the adaptability and effectiveness of the load prediction mechanism.

[0057] Step 206: In response to the routing adjustment signal, calculate the collaborative weight according to the network contribution degree of each agent, establish the task allocation mapping relationship between the agents based on the collaborative weight, and output the collaborative scheduling result.

[0058] The network contribution degree is a comprehensive evaluation index calculated by weighted summation of the path contribution value, the load balancing contribution value and the effect optimization contribution value. The collaborative weight is obtained by standardizing the network contribution degree, with a value range of 0 to 1. The larger the value, the stronger the collaborative ability of the agent. The task allocation mapping relationship establishes the correspondence between the agent identifier and the task, and clearly defines the specific responsibilities of each agent in the collaborative process.

[0059] Exemplarily, after receiving the routing adjustment signal output in step 204, the node identifier and load warning information contained in the signal are identified. Based on the historical performance data and current state information of each agent in the network, the contribution values of each agent in the path providing, load sharing and effect improving dimensions are calculated respectively. The contribution values in the three dimensions are weighted and summed according to the preset weight to obtain the network contribution degree of each agent, and then the network contribution degree is standardized by dividing the maximum value of all agent contribution degrees to obtain the cooperation weight value. According to the size of the cooperation weight value and the specific requirements of the routing adjustment, the corresponding scheduling task is allocated to each agent.

[0060] Further, the cooperation weight value is calculated according to the network contribution degree of each agent in step 206, including steps B1 to B4: Step B1, the number of path resources provided by each agent to the network is counted to determine the path contribution value.

[0061] Wherein, the path resource quantity refers to the number of available paths that the agent can provide for data transmission, including the main path and the standby path. The path contribution value is calculated by dividing the number of available paths provided by the agent by the total number of paths provided by all agents in the network, and the value range is between 0 and 1.

[0062] Exemplarily, the network topology structure is traversed to identify the number of adjacent nodes connected to each agent node and the link state. The total number of reachable paths from the agent node to other nodes in the network is counted, including direct connection paths and multi-hop paths. The calculation of reachable paths uses the path search algorithm in graph theory, and the search depth is limited within 3 hops to ensure the practicability of the path.

[0063] The calculation method of the path contribution value is: dividing the number of available paths provided by the agent by the total number of paths provided by all agents in the network to obtain the path contribution proportion of the agent. When the agent node fails or the link is interrupted, the number of available paths of the agent decreases accordingly, and the path contribution value is recalculated.

[0064] Step B2, the corresponding relationship between the network load shared by each agent and the system average load is calculated to determine the load balancing contribution value.

[0065] Wherein, the network load sharing refers to the size of the data flow borne by the agent in the data forwarding process, including the number of forwarded data packets and the occupied bandwidth resources. The system average load is the arithmetic mean of the load borne by all agents. The load balancing contribution value is determined by calculating the deviation degree of the actual load of the agent from the system average load, and the smaller the deviation, the greater the contribution value.

[0066] Exemplarily, load statistics of each agent in a recent time period are collected, including the number of processed data packets, occupied bandwidth, CPU utilization and other indicators. The average value of the load of all agents in the system is calculated as the baseline load. The actual load of each agent is compared with the baseline load, and the load deviation value is calculated.

[0067] The specific calculation method of the load balancing contribution value is: the baseline load divided by the absolute value of the difference between the actual load of the agent and the baseline load plus 1, and when the difference is zero, the contribution value is the baseline load. This calculation method ensures that agents with loads close to the average level obtain a higher contribution score.

[0068] Step B3, evaluate the improvement range of system effect after each agent joins the cooperation to determine the effect optimization contribution value.

[0069] Among them, the system effect includes network transmission delay, data packet loss rate, throughput and other performance indicators. The improvement range is measured by comparing the system performance before and after the agent participates in the cooperation. The effect optimization contribution value is calculated according to the weighted comprehensive score of the performance improvement.

[0070] Exemplarily, the system performance baseline data before the agent joins the cooperative scheduling is recorded, including average transmission delay, data packet loss rate and network throughput and other key indicators. After the agent participates in the cooperative scheduling, the same performance indicators are measured again, and the change range of each indicator is calculated. The greater the performance improvement, the more significant the effect optimization contribution of the agent.

[0071] The calculation of the effect optimization contribution value adopts the weighted comprehensive score method, and the transmission delay improvement, loss rate reduction and throughput improvement are weighted according to the weight ratio of 4:3:3. When the participation of a certain agent leads to the decline of system performance, the effect optimization contribution value of the agent is set to zero, avoiding the negative contribution to the overall evaluation.

[0072] Step B4, the path contribution value, the load balancing contribution value and the effect optimization contribution value are comprehensively calculated to obtain the cooperative weight value of each agent.

[0073] Among them, the cooperative weight value is calculated by weighting and summing the path contribution value, the load balancing contribution value and the effect optimization contribution value according to the preset weight coefficient, and then standardized to make the final result value range between 0 and 1.

[0074] Exemplarily, the three contribution values obtained in steps B1, B2 and B3 are weighted and calculated according to a weight ratio of 3:4:3 to obtain the original network contribution value of each agent. The maximum value in the original network contribution values of all agents is taken as a standardization benchmark, and the coordination weight of each agent is equal to the original network contribution value divided by the standardization benchmark. The setting of the weight ratio is based on the priority of network performance, and the highest contribution weight of load balancing reflects the importance of load sharing in coordination.

[0075] After the coordination weight is calculated, the agents are sorted according to the weight value, and the agents with higher weight values are given priority to participate in the cooperative scheduling task. When the weights of multiple agents are similar, the agent with a network position closer to the load abnormal node is selected to reduce the hop count and delay of data transmission.

[0076] Based on the coordination weight, a task allocation mapping relationship between agents is established, and the routing adjustment task is decomposed into subtasks such as path search, load sharing and performance monitoring. According to the coordination weight and capability characteristics of each agent, the corresponding subtasks are allocated. The cooperative scheduling result includes the list of agents participating in the cooperation, the specific tasks undertaken by each agent, the expected load allocation ratio and the time arrangement of the cooperative execution.

[0077] It should be noted that the calculation period of the coordination weight is synchronized with the collection period of the state information in step 202, so as to ensure that the weight reflects the current state of the agent. When the network topology changes or the state of the agent changes significantly, the coordination weight of each agent is recalculated to maintain the timeliness and accuracy of task allocation. The historical record of the coordination weight is used to analyze the performance trend of the agent and provide data support for long-term network optimization strategy.

[0078] Step 208: According to the cooperative scheduling result, determine the candidate path set through the hierarchical processing architecture, evaluate the paths in the candidate path set according to the indicators, and output the optimal routing path.

[0079] The hierarchical processing architecture is composed of three levels: global planning layer, local coordination layer and execution layer, each level undertakes different path processing tasks. The candidate path set is a combination of paths that meet the quality requirements after three layers of screening processing. The index evaluation is calculated by calculating the values of transmission delay, energy consumption and load balancing in three dimensions, and the comprehensive score is obtained by weighted summation. The optimal routing path is the single path with the highest comprehensive score in the candidate path set.

[0080] The cooperative scheduling result output by the receiving step 206 is exemplarily used to obtain a list of intelligent agents participating in the cooperation and task allocation information. A path search space of a network topology is established at a global planning layer to generate all possible transmission paths. The current node state information is exchanged by intelligent agents at a local coordination layer, and the path availability is updated according to the real-time load condition. The paths are quality-screened at an execution layer, and the paths meeting the basic transmission requirements are retained to form a candidate set. The performance indicators of the candidate paths are calculated one by one and weighted summed, and the path with the highest score is selected as the optimal routing scheme.

[0081] Further, the candidate path set is determined by the hierarchical processing architecture in step 208, including steps C1 to C3: Step C1, a path basic framework is established based on the network topology structure at the global planning layer to form an initial path set.

[0082] The network topology structure contains a graphical representation of all network nodes and their connection relationships. The path basic framework is a set of all theoretically reachable paths from the source node to the target node. The initial path set is generated by searching through a graph theory algorithm and contains all path combinations with a hop count not exceeding a preset upper limit.

[0083] Exemplarily, the source node requiring load distribution and the target node capable of receiving the distribution are determined according to the cooperative scheduling result. A breadth-first search algorithm is used to traverse the network topology, starting from the source node to expand layer by layer, and searching all paths reaching the target node. A maximum hop count limit of 5 hops is set during the search process to avoid a decrease in transmission efficiency caused by excessively long paths. All the searched paths are recorded according to the hop count and node sequence to form the initial path set.

[0084] The path record format is: {path identification path_id, source node source_node, target node target_node, intermediate node sequence middle_nodes, total hop count hop_count}. When there are multiple different paths from the source node to the target node, each path is assigned an independent path identification for differentiation.

[0085] Step C2, the intelligent agents exchange local information according to the node state data set at the local coordination layer to update the initial path set.

[0086] The local information includes the current load rate of the node where the intelligent agent is located, the link bandwidth occupation condition, and the connectivity state of the adjacent nodes. The information exchange is realized through a communication protocol between intelligent agents, and each intelligent agent sends a state query request to other intelligent agents through which the path passes and receives a response. The path set is updated according to the obtained real-time state information, and the unusable paths are deleted and the current quality state of the path is marked.

[0087] Exemplarily, for each path in the initial path set, the agent of the path start node sends a state query message to all intermediate nodes on the path. After receiving the query, the agent of the intermediate node replies the current load rate, available bandwidth and link connectivity state of the node. When the load rate of a certain intermediate node exceeds 90% or the link is interrupted, the path is deleted from the initial path set.

[0088] For the paths remaining in the set, the current quality index of the path is calculated according to the load conditions of the nodes on the path. The path quality index is equal to the arithmetic mean of the load rates of all nodes on the path, which is used for subsequent path optimization. The updated path set contains the topology information and the current quality index of the path.

[0089] Step C3, the execution layer selects paths meeting the quality requirements from the updated path set through a quality index screening mechanism to form a candidate path set.

[0090] The quality index screening mechanism screens the paths based on preset quality thresholds, including constraint conditions such as maximum allowed delay, minimum available bandwidth and maximum load rate. The path meeting the quality requirements is the path that meets all the constraint conditions at the same time. The candidate path set is the path combination remaining after quality screening, which is used for subsequent detailed evaluation.

[0091] Exemplarily, the constraint conditions for quality screening are set as follows: the average load rate of the path does not exceed 80%, the available bandwidth of the path is not less than 100 Mbps, and the estimated transmission delay of the path does not exceed 50 milliseconds. Each path in the updated path set is checked one by one to see if it meets all the constraint conditions. The average load rate of the path is calculated using the result of step C2, the available bandwidth of the path is the minimum value of the available bandwidths of all links on the path, and the estimated transmission delay of the path is calculated by accumulating the hop number and the link delay.

[0092] Only the paths meeting all the three constraint conditions are retained to form the candidate path set. When the number of paths in the candidate path set exceeds 10, the first 10 paths with the best quality are retained according to the path quality index sorting to avoid excessive calculation in subsequent evaluation. The candidate path set will be used as the input for subsequent detailed evaluation.

[0093] Further, the paths in the candidate path set are evaluated in step 208, including steps D1 to D4: Step D1, the transmission delay index, energy consumption index and load balancing index of each candidate path are obtained.

[0094] The transmission delay index is the total time required for data packets to be transmitted from the source node to the target node, including node processing delay and link transmission delay. The energy consumption index is the total energy consumed by all nodes on the path for forwarding data packets. The load balancing index is obtained by calculating the degree of influence of the path on the existing network load distribution, and the smaller the value, the smaller the influence on load balancing.

[0095] For each path in the candidate path set, the transmission delay index is calculated by adding the average processing delay of each node on the path and the transmission delay of each link. The node processing delay is calculated according to the current load rate and processing capacity of the node, and the higher the load rate, the greater the processing delay. The link transmission delay is calculated according to the link length and transmission rate.

[0096] The energy consumption index is calculated by counting the energy consumption required by each node on the path to forward a unit of data, calculating the energy consumption of each node according to the expected amount of data to be transmitted, and then summing up to obtain the total energy consumption of the path. The load balancing index is obtained by calculating the standard deviation of the load rate of each node after the path is used, and the larger the standard deviation, the more uneven the load distribution.

[0097] Step D2, determine the calculation coefficient of each index according to the network application requirement.

[0098] The calculation coefficient is the weight value of each performance index in the comprehensive evaluation, which is determined according to the current network application sensitivity to delay, energy consumption control requirement and load balancing importance. The delay-sensitive application increases the calculation coefficient of the transmission delay index, the energy-sensitive application increases the calculation coefficient of the energy consumption index, and the load-sensitive application increases the calculation coefficient of the load balancing index.

[0099] For example, the calculation coefficient is determined according to the type of data flow currently being processed. For real-time audio and video transmission, the transmission delay index coefficient is set to 0.6, the energy consumption index coefficient is set to 0.2, and the load balancing index coefficient is set to 0.2, highlighting the importance of delay performance. For large file transmission, the transmission delay index coefficient is set to 0.2, the energy consumption index coefficient is set to 0.3, and the load balancing index coefficient is set to 0.5, emphasizing load balancing and energy consumption control.

[0100] The sum of the calculation coefficients is always equal to 1.0, ensuring that the numerical range of the comprehensive evaluation result is stable. When the network application requirement changes, the calculation coefficients are redistributed according to the new application characteristics, keeping the evaluation standard matched with the application requirement.

[0101] Step D3, the index values are integrated with the corresponding calculation coefficients to obtain the comprehensive quality score of each candidate path.

[0102] The comprehensive quality score is obtained by multiplying each performance indicator by its calculation coefficient and summing the results. To ensure the comparability of different indicators, each indicator is standardized before calculation, converting the indicator value to a standardized value between 0 and 1. The comprehensive quality score has a value range of 0 to 1, and the larger the value, the better the comprehensive performance of the path.

[0103] By way of example, the same indicator of all paths in the candidate path set is standardized: the minimum value of the indicator is set to 0, the maximum value is set to 1, and other values are converted proportionally. The transmission delay indicator is inversely standardized, i.e., the minimum delay corresponds to 1 and the maximum delay corresponds to 0. The energy consumption indicator is also inversely standardized. The load balancing indicator is inversely standardized, i.e., the minimum standard deviation corresponds to 1.

[0104] The comprehensive quality score calculation formula is: the standardized transmission delay indicator multiplied by its calculation coefficient, plus the standardized energy consumption indicator multiplied by its calculation coefficient, plus the standardized load balancing indicator multiplied by its calculation coefficient. After calculation, each candidate path obtains a comprehensive quality score between 0 and 1.

[0105] Step D4, selecting the path with the highest comprehensive quality score as the optimal routing path.

[0106] The optimal routing path is the single path with the highest comprehensive quality score in the candidate path set. When there are multiple paths with the same score and the highest score, the path with fewer hops is selected as the optimal path to reduce transmission complexity. After the optimal routing path is determined, the path information is sent to the relevant agents for routing configuration.

[0107] By way of example, the comprehensive quality scores of all paths in the candidate path set are compared numerically to identify the path with the highest score. If path A has a score of 0.85, path B has a score of 0.82, and path C has a score of 0.85, then among paths A and C with the same score, the path with fewer hops is selected. After the optimal routing path is determined, a routing configuration instruction is generated and sent to each agent node on the path.

[0108] The routing configuration instruction includes path identification, source node, target node, complete node forwarding sequence, and expected data traffic size. After receiving the instruction, each agent adds the corresponding forwarding rule to the local routing table, preparing for the upcoming data traffic.

[0109] It should be noted that step 208 ensures that the optimal routing path selected not only meets the basic transmission requirements but also achieves balanced optimization in multiple performance dimensions through layer-by-layer screening and multi-indicator comprehensive evaluation of the hierarchical architecture. The processing time of each level is controlled within milliseconds, ensuring the real-time requirements of routing selection.

[0110] Step 210: Perform data transmission based on the optimal routing path, obtain the quality difference between the actual transmission effect and the expected effect, and make parameter correction to the routing decision according to the quality difference.

[0111] The actual transmission effect is obtained by monitoring the actual time delay and energy consumption during data transmission. The expected effect is the theoretical value of the transmission time delay index and energy consumption index calculated in step D1. The quality difference is obtained by calculating the difference between the actual effect and the expected effect, including the time delay difference and the energy consumption difference. The parameter correction adjusts the prediction model parameters used in the subsequent routing decision according to the size and direction of the quality difference.

[0112] For example, after starting data transmission, the performance indicators are monitored in real time during transmission. The actual total time delay of the data packet from the source node to the target node is recorded, and the actual energy consumption of each node on the path is counted. The actual measured performance data is compared with the expected value calculated in step D1.

[0113] The quality difference calculation method is: actual time delay - expected time delay = time delay difference, actual energy consumption - expected energy consumption = energy consumption difference. When the difference is positive, the actual performance is lower than expected, and when the difference is negative, the actual performance is better than expected. The absolute value of the difference determines the amplitude of the parameter correction, the larger the difference, the larger the correction amplitude.

[0114] The parameter correction specifically includes adjusting the prediction benchmark of node processing time delay and correcting the calculation benchmark of energy consumption. The corrected parameters are used in the next routing decision process, gradually improving the accuracy of routing performance prediction and the effectiveness of routing selection.

[0115] Further, the parameter correction to the routing decision in step 210 according to the quality difference includes steps E1 to E4: Step E1, calculate the deviation of the actual transmission time delay and the expected transmission time delay to obtain the first quality deviation.

[0116] The actual transmission time delay is the measured time of data packet transmission on the optimal routing path, starting from the sending of the data packet by the source node to the successful reception of the data packet by the target node. The expected transmission time delay is the theoretical time delay value calculated in step D1 based on the path node processing capacity and link transmission rate. The first quality deviation is calculated by subtracting the two, a positive value indicating that the actual performance is lower than expected, and a negative value indicating that the actual performance is better than expected.

[0117] Exemplarily, at the beginning of data transmission, the source node records the accurate timestamp of data packet sending. The data packet is forwarded through each intermediate node on the optimal routing path and records the receiving timestamp when reaching the target node. The actual transmission delay = the time difference between the receiving timestamp and the sending timestamp. Numerical comparison is made between the actual transmission delay and the expected transmission delay calculated in step D1.

[0118] The calculation formula of the first quality deviation is: actual transmission delay - expected transmission delay. For example, if the actual transmission delay is 45 milliseconds and the expected transmission delay is 40 milliseconds, then the first quality deviation is 5 milliseconds, indicating that the actual delay is 5 milliseconds higher than the expected delay. If the actual transmission delay is 35 milliseconds, then the first quality deviation is -5 milliseconds, indicating that the actual delay is 5 milliseconds lower than the expected delay.

[0119] Step E2, calculate the deviation between the actual energy consumption and the expected energy consumption to obtain the second quality deviation.

[0120] The actual energy consumption is the total energy consumed by all nodes participating in data forwarding on the path during transmission, which is accumulated by reading the energy monitoring data of each node. The expected energy consumption is the estimated value calculated in step D1 based on the theoretical energy consumption model of each node. The second quality deviation is calculated by subtracting the expected energy consumption from the actual energy consumption.

[0121] Exemplarily, after the completion of data transmission, the actual energy consumption data of each node on the path during transmission is collected. Each node records the electric energy consumed during data processing and forwarding through the built-in energy monitoring module, and the value is counted in joules. The energy consumption values of all nodes on the path are summed to obtain the total actual energy consumption.

[0122] The calculation method of the second quality deviation is: the total value of the actual energy consumption - the total value of the expected energy consumption calculated in step D1. For example, if the actual energy consumption is 12.8 joules and the expected energy consumption is 10.5 joules, then the second quality deviation is 2.3 joules, indicating that the actual energy consumption exceeds the expected value by 2.3 joules. The accuracy of energy monitoring is maintained within 0.1 joules to ensure the accuracy of the deviation calculation.

[0123] Step E3, combine the first quality deviation with the first correction factor and combine the second quality deviation with the second correction factor to generate the parameter adjustment value.

[0124] The first correction factor is a conversion coefficient for converting the delay deviation into a delay prediction parameter adjustment amount, and its value is determined according to statistical analysis of historical deviation data. The second correction factor is a conversion coefficient for converting the energy consumption deviation into an energy consumption prediction parameter adjustment amount. The parameter adjustment value is calculated by multiplying the quality deviation by the corresponding correction factor, which is used for subsequent parameter reference value correction.

[0125] For example, the first correction factor is determined by collecting the delay deviation data of the last 100 routing decisions, calculating the standard deviation of the deviation data, taking the reciprocal of the standard deviation as the base value of the first correction factor, and multiplying it by a regulation coefficient of 0.1 to obtain the first correction factor. When the historical standard deviation of the delay deviation is 8 milliseconds, the first correction factor is equal to 1 / 8 x 0.1 = 0.0125.

[0126] The second correction factor is calculated based on the standard deviation of the historical energy consumption deviation data using the same determination method. When the historical standard deviation of the energy consumption deviation is 2.5 joules, the second correction factor is equal to 1 / 2.5 x 0.1 = 0.04. The value of the correction factor is updated regularly, and is recalculated every 50 routing decisions to maintain the adaptability of the correction mechanism.

[0127] The calculation method of the parameter adjustment value is: the first quality deviation x the first correction factor to obtain the delay adjustment value, and the second quality deviation x the second correction factor to obtain the energy consumption adjustment value. Using the aforementioned example data, the delay adjustment value is 5 milliseconds x 0.0125 = 0.0625, and the energy consumption adjustment value is 2.3 joules x 0.04 = 0.092.

[0128] Step E4, apply the parameter adjustment value to the baseline value of the corresponding parameter in the routing decision process.

[0129] The baseline value of the corresponding parameter includes the predicted baseline of the node processing delay and the predicted baseline of the node energy consumption, which are used in the performance index calculation in step D1. The application of the parameter adjustment value is realized by adding the adjustment value to the corresponding baseline value, and the positive adjustment value indicates an increase in the baseline value, and the negative adjustment value indicates a decrease in the baseline value. The adjusted baseline value is used for performance prediction calculation in the next routing decision process.

[0130] For example, the delay adjustment value is applied to the processing delay prediction baseline of each node on the path. Assuming that the processing delay prediction baseline of a certain node is 3 milliseconds, and the delay adjustment value is 0.0625, then the adjusted prediction baseline is 3.0625 milliseconds. This adjusted value will be used in the next path delay calculation involving this node, improving the accuracy of the delay prediction.

[0131] The energy consumption adjustment value is distributed according to the proportion of each node in the total energy consumption of the path. If the path contains 3 nodes, the energy consumption proportion of each node is 40%, 35%, and 25% respectively, and the energy consumption adjustment value is 0.092 joules, then the energy consumption baseline of the three nodes is increased by 0.092 x 40% = 0.0368, 0.092 x 35% = 0.0322, and 0.092 x 25% = 0.023 joules respectively.

[0132] The upper and lower limits of the amplitude of the reference value adjustment are set to constrain the amplitude of a single adjustment to not exceed 20% of the original reference value, avoiding excessive adjustment caused by abnormal data. The adjusted reference value is saved in the parameter database of each node, providing more accurate prediction basis for subsequent routing decisions. When the cumulative adjustment of the reference value exceeds 50% of the original value, the reference value reset mechanism is triggered, and the reference value is recalibrated using the latest standard test data.

[0133] It should be noted that the parameter correction of the present application realizes self-optimization of routing performance prediction through continuous actual measurement feedback and reference adjustment. The historical data generated during the correction process is used to analyze the network performance change trend, providing data support for network capacity planning and performance optimization. Steps 202 to 210 constitute a complete multi-agent collaborative routing optimization closed loop, which realizes continuous optimization and adjustment improvement of routing decisions through the combination of predictive analysis, collaborative scheduling, hierarchical processing and parameter correction.

[0134] It should be understood that although each step in the flowchart involved in each embodiment as described above is displayed in sequence according to the arrow, these steps are not necessarily executed in sequence according to the arrow. Unless otherwise specified herein, the execution of these steps is not strictly limited in sequence, and these steps can be executed in other sequences. Moreover, at least part of the steps in the flowchart involved in each embodiment as described above can include multiple steps or stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution sequence of these steps or stages is not necessarily sequential, but can be executed in rotation or alternation with at least part of other steps or steps or stages in other steps.

[0135] Based on the same inventive concept, the present application also provides a multi-agent collaborative data exchange dynamic routing optimization system. The implementation scheme for solving the problem provided by the system is similar to the implementation scheme described in the above method, so the specific limitations in one or more multi-agent collaborative data exchange dynamic routing optimization system embodiments provided below can refer to the limitations of the multi-agent collaborative data exchange dynamic routing optimization method in the above, which will not be repeated here.

[0136] In one exemplary embodiment, as shown in Figure 3 a multi-agent collaborative data exchange dynamic routing optimization system is provided, comprising: a state acquisition module for acquiring real-time state information of each node in the network, and outputting a node state data set; The analysis and early warning module is configured to construct a historical state sequence of each node based on the node state data set, calculate a load change trend of each node through sequence analysis, and output a routing adjustment signal when the load change trend reaches a load critical value; The cooperative calculation module is configured to calculate a cooperative weight according to a network contribution degree of each agent in response to the routing adjustment signal, establish a task allocation mapping relationship among the agents based on the cooperative weight, and output a cooperative scheduling result; The hierarchical decision module is configured to determine a candidate path set through a hierarchical processing architecture according to the cooperative scheduling result, perform index evaluation on paths in the candidate path set, and output an optimal routing path. The parameter correction module is configured to perform data transmission based on the optimal routing path, obtain a quality difference between an actual transmission effect and an expected effect, and perform parameter correction on routing decision according to the quality difference.

[0137] Each module in the multi-agent cooperative data exchange dynamic routing optimization system described above can be implemented wholly or partially through software, hardware, or a combination thereof. Each module described above can be embedded in or independent of a processor in a computer device in hardware form, or can be stored in a memory in the computer device in software form, so as to be called and executed by a processor to perform operations corresponding to each module.

[0138] In an exemplary embodiment, a computer device, which can be a terminal, is provided, and an internal structure diagram of the computer device can be as shown in Figure 4As shown in the figure. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit and an input device. Among them, the processor, the memory and the input / output interface are connected through the system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capability. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operating system and the computer program in the non-volatile storage medium to run. The input / output interface of the computer device is used to exchange information between the processor and the external device. The communication interface of the computer device is used to communicate with the external terminal in a wired or wireless manner. The wireless manner can be realized through WIFI, mobile cellular network, near field communication (Near Field Communication, NFC) or other technologies. The computer program is executed by the processor to realize a multi-agent collaborative data exchange dynamic routing optimization method. The display unit of the computer device is used to form a visually visible picture, which can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer overlaid on the display screen, or a key, trackball or touchpad arranged on the shell of the computer device, or an external keyboard, touchpad or mouse, etc.

[0139] Those skilled in the art can understand that, Figure 4 The structure shown in the figure is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied. The specific computer device can include more or fewer components than those shown in the figure, or combine certain components, or have a different component arrangement.

[0140] In one embodiment, a computer device is also provided, including a memory and a processor, the memory storing a computer program, and the processor executing the computer program to realize the steps in each of the above method embodiments.

[0141] In one embodiment, a computer readable storage medium is provided, which stores a computer program, and the computer program is executed by a processor to realize the steps in each of the above method embodiments.

[0142] In one embodiment, a computer program product is provided, including a computer program, and the computer program is executed by a processor to realize the steps in each of the above method embodiments.

[0143] It should be noted that the user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties, and the collection, use and processing of related data need to comply with relevant regulations.

[0144] It can be understood by those skilled in the art that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing related hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium. When the computer program is executed, it can include the processes of the above-mentioned embodiments of each method. In the embodiments provided in the present application, any reference to memory, database or other medium can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The database involved in the embodiments provided in the present application can include at least one of a relational database and a non-relational database. The non-relational database can include a distributed database based on a block chain, etc., without being limited thereto. The processor involved in the embodiments provided in the present application can be a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., without being limited thereto.

[0145] Any technical features in the above embodiments can be combined, and for the sake of brevity, not all possible combinations are described above, however, any combination of these technical features is deemed to be within the scope of the present application.

[0146] The above embodiments only express several implementation manners of the present application, and the description is relatively specific and detailed, but it should not be understood as a limitation on the patent scope of the present application. It should be pointed out that, for ordinary skilled persons in the art, several modifications and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. A multi-agent cooperative data exchange dynamic routing optimization method, characterized in that, The method comprises the following steps: acquiring real-time state information of each node in the network, and outputting a node state data set; constructing a historical state sequence of each node based on the node state data set, calculating a load change trend of each node through sequence analysis, and outputting a routing adjustment signal when the load change trend reaches a load threshold value; in response to the routing adjustment signal, calculating a collaboration weight value according to the network contribution degree of each agent, establishing a task allocation mapping relationship between agents based on the collaboration weight value, and outputting a collaborative scheduling result; determining a candidate path set through a hierarchical processing architecture according to the collaborative scheduling result, evaluating the paths in the candidate path set, and outputting an optimal routing path; based on the optimal routing path, performing data transmission, acquiring a quality difference between an actual transmission effect and an expected effect, and performing parameter correction on the routing decision according to the quality difference.

2. The multi-agent cooperative data exchange dynamic route optimization method according to claim 1, characterized in that: The method of constructing a historical state sequence of each node based on the node state data set, and calculating a load change trend of each node through sequence analysis comprises the following steps: extracting state data of each node within a preset time period, and arranging the state data in chronological order to form the historical state sequence; analyzing the historical state sequence to identify the regular characteristics of load changes; calculating a load prediction value at a target time based on the regular characteristics; comparing the load prediction value with the load threshold value, and generating the routing adjustment signal when the load prediction value exceeds the load threshold value.

3. The multi-agent cooperative data exchange dynamic route optimization method of claim 2, wherein: The method of calculating a collaboration weight value according to the network contribution degree of each agent comprises the following steps: determining a path contribution value by counting the number of path resources provided by each agent to the network; determining a load balancing contribution value by calculating the correspondence between the network load shared by each agent and the system average load; determining an effect optimization contribution value by evaluating the improvement amplitude of the system effect after each agent joins the collaboration; comprehensively calculating the path contribution value, the load balancing contribution value and the effect optimization contribution value to obtain the collaboration weight value of each agent.

4. The multi-agent cooperative data exchange dynamic route optimization method of claim 3, wherein: The method of determining a candidate path set through a hierarchical processing architecture comprises the following steps: establishing a path basic framework based on the network topology structure at the global planning layer to form an initial path set; updating the initial path set by exchanging local information based on the node state data set at the local coordination layer; selecting paths meeting the quality requirements from the updated path set through a quality index screening mechanism at the execution layer to form the candidate path set.

5. The multi-agent cooperative data exchange dynamic route optimization method of claim 4, wherein: The method of evaluating the paths in the candidate path set comprises the following steps: acquiring transmission delay indicators, energy consumption indicators and load balancing indicators of each candidate path; determining calculation coefficients of each indicator according to network application requirements; comprehensively calculating each indicator value and the corresponding calculation coefficient to obtain a comprehensive quality score of each candidate path; selecting the path with the highest comprehensive quality score as the optimal routing path.

6. The multi-agent cooperative data exchange dynamic route optimization method of claim 5, wherein: The method of performing parameter correction on the routing decision according to the quality difference comprises the following steps: calculating the deviation between the actual transmission delay and the expected transmission delay to obtain a first quality deviation; calculating the deviation between the actual energy consumption and the expected energy consumption to obtain a second quality deviation; combining the first mass deviation with a first correction factor and combining the second mass deviation with a second correction factor to generate a parameter adjustment value; applying the parameter adjustment value to a reference value of a corresponding parameter in a routing decision process.

7. The multi-agent cooperative data exchange dynamic route optimization method of claim 6, wherein: The length of the preset time period is set according to the network state fluctuation frequency, when the network state fluctuation frequency is higher than the fluctuation threshold, the length of the preset time period is reduced, when the network state fluctuation frequency is lower than the fluctuation threshold, the length of the preset time period is increased.

8. A multi-agent cooperative data exchange dynamic routing optimization system, adopting the multi-agent cooperative data exchange dynamic routing optimization method according to any one of claims 1-7. Comprising: The state acquisition module is used for acquiring the real-time state information of each node in the network, and outputs a node state data set; The analysis and early warning module is used for constructing a historical state sequence of each node based on the node state data set, calculating a load change trend of each node through sequence analysis, and outputting a routing adjustment signal when the load change trend reaches a load critical value; The collaborative calculation module is used for responding to the routing adjustment signal, calculating a collaborative weight according to the network contribution degree of each agent, establishing a task allocation mapping relationship between agents based on the collaborative weight, and outputting a collaborative scheduling result; The hierarchical decision module is used for determining a candidate path set according to the collaborative scheduling result through a hierarchical processing architecture, evaluating the paths in the candidate path set according to the index, and outputting an optimal routing path; The parameter correction module is used for executing data transmission based on the optimal routing path, obtaining the quality difference between the actual transmission effect and the expected effect, and performing parameter correction on the routing decision according to the quality difference. 9.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-8 when the computer program is executed by the processor. The processor executes the computer program to realize the steps of the multi-agent collaborative data exchange dynamic routing optimization method of any one of claims 1 to 7.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to realize the steps of the multi-agent collaborative data exchange dynamic routing optimization method of any one of claims 1 to 7.

Images