Adaptive sensing and cooperative transmission method and system based on industrial internet of things

Abstract

The application discloses an adaptive sensing and cooperative transmission method and system based on an industrial Internet of Things, relates to the technical field of the industrial Internet of Things, and comprises the following steps: modeling each heterogeneous service flow accessing an edge gateway into a resource demand agent, allocating a virtual budget to each resource demand agent, and predefining a differentiated utility function for each category of heterogeneous service flow; constructing the edge gateway into a resource transaction market with the resource demand agent as a participant; dynamically calculating a resource price according to current network resource state information and broadcasting the resource price to each resource demand agent; updating the resource price according to the bidding request of each resource demand agent and rebroadcasting; and performing allocation and scheduling of heterogeneous resources according to the resource demand of each resource demand agent in a Nash equilibrium state and deducting the virtual budget corresponding to each resource demand agent. The application has the effect of improving the flexibility of resource scheduling.

Description

Technical Field

[0001] This application relates to the technical field of industrial Internet of Things (IIoT), and in particular to adaptive sensing and collaborative transmission methods and systems based on IIoT. Background Technology

[0002] With the deepening development of industry and intelligent manufacturing, the Industrial Internet of Things (IIoT), as a key supporting technology, is experiencing continuous expansion in network scale, a surge in the number of connected devices, and increasingly complex service types. In typical IIoT systems, various heterogeneous service flows coexist, such as periodic control command flows for precision motion control, high-definition video streams for equipment status monitoring and fault diagnosis, and massive sensor data streams for environmental perception. These heterogeneous service flows exhibit significant differences in data rate, packet size, latency tolerance, and reliability requirements, and their workloads and network environments are both time-varying and bursty.

[0003] Currently, resource allocation and scheduling in the Industrial Internet of Things (IIoT) mostly adopt static configuration methods based on traditional network protocols or rely on centralized cloud platforms for centralized optimization. However, these methods have significant shortcomings when facing the aforementioned complex heterogeneous scenarios. Static configuration methods lack flexibility and cannot dynamically adjust resources according to instantaneous channel changes or service priorities, easily leading to resource waste or Quality of Service (QoS) breaches. While centralized optimization schemes can theoretically obtain globally optimal solutions, they rely on aggregating all sensor data and service status to a central node, which not only generates huge backhaul link overhead, but also causes the computational complexity of the optimization algorithm to increase exponentially with network scale, making it difficult to meet the millisecond-level real-time response requirements of industrial sites. In addition, existing methods generally lack an effective trade-off mechanism between "service value" and "resource cost," making it difficult to achieve fair and efficient collaboration among various heterogeneous service flows while ensuring high-priority services.

[0004] Therefore, designing a distributed resource collaborative transmission method that can adaptively sense network dynamic changes, support differentiated service quality requirements of heterogeneous business flows, and has low overhead and high scalability has become a key technical challenge that urgently needs to be solved in the current industrial Internet of Things field. Summary of the Invention

[0005] To improve the flexibility of resource scheduling, this application provides an adaptive sensing and collaborative transmission method and system based on the Industrial Internet of Things.

[0006] Firstly, this application provides an adaptive sensing and collaborative transmission method based on the Industrial Internet of Things, employing the following technical solution: An adaptive sensing and collaborative transmission method based on the Industrial Internet of Things (IIoT) is applied to an IIoT system, which includes a management platform, a sensor network platform, and an object platform connected in sequence. The method is executed by the edge gateway of the sensor network platform and includes: Each heterogeneous service flow accessing the edge gateway is modeled as a resource demand agent with independent decision-making capabilities. A virtual budget is allocated to each resource demand agent, and a differentiated utility function is predefined for each type of heterogeneous service flow. The edge gateway is constructed as a resource trading market with the resource demand agents as the main participants; The resource price is dynamically calculated based on the current network resource status information, and the resource price is broadcast to each of the resource demand agents. After receiving the resource price, each of the resource demand agents calculates the resource demand amount through a distributed optimization algorithm based on the corresponding utility function, the current virtual budget balance, and the resource price, and submits a bidding request containing the resource demand amount to the edge gateway. The resource price is updated according to the bidding requests of each resource demand agent and rebroadcast. Each resource demand agent, after receiving the updated resource price, adjusts its own resource demand according to the updated resource price and repeats the broadcast process until the bidding strategies of all resource demand agents reach a Nash equilibrium. Based on the resource demand of each resource demand agent under the Nash equilibrium state, heterogeneous resources are allocated and scheduled, and the virtual budget corresponding to each resource demand agent is deducted to complete adaptive perception and collaborative transmission.

[0007] By adopting the above technical solution, each heterogeneous service flow accessing the edge gateway is modeled as a resource demand agent with independent decision-making capabilities. A virtual budget is allocated to each resource demand agent, and a differentiated utility function is predefined for each type of heterogeneous service flow. Then, the edge gateway is constructed as a resource trading market with resource demand agents as the main participants. The resource price is then dynamically calculated based on the current network resource status information and broadcast to each resource demand agent. Upon receiving the resource price, each resource demand agent, based on its corresponding utility function, current virtual budget balance, and resource price, performs distributed optimization calculations. The method calculates resource demand and submits bidding requests containing resource demand to the edge gateway. Then, it updates resource prices based on the bidding requests of each resource-demanding agent and rebroadcasts the changes. Each resource-demanding agent, upon receiving the updated resource price, adjusts its own resource demand and repeats the broadcast process until all bidding strategies of all resource-demanding agents reach a Nash equilibrium. Finally, based on the resource demands of each agent in the Nash equilibrium state, heterogeneous resources are allocated and scheduled, and the virtual budget corresponding to each agent is deducted to achieve adaptive perception and collaborative transmission. This invention models heterogeneous service flows as autonomous decision-making resource-demanding agents and constructs a resource trading market. By replacing centralized optimization with a distributed bidding mechanism, it solves the problems of poor flexibility in static configuration and high signaling overhead and insufficient real-time performance in centralized solutions. Through price signals and differentiated utility functions, it achieves an effective trade-off between service value and resource cost, improving the flexibility of resource scheduling and system scalability while ensuring differentiated service quality.

[0008] Optionally, the step of modeling each heterogeneous service flow accessing the edge gateway as a resource demand agent with independent decision-making capabilities includes: Identify each heterogeneous service flow that is accessed, and extract the service flow feature parameters of each heterogeneous service flow, wherein the service flow feature parameters include category identifier, service quality level, latency sensitivity and service flow priority; A resource demand agent is instantiated for each heterogeneous service flow, and the corresponding service flow feature parameters are injected into the decision configuration file of the resource demand agent. Each resource-demanding agent is configured with an independent decision engine, wherein the decision engine has a built-in execution environment for distributed optimization algorithms to support the resource-demanding agents in making independent decisions; Establish a bidirectional communication channel between the resource demand agent and the edge gateway.

[0009] By adopting the above technical solution, in order to model the resource demand agent, each heterogeneous service flow is identified, and the service flow feature parameters of each heterogeneous service flow are extracted. The service flow feature parameters include category identifier, service quality level, latency sensitivity, and service flow priority. Then, a resource demand agent is instantiated for each heterogeneous service flow, and the corresponding service flow feature parameters are injected into the decision configuration file of the resource demand agent. Then, an independent decision engine is configured for each resource demand agent. The decision engine has a built-in distributed optimization algorithm execution environment to support the resource demand agent to make independent decisions. Finally, a bidirectional communication channel is established between the resource demand agent and the edge gateway.

[0010] Optionally, the step of allocating a virtual budget to each of the resource-demanding agents includes: Obtain the budget allocation basis parameters for each of the heterogeneous service flows, wherein the budget allocation basis parameters include service level agreement weights, historical data traffic statistics, and preset priority coefficients; Based on the budget allocation parameters, a weighted allocation algorithm is used to calculate the corresponding virtual budget, and the virtual budget is bound to the corresponding resource demand agent. Configure a dynamic budget adjustment strategy, wherein the dynamic budget adjustment strategy includes a periodic replenishment mechanism, an on-demand recharge interface, and penalty deduction rules; Configure budget consumption rules, wherein the budget consumption rules are used to determine the deduction amount of the virtual budget based on the actual allocated amount of resources and according to a preset resource pricing benchmark.

[0011] By adopting the above technical solution, in order to achieve the allocation of virtual budget, the budget allocation basis parameters of each heterogeneous business flow are obtained. The budget allocation basis parameters include service level agreement weights, historical data traffic statistics, and preset priority coefficients. Then, based on the budget allocation basis parameters, a weighted allocation algorithm is used to calculate the corresponding virtual budget, and the virtual budget is bound to the corresponding resource demand agent. Then, a dynamic budget adjustment strategy is configured, which includes a periodic replenishment mechanism, an on-demand recharge interface, and a penalty deduction rule. Finally, a budget consumption rule is configured, which is used to determine the deduction amount of the virtual budget based on the actual allocated resource amount and according to the preset resource pricing benchmark.

[0012] Optionally, the step of predefining differentiated utility functions for each category of heterogeneous service flows includes: Based on the service quality level identifier of the heterogeneous service flow, the heterogeneous service flow is divided into three categories: high real-time service flow, medium real-time service flow, and low real-time service flow. Different utility functions are configured for each category of heterogeneous service flows. Specifically, a concave function with high marginal returns and a slow rate of decline in marginal returns is configured for the high real-time service flows, a standard concave function is configured for the medium real-time service flows, and a function with rapidly decreasing marginal returns as resource consumption increases is configured for the low real-time service flows. The functional form of the utility function and its corresponding parameterized configuration are encapsulated into a utility function description file and sent to the corresponding resource-demanding agent.

[0013] By adopting the above technical solution, in order to define the utility function, heterogeneous service flows are divided into three categories—high real-time service flows, medium real-time service flows, and low real-time service flows—based on the service quality level identifier of the heterogeneous service flows. Then, different utility functions are configured for each category of heterogeneous service flows. Specifically, a concave function with high marginal returns and a slow rate of decline in marginal returns is configured for high real-time service flows, a standard concave function is configured for medium real-time service flows, and a function with rapidly decreasing marginal returns as resource consumption increases is configured for low real-time service flows. The functional form of the utility function and the corresponding parameterized configuration are then encapsulated into a utility function description file and distributed to the corresponding resource demand agent.

[0014] Optionally, the step of constructing the edge gateway as a resource trading market with the resource demand agent as the main participant includes: The edge gateway establishes a unified resource abstraction layer internally, which standardizes and encapsulates the managed computing resources, communication bandwidth resources, and storage resources, and defines tradable commodities in units of resource shares. Initialize market rules and broadcast the market rules to each resource-demanding agent, wherein the market rules include pricing cycle, bidding time window, resource allocation granularity, and virtual budget settlement mechanism; Create an independent market account for each resource demand agent, record the corresponding virtual budget balance, historical transaction records and credit score, and establish a mapping relationship between the resource demand agent and the business flow category; Start the market operation state machine and set the market state sequentially to the initialization stage, bidding stage, settlement stage and scheduling execution stage. In each of the market states, interact with each of the resource demand agents according to the market rules.

[0015] By adopting the above technical solution, in order to build a resource trading market, the edge gateway establishes a unified resource abstraction layer internally, standardizes and encapsulates the managed computing resources, communication bandwidth resources, and storage resources, defines tradable commodities in units of resource shares, initializes market rules, and broadcasts these rules to each resource-demanding agent. These market rules include pricing cycles, bidding time windows, resource allocation granularity, and a virtual budget settlement mechanism. Then, an independent market account is created for each resource-demanding agent, recording the corresponding virtual budget balance, historical transaction records, and reputation score. A mapping relationship is established between resource-demanding agents and business flow categories. Finally, the market operation state machine is started, sequentially setting the market state to the initialization phase, bidding phase, settlement phase, and scheduling execution phase. In each market state, the gateway interacts with each resource-demanding agent according to the market rules.

[0016] Optionally, the step of dynamically calculating resource prices based on current network resource status information includes: Real-time collection of current network resource status information, including bandwidth utilization, computing resource utilization, storage resource utilization, and queue backlog length; Based on the supply and demand of various network resources, the resource price for this round is calculated using a price update function, wherein the price update function is set to: increase the resource price when the total resource demand exceeds the available resource capacity, and decrease the resource price when the total resource demand is lower than the available resource capacity; The calculated resource price is smoothed compared with the resource price in the previous round, and the smoothed resource price is broadcast to each resource-demanding agent.

[0017] By adopting the above technical solution, in order to calculate resource prices, the current network resource status information is collected in real time. The current network resource status information includes bandwidth utilization, computing resource utilization, storage resource utilization, and queue backlog length. Then, based on the supply and demand of various network resources, the resource price for this round is calculated using a price update function. The price update function is set to: increase the resource price when the total resource demand exceeds the available resource capacity, and decrease the resource price when the total resource demand is lower than the available resource capacity. Then, the calculated resource price is smoothed with the resource price of the previous round, and the smoothed resource price is broadcast to each resource demand agent.

[0018] Optionally, the step of calculating the resource demand using a distributed optimization algorithm based on the corresponding utility function, the current virtual budget balance, and the resource price includes: The resource-demanding agent reads its own stored utility function and virtual budget balance; Construct an objective function and set a virtual budget balance as an upper limit constraint on resource demand. The objective function is used to characterize the net benefit between resource utility and usage cost. The objective function is the utility function minus the product of the resource price and the resource demand. The objective function is solved using the gradient descent method, and the resource requirement that maximizes the value of the objective function is obtained through iterative calculation.

[0019] By adopting the above technical solution, in order to calculate the resource demand, the resource demand agent reads its own stored utility function and virtual budget balance, then constructs the objective function, and sets the virtual budget balance as the upper limit constraint of the resource demand. The objective function is used to characterize the net benefit between resource utility and usage cost. The objective function is the utility function minus the product of resource price and resource demand. Then, the gradient descent method is used to solve the objective function, and the resource demand that maximizes the value of the objective function is obtained through iterative calculation.

[0020] Optionally, the step of updating the resource price based on the bidding requests of each of the resource-demanding agents and rebroadcasting it includes: Receive all the bidding requests submitted by the resource demand agents in this round, and extract the resource demand from the bidding requests; The resource demands are aggregated to calculate the total demand for various network resources. Based on the deviation between the total demand for various network resources and the current available resource capacity, a subgradient method is used to update the resource price. The deviation is used to characterize the supply-demand gap. The subgradient method aims to minimize the deviation. When the total demand exceeds the available resource capacity, the resource price is increased; when the total demand is less than the available resource capacity, the resource price is decreased. The updated resource price is broadcast to each of the resource-demanding agents.

[0021] By adopting the above technical solution, in order to update resource prices, the system receives bidding requests submitted by all resource-demanding agents in the current round, extracts the resource demand from the bidding requests, aggregates the resource demands, calculates the total demand for various network resources, and then updates the resource prices using a subgradient method based on the deviation between the total demand for various network resources and the current available resource capacity. Here, the deviation is used to represent the supply-demand gap, and the subgradient method aims to minimize the deviation. When the total demand exceeds the available resource capacity, the resource price is increased; when the total demand is lower than the available resource capacity, the resource price is decreased. The updated resource prices are then broadcast to all resource-demanding agents.

[0022] Optionally, the step of adjusting its own resource demand based on the updated resource price and repeatedly executing the broadcast process until the bidding strategies of all resource demand agents reach a Nash equilibrium includes: Set the maximum number of iterations and the convergence accuracy threshold; In each iteration, each resource-demanding agent adjusts its own resource demand based on the updated resource price and submits an updated bidding request to the edge gateway. Determine whether the difference between the resource demand of each resource demand agent in this round and the resource demand in the previous round is less than the convergence accuracy threshold, and whether the difference between the resource price in this round and the resource price in the previous round is less than the preset price fluctuation threshold. If the determination result is yes, or the maximum number of iterations has been reached, then stop iterating and determine the current state as a Nash equilibrium state; If all the results are negative, then continue to the next iteration.

[0023] By adopting the above technical solution, in order to determine the Nash equilibrium state, a maximum number of iterations and a convergence accuracy threshold are set. In each iteration, each resource demand agent adjusts its own resource demand according to the updated resource price and submits an updated bidding request to the edge gateway. Then, it is determined whether the difference between the resource demand of each resource demand agent in this round and the resource demand in the previous round is less than the convergence accuracy threshold, and whether the difference between the resource price in this round and the resource price in the previous round is less than a preset price fluctuation threshold. If the determination result is yes, or the maximum number of iterations has been reached, the iteration stops and the current state is determined to be the Nash equilibrium state. If the determination result is no, the next iteration continues.

[0024] Secondly, this application also provides an adaptive sensing and collaborative transmission system based on the Industrial Internet of Things, which adopts the following technical solution: An adaptive sensing and collaborative transmission system based on the Industrial Internet of Things (IIoT) is applied to an IIoT system, which includes a management platform, a sensor network platform, and an object platform connected in sequence. The system is located in an edge gateway of the sensor network platform, and the edge gateway is configured with: The agent modeling and initialization module is used to model each heterogeneous service flow accessing the edge gateway as a resource demand agent with independent decision-making ability, allocate a virtual budget to each resource demand agent, and predefine differentiated utility functions for each type of heterogeneous service flow. The resource trading market construction module is used to construct the edge gateway into a resource trading market with the resource demand agents as the main participants. The broadcast module is used to dynamically calculate the resource price based on the current network resource status information and broadcast the resource price to each of the resource demand agents. After receiving the resource price, each of the resource demand agents calculates the resource demand amount through a distributed optimization algorithm based on the corresponding utility function, the current virtual budget balance and the resource price, and submits a bidding request containing the resource demand amount to the edge gateway. An optimization module is used to update the resource price according to the bidding requests of each resource demand agent and rebroadcast it. Each resource demand agent, after receiving the updated resource price, adjusts its own resource demand according to the updated resource price and repeatedly executes the broadcast process until the bidding strategies of all resource demand agents reach a Nash equilibrium. The scheduling and transmission module is used to perform heterogeneous resource allocation and scheduling based on the resource demand of each resource demand agent in the Nash equilibrium state, and deduct the virtual budget corresponding to each resource demand agent to complete adaptive perception and collaborative transmission.

[0025] In summary, this application includes at least the following beneficial technical effects: Each heterogeneous service flow accessing the edge gateway is modeled as a resource demand agent with independent decision-making capabilities; a virtual budget is allocated to each resource demand agent; and differentiated utility functions are predefined for each type of heterogeneous service flow. Then, the edge gateway is constructed as a resource trading market with resource demand agents as the main participants. Resource prices are then dynamically calculated based on the current network resource status information and broadcast to each resource demand agent. Upon receiving the resource price, each resource demand agent, based on its corresponding utility function, current virtual budget balance, and resource price, allocates a virtual budget to each resource demand agent. A distributed optimization algorithm calculates resource demand and submits bidding requests containing this demand to the edge gateway. Then, it updates resource prices based on the bidding requests of each resource-demanding agent and rebroadcasts the changes. Each resource-demanding agent, upon receiving the updated price, adjusts its own resource demand and repeats the broadcast process until all agents reach a Nash equilibrium. Finally, based on the resource demands of each agent in this Nash equilibrium, heterogeneous resources are allocated and scheduled, and the virtual budget corresponding to each agent is deducted to achieve adaptive perception and collaborative transmission. This invention models heterogeneous service flows as autonomous decision-making resource-demanding agents and constructs a resource trading market. By replacing centralized optimization with a distributed bidding mechanism, it solves the problems of poor flexibility in static configuration and high signaling overhead and insufficient real-time performance in centralized solutions. Through price signals and differentiated utility functions, it achieves an effective trade-off between service value and resource cost, improving resource scheduling flexibility and system scalability while ensuring differentiated service quality. Attached Figure Description

[0026] Figure 1 This is a schematic diagram of the overall process of an embodiment of this application.

[0027] Figure 2 This is a structural diagram of one application scenario of the system in this application embodiment.

[0028] Figure 3 This is a structural diagram of another application scenario of the system according to an embodiment of this application. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0030] This application discloses an adaptive sensing and collaborative transmission method based on the Industrial Internet of Things.

[0031] Reference Figure 1 An adaptive sensing and collaborative transmission method based on the Industrial Internet of Things (IIoT) is applied to an IIoT system, which includes a management platform, a sensor network platform, and an object platform connected in sequence. The method is executed by the edge gateway of the sensor network platform and includes: Step S11: Model each heterogeneous service flow accessing the edge gateway as a resource demand agent with independent decision-making capabilities, allocate a virtual budget to each resource demand agent, and predefine differentiated utility functions for each type of heterogeneous service flow.

[0032] It should be noted that step S11 breaks away from the traditional centralized scheduling model of processing all services uniformly by abstracting each heterogeneous service flow into an independent resource-demanding agent. Each agent possesses autonomous decision-making capabilities and can independently participate in resource competition based on its own service characteristics and real-time needs. The introduction of a virtual budget sets economic constraints on the agents' resource acquisition, simulating rational consumption behavior under limited resources. Differentiated utility functions are predefined according to service categories (such as high-reliability low-latency control flows, high-bandwidth video streams, and massive data acquisition streams) to quantify the satisfaction or benefits of different services when acquiring resources.

[0033] Step S12: Construct the edge gateway into a resource trading market with resource demand agents as the main participants.

[0034] It should be noted that step S12 transforms the edge gateway from a traditional resource scheduler to a market organizer. By constructing a resource trading market, the edge gateway no longer directly allocates resources to various service flows, but instead provides a platform and rules for resource trading. This market uses the resource demand agent established in step S11 as the buyer and the heterogeneous network resources (such as bandwidth, time slots, and computing resources) managed by the edge gateway as the commodity, forming a mechanism for supply and demand to interact through price signals. This transformation shifts resource allocation from "centralized decision-making" to "distributed game theory," providing architectural support for achieving low-overhead, highly flexible adaptive scheduling in the future.

[0035] Step S13: Dynamically calculate the resource price based on the current network resource status information, and broadcast the resource price to each resource demand agent. After receiving the resource price, each resource demand agent calculates the resource demand amount through a distributed optimization algorithm based on the corresponding utility function, the current virtual budget balance, and the resource price, and submits a bidding request containing the resource demand amount to the edge gateway.

[0036] It should be noted that in step S13, the edge gateway first dynamically calculates the resource price based on the current network resource usage (such as remaining bandwidth, channel quality, queue backlog, etc.). This price reflects the real-time scarcity of resources and is broadcast to all resource-demanding agents. After learning the price, each agent, combining its own differentiated utility function and remaining virtual budget, independently solves the resource demand that maximizes its utility at a given price using a distributed optimization algorithm, forming a bidding request and submitting it to the edge gateway. This process enables the business side to perceive and respond autonomously to resource costs in real time, avoiding the communication overhead caused by uploading all raw states to the central node.

[0037] Step S14: Update the resource price according to the bidding requests of each resource demand agent and rebroadcast it. After receiving the updated resource price, each resource demand agent adjusts its own resource demand according to the updated resource price and repeats the broadcast process until the bidding strategies of all resource demand agents reach a Nash equilibrium.

[0038] It should be noted that step S14 achieves dynamic matching and equilibrium convergence of resource supply and demand through an iterative bidding mechanism. After collecting bidding requests from all agents, the edge gateway adjusts resource prices based on supply and demand: if demand exceeds available resources, the price is increased to suppress demand; conversely, the price is decreased to incentivize demand. The updated price is broadcast again, and agents recalculate and submit new demands accordingly. This process iterates until no agent can unilaterally change its demand to obtain higher utility, reaching a Nash equilibrium. In this equilibrium, resource prices and the demands of each agent are mutually adapted, achieving stable and efficient overall resource allocation under individual budget constraints.

[0039] Step S15: Based on the resource demand of each resource demand agent under the Nash equilibrium state, perform heterogeneous resource allocation and scheduling, and deduct the virtual budget corresponding to each resource demand agent to complete adaptive perception and collaborative transmission.

[0040] It should be noted that step S15 translates the game equilibrium result into actual resource allocation and execution. In the Nash equilibrium state, each agent's resource demand reflects its most advantageous resource acquisition strategy guided by price signals. Based on this, the edge gateway performs the final heterogeneous resource mapping and scheduling, allocating bandwidth, time slots, and computing resources to the corresponding business flows according to demand. Simultaneously, by deducting each agent's corresponding virtual budget based on the traded resource quantity and current price, the "payment" process is completed. This budget deduction mechanism makes the virtual budget a constraint on the agents' long-term decisions, prompting each business flow to rationally participate in competition in subsequent cycles. Thus, the system completes a full cycle of adaptive perception and collaborative transmission.

[0041] In the above implementation, each heterogeneous service flow accessing the edge gateway is modeled as a resource demand agent with independent decision-making capabilities. A virtual budget is allocated to each resource demand agent, and a differentiated utility function is predefined for each type of heterogeneous service flow. Then, the edge gateway is constructed as a resource trading market with resource demand agents as the main participants. The resource price is then dynamically calculated based on the current network resource status information and broadcast to each resource demand agent. Upon receiving the resource price, each resource demand agent, based on its corresponding utility function, current virtual budget balance, and resource price, uses a distributed optimization algorithm to determine its target price. The system calculates resource demand and submits bidding requests containing this demand to the edge gateway. It then updates resource prices based on the bidding requests of each resource-demanding agent and rebroadcasts the changes. Each resource-demanding agent, upon receiving the updated price, adjusts its own resource demand and repeats the broadcast process until all agents reach a Nash equilibrium. Finally, based on the resource demands of each agent in this Nash equilibrium, heterogeneous resources are allocated and scheduled, and the virtual budget corresponding to each agent is deducted to achieve adaptive perception and collaborative transmission. This invention models heterogeneous service flows as autonomous decision-making resource-demanding agents and constructs a resource trading market. By replacing centralized optimization with a distributed bidding mechanism, it solves the problems of poor flexibility in static configuration and high signaling overhead and insufficient real-time performance in centralized solutions. Through price signals and differentiated utility functions, it achieves an effective trade-off between service value and resource cost, improving resource scheduling flexibility and system scalability while ensuring differentiated service quality.

[0042] As a further implementation of the method, the step of modeling each heterogeneous service flow accessing the edge gateway as a resource demand agent with independent decision-making capabilities includes: Step S21: Identify each heterogeneous service flow that has been accessed and extract the service flow feature parameters of each heterogeneous service flow. The service flow feature parameters include category identifier, service quality level, latency sensitivity and service flow priority.

[0043] Step S22: Instantiate a resource demand agent for each heterogeneous business flow and inject the corresponding business flow feature parameters into the decision configuration file of the resource demand agent.

[0044] Step S23: Configure an independent decision engine for each resource demand agent. The decision engine has a built-in execution environment for distributed optimization algorithms to support the resource demand agents in making independent decisions.

[0045] Step S24: Establish a bidirectional communication channel between the resource demand agent and the edge gateway.

[0046] In the above implementation, in order to model the resource demand agent, each heterogeneous service flow is identified, and the service flow feature parameters of each heterogeneous service flow are extracted. The service flow feature parameters include category identifier, service quality level, latency sensitivity, and service flow priority. Then, a resource demand agent is instantiated for each heterogeneous service flow, and the corresponding service flow feature parameters are injected into the decision configuration file of the resource demand agent. Then, an independent decision engine is configured for each resource demand agent. The decision engine has a built-in distributed optimization algorithm execution environment to support the resource demand agent to make independent decisions. Finally, a bidirectional communication channel is established between the resource demand agent and the edge gateway.

[0047] As a further implementation of the method, the step of allocating a virtual budget to each resource-demanding agent includes: Step S31: Obtain the budget allocation basis parameters for each heterogeneous business flow. The budget allocation basis parameters include service level agreement weights, historical data traffic statistics, and preset priority coefficients.

[0048] Step S32: Calculate the corresponding virtual budget using a weighted allocation algorithm based on the budget allocation parameters, and bind the virtual budget to the corresponding resource demand agent.

[0049] Step S33: Configure the budget dynamic adjustment strategy, which includes a periodic replenishment mechanism, an on-demand recharge interface, and penalty deduction rules.

[0050] Step S34: Configure budget consumption rules, wherein the budget consumption rules are used to determine the deduction amount of the virtual budget based on the actual allocated amount of resources and according to the preset resource pricing benchmark.

[0051] In the above implementation, in order to allocate the virtual budget, the budget allocation basis parameters for each heterogeneous business flow are obtained. These parameters include service level agreement weights, historical data traffic statistics, and preset priority coefficients. Then, based on the budget allocation basis parameters, a weighted allocation algorithm is used to calculate the corresponding virtual budget, and the virtual budget is bound to the corresponding resource demand agent. Next, a dynamic budget adjustment strategy is configured, which includes a periodic replenishment mechanism, an on-demand recharge interface, and a penalty deduction rule. Finally, a budget consumption rule is configured, which is used to determine the deduction amount of the virtual budget based on the actual allocated resource amount and a preset resource pricing benchmark.

[0052] As a further implementation of the method, the step of predefining differentiated utility functions for each category of heterogeneous service flows includes: Step S41: Based on the service quality level identifier of the heterogeneous service flow, the heterogeneous service flow is divided into three categories: high real-time service flow, medium real-time service flow, and low real-time service flow.

[0053] Step S42: Configure different utility functions for each category of heterogeneous service flow. Specifically, configure a concave function with high marginal returns and a slow rate of decline in marginal returns for high real-time service flows, configure a standard concave function for medium real-time service flows, and configure a function with rapidly decreasing marginal returns as resource consumption decreases for low real-time service flows.

[0054] Step S43: Encapsulate the functional form of the utility function and its corresponding parameterized configuration into a utility function description file and distribute it to the corresponding resource requirement agent.

[0055] In the above implementation, in order to define the utility function, the heterogeneous service flows are divided into three categories—high real-time service flows, medium real-time service flows, and low real-time service flows—based on the service quality level identifier of the heterogeneous service flows. Then, different utility functions are configured for each category of heterogeneous service flows. Specifically, a concave function with high marginal returns and a slow rate of decline in marginal returns is configured for high real-time service flows, a standard concave function is configured for medium real-time service flows, and a function with rapidly decreasing marginal returns as resource consumption decreases for low real-time service flows. The functional form of the utility function and the corresponding parameterized configuration are then encapsulated into a utility function description file and distributed to the corresponding resource demand agent.

[0056] As a further implementation of the method, the step of constructing the edge gateway as a resource trading market with resource-demanding agents as the main participants includes: Step S51: The edge gateway establishes a unified resource abstraction layer internally, standardizes and encapsulates the managed computing resources, communication bandwidth resources and storage resources, and defines tradable commodities in units of resource shares.

[0057] Step S52: Initialize market rules and broadcast them to each resource-demanding agent. The market rules include pricing cycle, bidding time window, resource allocation granularity, and virtual budget settlement mechanism.

[0058] Step S53: Create an independent market account for each resource demand agent, record the corresponding virtual budget balance, historical transaction records and reputation score, and establish a mapping relationship between resource demand agents and business flow categories.

[0059] Step S54: Start the market operation state machine and set the market state sequentially to the initialization stage, bidding stage, settlement stage and scheduling execution stage. In each market state, interact with each resource demand agent according to the market rules.

[0060] In the above implementation, in order to build a resource trading market, the edge gateway establishes a unified resource abstraction layer internally, standardizes and encapsulates the managed computing resources, communication bandwidth resources, and storage resources, defines tradable commodities in units of resource shares, initializes market rules, and broadcasts the market rules to each resource-demanding agent. The market rules include pricing cycles, bidding time windows, resource allocation granularity, and virtual budget settlement mechanisms. Then, an independent market account is created for each resource-demanding agent, recording the corresponding virtual budget balance, historical transaction records, and reputation score, and establishing a mapping relationship between resource-demanding agents and business flow categories. Then, the market operation state machine is started, and the market state is sequentially set to the initialization phase, bidding phase, settlement phase, and scheduling execution phase. In each market state, the gateway interacts with each resource-demanding agent according to the market rules.

[0061] As a further implementation of the method, the step of dynamically calculating resource prices based on current network resource status information includes: Step S61: Collect current network resource status information in real time, including bandwidth utilization, computing resource utilization, storage resource utilization, and queue backlog length.

[0062] Step S62: Based on the supply and demand of various network resources, calculate the resource price for this round using a price update function. The price update function is set to: increase the resource price when the total resource demand exceeds the available resource capacity, and decrease the resource price when the total resource demand is lower than the available resource capacity.

[0063] Step S63: Smooth the calculated resource price with the resource price of the previous round, and broadcast the smoothed resource price to each resource demand agent.

[0064] In the above implementation, in order to calculate resource prices, current network resource status information is collected in real time. This information includes bandwidth utilization, computing resource utilization, storage resource utilization, and queue backlog length. Then, based on the supply and demand of various network resources, a price update function is used to calculate the resource price for the current round. The price update function is set to: increase the resource price when the total resource demand exceeds the available resource capacity, and decrease the resource price when the total resource demand is lower than the available resource capacity. The calculated resource price is then smoothed with the resource price of the previous round, and the smoothed resource price is broadcast to each resource demand agent.

[0065] As a further implementation of the method, the step of calculating resource demand using a distributed optimization algorithm based on the corresponding utility function, the current virtual budget balance, and resource prices includes: Step S71: The resource-demanding agent reads its stored utility function and virtual budget balance.

[0066] Step S72: Construct the objective function and set the virtual budget balance as the upper limit constraint of resource demand. The objective function is used to characterize the net benefit between resource utility and usage cost. The objective function is the utility function minus the product of resource price and resource demand.

[0067] Step S73: The objective function is solved using the gradient descent method, and the resource requirement that maximizes the value of the objective function is obtained through iterative calculation.

[0068] In the above implementation, in order to calculate the resource demand, the resource demand agent reads its stored utility function and virtual budget balance, then constructs an objective function, and sets the virtual budget balance as an upper limit constraint on the resource demand. The objective function is used to characterize the net benefit between resource utility and usage cost. The objective function is the utility function minus the product of resource price and resource demand. Then, the gradient descent method is used to solve the objective function, and the resource demand that maximizes the value of the objective function is obtained through iterative calculation.

[0069] As a further implementation of the method, the steps of updating resource prices based on the bidding requests of each resource-demanding agent and rebroadcasting include: Step S81: Receive bidding requests submitted by all resource-demanding agents in this round, and extract the resource demand from the bidding requests.

[0070] Step S82: Aggregate the resource requirements and calculate the total requirements for various network resources.

[0071] Step S83: Based on the deviation between the total demand for various network resources and the current available resource capacity, update the resource price using a subgradient method. The deviation is used to represent the supply-demand gap. The subgradient method aims to minimize the deviation. When the total demand exceeds the available resource capacity, the resource price is increased; when the total demand is lower than the available resource capacity, the resource price is decreased.

[0072] Step S84: Broadcast the updated resource price to each resource-demanding agent.

[0073] In the above implementation, in order to update resource prices, the system receives bidding requests submitted by all resource-demanding agents in the current round, extracts resource demand from the bidding requests, aggregates the resource demand, calculates the total demand for various types of network resources, and then updates resource prices using a subgradient method based on the deviation between the total demand for various types of network resources and the current available resource capacity. Here, the deviation is used to represent the supply-demand gap, and the subgradient method aims to minimize the deviation. When the total demand exceeds the available resource capacity, the resource price is increased; when the total demand is lower than the available resource capacity, the resource price is decreased. Then, the updated resource price is broadcast to each resource-demanding agent.

[0074] As a further implementation of the method, the steps of adjusting one's own resource demand based on the updated resource price and repeatedly executing the broadcast process until the bidding strategies of all resource-demanding agents reach a Nash equilibrium include: Step S91: Set the maximum number of iterations and the convergence accuracy threshold.

[0075] In step S92, in each iteration, each resource demand agent adjusts its own resource demand according to the updated resource price and submits an updated bidding request to the edge gateway.

[0076] Step S93: Determine whether the difference between the resource demand of each resource demand agent in this round and the resource demand in the previous round is less than the convergence accuracy threshold, and whether the difference between the resource price in this round and the resource price in the previous round is less than the preset price fluctuation threshold.

[0077] Step S94: If the determination result is yes, or the maximum number of iterations has been reached, then stop the iteration and determine the current state as a Nash equilibrium state.

[0078] In step S95, if all the determination results are negative, then continue to execute the next round of iteration.

[0079] In the above implementation, in order to determine the Nash equilibrium state, a maximum number of iterations and a convergence accuracy threshold are set. In each iteration, each resource demand agent adjusts its own resource demand according to the updated resource price and submits an updated bidding request to the edge gateway. Then, it is determined whether the difference between the resource demand of each resource demand agent in this round and the resource demand in the previous round is less than the convergence accuracy threshold, and whether the difference between the resource price in this round and the resource price in the previous round is less than a preset price fluctuation threshold. If the determination result is yes, or the maximum number of iterations has been reached, the iteration stops and the current state is determined to be the Nash equilibrium state. If the determination results are no, the next iteration continues.

[0080] This application also discloses an adaptive sensing and collaborative transmission system based on the Industrial Internet of Things.

[0081] refer to Figure 2 An adaptive sensing and collaborative transmission system based on the Industrial Internet of Things (IIoT) is applied to an IIoT system. The IIoT system includes a management platform, a sensor network platform, and an object platform that are sequentially connected. The system is located in the edge gateway of the sensor network platform, and the edge gateway is configured with: The agent modeling and initialization module is used to model each heterogeneous service flow accessing the edge gateway as a resource demand agent with independent decision-making ability, allocate a virtual budget to each resource demand agent, and predefine differentiated utility functions for each type of heterogeneous service flow. The resource trading market construction module is used to build the edge gateway into a resource trading market with resource demand agents as the main participants; The broadcast module is used to dynamically calculate resource prices based on the current network resource status information and broadcast the resource prices to each resource demand agent. After receiving the resource prices, each resource demand agent calculates the resource demand based on the corresponding utility function, the current virtual budget balance, and the resource prices using a distributed optimization algorithm, and submits a bidding request containing the resource demand to the edge gateway. The optimization module is used to update the resource price based on the bidding requests of each resource demand agent and rebroadcast it. After receiving the updated resource price, each resource demand agent adjusts its own resource demand according to the updated resource price and executes the broadcast process in a loop until the bidding strategies of all resource demand agents reach a Nash equilibrium. The scheduling and transmission module is used to allocate and schedule heterogeneous resources according to the resource demand of each resource demand agent under the Nash equilibrium state, and deduct the virtual budget corresponding to each resource demand agent to complete adaptive perception and collaborative transmission.

[0082] The overall framework of another application scenario of the adaptive sensing and cooperative transmission system based on the Industrial Internet of Things in this application is as follows: Figure 3 As shown, it can include a user platform, service platform, management platform, sensor network platform, and object platform that interact sequentially, forming a five-platform architecture based on the Industrial Internet of Things. The system is located in the edge gateway of the sensor network platform, and the edge gateway is configured with: an intelligent agent modeling and initialization module, a resource trading market construction module, a broadcast module, an optimization module, and a scheduling and transmission module.

[0083] Specifically, in the aforementioned application scenario, the adaptive sensing and collaborative transmission system based on the Industrial Internet of Things (IIoT) includes a sensor network platform, which includes an edge gateway. The edge gateway is configured to: model each heterogeneous service flow accessing the edge gateway as a resource-demanding agent with independent decision-making capabilities; allocate a virtual budget to each resource-demanding agent; and predefine differentiated utility functions for each type of heterogeneous service flow; construct the edge gateway as a resource trading market with resource-demanding agents as the main participants; dynamically calculate resource prices based on current network resource status information and broadcast the resource prices to each resource-demanding agent. Upon receiving the resource price, each resource-demanding agent, according to its corresponding... The utility function, current virtual budget balance, and resource price are used to calculate resource demand using a distributed optimization algorithm. A bidding request containing the resource demand is then submitted to the edge gateway. The resource price is updated based on the bidding requests of each resource-demanding agent and rebroadcast. Upon receiving the updated resource price, each resource-demanding agent adjusts its own resource demand accordingly and repeats the broadcast process until all bidding strategies of all resource-demanding agents reach a Nash equilibrium. Based on the resource demands of each resource-demanding agent in the Nash equilibrium state, heterogeneous resources are allocated and scheduled, and the virtual budget corresponding to each resource-demanding agent is deducted to complete adaptive perception and collaborative transmission.

[0084] By leveraging the interaction between various functional platforms of the industrial IoT-based adaptive sensing and collaborative transmission system, which is based on the aforementioned three or five platforms, a complete closed-loop information operation logic is established, ensuring the orderly operation of sensing and control information and realizing intelligent equipment management.

[0085] The adaptive sensing and collaborative transmission system based on the Industrial Internet of Things (IIoT) of the present invention can implement any of the adaptive sensing and collaborative transmission methods based on the Industrial Internet of Things (IIoT), and the specific working process of the adaptive sensing and collaborative transmission system based on the Industrial Internet of Things of the present invention can refer to the corresponding process in the above-mentioned adaptive sensing and collaborative transmission methods based on the Industrial Internet of Things.

[0086] The above are all preferred embodiments of this application and are not intended to limit the scope of protection of this application. Any feature disclosed in this specification (including the abstract and drawings) may be replaced by other equivalent or similar features unless specifically stated otherwise. That is, unless specifically stated otherwise, each feature is only one example of a series of equivalent or similar features.

Claims

1. An adaptive sensing and collaborative transmission method based on the Industrial Internet of Things, characterized in that, Applied to an industrial Internet of Things (IIoT) system, the IIoT system includes a management platform, a sensor network platform, and an object platform that are sequentially and communicatively connected. The method is executed by an edge gateway of the sensor network platform and includes: Each heterogeneous service flow accessing the edge gateway is modeled as a resource demand agent with independent decision-making capabilities. A virtual budget is allocated to each resource demand agent, and a differentiated utility function is predefined for each type of heterogeneous service flow. The edge gateway is constructed as a resource trading market with the resource demand agents as the main participants; The resource price is dynamically calculated based on the current network resource status information, and the resource price is broadcast to each of the resource demand agents. After receiving the resource price, each of the resource demand agents calculates the resource demand amount through a distributed optimization algorithm based on the corresponding utility function, the current virtual budget balance, and the resource price, and submits a bidding request containing the resource demand amount to the edge gateway. The resource price is updated according to the bidding requests of each resource demand agent and rebroadcast. Each resource demand agent, after receiving the updated resource price, adjusts its own resource demand according to the updated resource price and repeats the broadcast process until the bidding strategies of all resource demand agents reach a Nash equilibrium. Based on the resource demand of each resource demand agent under the Nash equilibrium state, heterogeneous resources are allocated and scheduled, and the virtual budget corresponding to each resource demand agent is deducted to complete adaptive perception and collaborative transmission.

2. The adaptive sensing and collaborative transmission method based on the Industrial Internet of Things according to claim 1, characterized in that, The step of modeling each heterogeneous service flow accessing the edge gateway as a resource demand agent with independent decision-making capabilities includes: Identify each heterogeneous service flow that is accessed, and extract the service flow feature parameters of each heterogeneous service flow, wherein the service flow feature parameters include category identifier, service quality level, latency sensitivity and service flow priority; A resource demand agent is instantiated for each heterogeneous service flow, and the corresponding service flow feature parameters are injected into the decision configuration file of the resource demand agent. Each resource-demanding agent is configured with an independent decision engine, wherein the decision engine has a built-in execution environment for distributed optimization algorithms to support the resource-demanding agents in making independent decisions; Establish a bidirectional communication channel between the resource demand agent and the edge gateway.

3. The adaptive sensing and collaborative transmission method based on the Industrial Internet of Things according to claim 1, characterized in that, The step of allocating a virtual budget to each of the resource-demanding agents includes: Obtain the budget allocation basis parameters for each of the heterogeneous service flows, wherein the budget allocation basis parameters include service level agreement weights, historical data traffic statistics, and preset priority coefficients; Based on the budget allocation parameters, a weighted allocation algorithm is used to calculate the corresponding virtual budget, and the virtual budget is then bound to the corresponding resource demand agent. Configure a dynamic budget adjustment strategy, wherein the dynamic budget adjustment strategy includes a periodic replenishment mechanism, an on-demand recharge interface, and penalty deduction rules; Configure budget consumption rules, wherein the budget consumption rules are used to determine the deduction amount of the virtual budget based on the actual allocated amount of resources and according to a preset resource pricing benchmark.

4. The adaptive sensing and collaborative transmission method based on the Industrial Internet of Things according to claim 1, characterized in that, The step of predefining differentiated utility functions for each category of heterogeneous service flows includes: Based on the service quality level identifier of the heterogeneous service flow, the heterogeneous service flow is divided into three categories: high real-time service flow, medium real-time service flow, and low real-time service flow. Different utility functions are configured for each category of heterogeneous service flows. Specifically, a concave function with high marginal returns and a slow rate of decline in marginal returns is configured for the high real-time service flows, a standard concave function is configured for the medium real-time service flows, and a function with rapidly decreasing marginal returns as resource consumption increases is configured for the low real-time service flows. The functional form of the utility function and its corresponding parameterized configuration are encapsulated into a utility function description file and sent to the corresponding resource-demanding agent.

5. The adaptive sensing and collaborative transmission method based on the Industrial Internet of Things according to claim 1, characterized in that, The step of constructing the edge gateway into a resource trading market with the resource demand agent as the main participant includes: The edge gateway establishes a unified resource abstraction layer internally, which standardizes and encapsulates the managed computing resources, communication bandwidth resources, and storage resources, and defines tradable commodities in units of resource shares. Initialize market rules and broadcast the market rules to each resource-demanding agent, wherein the market rules include pricing cycle, bidding time window, resource allocation granularity, and virtual budget settlement mechanism; Create an independent market account for each resource demand agent, record the corresponding virtual budget balance, historical transaction records and credit score, and establish a mapping relationship between the resource demand agent and the business flow category; Start the market operation state machine and set the market state sequentially to the initialization stage, bidding stage, settlement stage and scheduling execution stage. In each of the market states, interact with each of the resource demand agents according to the market rules.

6. The adaptive sensing and collaborative transmission method based on the Industrial Internet of Things according to claim 1, characterized in that, The step of dynamically calculating resource prices based on current network resource status information includes: Real-time collection of current network resource status information, including bandwidth utilization, computing resource utilization, storage resource utilization, and queue backlog length; Based on the supply and demand of various network resources, the resource price for this round is calculated using a price update function, wherein the price update function is set to: increase the resource price when the total resource demand exceeds the available resource capacity, and decrease the resource price when the total resource demand is lower than the available resource capacity; The calculated resource price is smoothed compared with the resource price in the previous round, and the smoothed resource price is broadcast to each resource-demanding agent.

7. The adaptive sensing and collaborative transmission method based on the Industrial Internet of Things according to claim 1, characterized in that, The step of calculating resource demand using a distributed optimization algorithm based on the corresponding utility function, the current virtual budget balance, and the resource price includes: The resource-demanding agent reads its own stored utility function and virtual budget balance; Construct an objective function and set a virtual budget balance as an upper limit constraint on resource demand. The objective function is used to characterize the net benefit between resource utility and usage cost. The objective function is the utility function minus the product of the resource price and the resource demand. The objective function is solved using the gradient descent method, and the resource requirement that maximizes the value of the objective function is obtained through iterative calculation.

8. The adaptive sensing and collaborative transmission method based on the Industrial Internet of Things according to claim 1, characterized in that, The step of updating the resource price based on the bidding requests of each of the resource-demanding agents and rebroadcasting it includes: Receive all the bidding requests submitted by the resource demand agents in this round, and extract the resource demand from the bidding requests; The resource demands are aggregated to calculate the total demand for various network resources. Based on the deviation between the total demand for various network resources and the current available resource capacity, a subgradient method is used to update the resource price. The deviation is used to characterize the supply-demand gap. The subgradient method aims to minimize the deviation. When the total demand exceeds the available resource capacity, the resource price is increased; when the total demand is less than the available resource capacity, the resource price is decreased. The updated resource price is broadcast to each of the resource-demanding agents.

9. The adaptive sensing and collaborative transmission method based on the Industrial Internet of Things according to claim 1, characterized in that, The step of adjusting its own resource demand based on the updated resource price and repeatedly executing the broadcast process until the bidding strategies of all resource demand agents reach a Nash equilibrium includes: Set the maximum number of iterations and the convergence accuracy threshold; In each iteration, each resource-demanding agent adjusts its own resource demand based on the updated resource price and submits an updated bidding request to the edge gateway. Determine whether the difference between the resource demand of each resource demand agent in this round and the resource demand in the previous round is less than the convergence accuracy threshold, and whether the difference between the resource price in this round and the resource price in the previous round is less than the preset price fluctuation threshold. If the determination result is yes, or the maximum number of iterations has been reached, then stop iterating and determine the current state as a Nash equilibrium state; If all the results are negative, then continue to the next iteration.

10. An adaptive sensing and collaborative transmission system based on the Industrial Internet of Things, characterized in that, This system is applied to an industrial Internet of Things (IIoT) system, which includes a management platform, a sensor network platform, and an object platform connected in sequence via communication. The system is located in an edge gateway of the sensor network platform, and the edge gateway is configured with: The agent modeling and initialization module is used to model each heterogeneous service flow accessing the edge gateway as a resource demand agent with independent decision-making ability, allocate a virtual budget to each resource demand agent, and predefine differentiated utility functions for each type of heterogeneous service flow. The resource trading market construction module is used to construct the edge gateway into a resource trading market with the resource demand agent as the main participant. The broadcast module is used to dynamically calculate the resource price based on the current network resource status information and broadcast the resource price to each of the resource demand agents. After receiving the resource price, each of the resource demand agents calculates the resource demand amount through a distributed optimization algorithm based on the corresponding utility function, the current virtual budget balance and the resource price, and submits a bidding request containing the resource demand amount to the edge gateway. An optimization module is used to update the resource price according to the bidding requests of each resource demand agent and rebroadcast it. Each resource demand agent, after receiving the updated resource price, adjusts its own resource demand according to the updated resource price and repeatedly executes the broadcast process until the bidding strategies of all resource demand agents reach a Nash equilibrium. The scheduling and transmission module is used to perform heterogeneous resource allocation and scheduling based on the resource demand of each resource demand agent in the Nash equilibrium state, and deduct the virtual budget corresponding to each resource demand agent to complete adaptive perception and collaborative transmission.

Images