Transmission power and task allocation optimization method and system under port cloud edge collaborative architecture

Abstract

The invention discloses a transmission power and task allocation optimization method and system under port cloud edge collaborative architecture, and belongs to the technical field of communication resource management. The method comprises the following steps: constructing a three-layer cloud edge collaborative architecture comprising port terminal equipment, edge nodes and a cloud center; the three-layer cloud edge collaborative architecture is mapped, a double-layer game optimization problem model is constructed, and the double-layer game optimization problem model comprises the steps that an upper-layer non-cooperative game model optimizes the transmitting power of terminal equipment to minimize energy consumption, and a lower-layer auction game model achieves resource bidding and task allocation of edge nodes; and performing double-layer joint solution on the double-layer game optimization problem model by using an improved non-dominated sorting genetic algorithm to realize global optimization of power control and task allocation. According to the method, the problems of unreasonable resource allocation and high energy consumption in a complex port environment are effectively solved, collaborative global optimization of transmission power and computing resources is realized, the energy consumption of equipment is remarkably reduced, and the system effectiveness and the communication reliability are improved.

Description

Technical Field

[0001] This invention belongs to the field of communication resource management technology, and in particular relates to a method and system for collaboratively optimizing the power and task allocation of multi-source data transmission in ports under a cloud-edge collaborative architecture. Background Technology

[0002] Smart ports rely on information and automation technologies to achieve intelligent management of business processes such as ship berthing and cargo loading and unloading. However, port communication systems face many challenges: frequent data interaction, strong equipment heterogeneity, and complex communication environments (such as salt spray, dust, and high humidity) lead to poor communication stability, high transmission latency, and high energy consumption. In addition, traditional network architectures have limited coverage and weak anti-interference capabilities, making it difficult to meet the needs of large-scale multi-terminal mobile communication and easily resulting in information silos, which restricts the improvement of the overall intelligence level of ports.

[0003] To address these issues, existing research has proposed several solutions. For example, integrating port equipment status and environmental data through an industrial internet platform to build an AGV cluster communication system based on protocols such as TCP / MQTT can enhance collaborative operation capabilities; utilizing 5G technology to build a remote detection and control platform can enhance real-time performance; and employing methods such as federated learning and edge computing can optimize communication efficiency and privacy protection.

[0004] Among them, the closest prior art to this application is a method for improving communication quality and resource utilization in port areas by jointly optimizing the relay position and communication block length of unmanned aerial vehicles (UAVs). This method optimizes the transmission channel by adjusting the spatial position of the UAV and balances latency and throughput with dynamic communication block allocation.

[0005] However, this existing technology focuses on transmission optimization and block scheduling at the UAV relay level. It does not achieve end-to-end collaborative optimization of terminal device transmission power and edge computing task resources under the cloud-edge collaborative architecture consisting of terminal devices, edge nodes and cloud centers. It also does not consider the dynamic impact of the complex weather environment in ports on the channel model, resulting in unreasonable resource allocation in the actual complex environment of ports. The overall system energy efficiency and reliability still have considerable room for improvement.

[0006] Therefore, designing a joint optimization method that can take into account transmission power control, task offloading decision and resource allocation strategy under the cloud-edge collaborative architecture has become a key issue in improving the performance of port communication systems. Summary of the Invention

[0007] This invention provides a method and system for optimizing transmission power and task allocation in a port cloud-edge collaborative architecture. By constructing a three-layer system model consisting of terminal devices, edge nodes, and a cloud center, and combining this with a communication rate and task latency model tailored to the specific meteorological environment of ports, a latency-based utility function is designed. An upper-layer non-cooperative game theory approach is used to optimize device transmission power, while a lower-layer auction game theory approach is used to allocate task resources, resulting in a joint solution. This invention effectively reduces device transmission energy consumption, improves system resource utilization efficiency and real-time service response, and is suitable for low-latency, high-reliability service scenarios such as port remote control.

[0008] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: On the one hand, this invention provides a method for optimizing transmission power and task allocation under a port cloud-edge collaborative architecture, including the following steps: Construct a three-layer cloud-edge collaborative architecture that includes port terminal equipment, edge nodes, and cloud center; The three-layer cloud-edge collaborative architecture is mapped to construct a two-layer game optimization problem model. The two-layer game optimization problem model includes: an upper-layer non-cooperative game model to optimize the transmission power of the terminal device to minimize energy consumption, and a lower-layer auction game model to realize resource bidding and task allocation of the edge nodes. An improved non-dominated sorting genetic algorithm is used to solve the two-layer game optimization problem model in a two-layer joint manner, so as to achieve global optimization of power control and task allocation. The improvement includes: introducing a decision space Euclidean distance term into the crowding calculation of the non-dominated sorting genetic algorithm, balancing the convergence of the target space and the diversity of the solution set through weight parameters; and dynamically adjusting the crossover rate according to the parent fitness and the overall level of the population.

[0009] Furthermore, the three-layer cloud-edge collaborative architecture includes: a terminal device layer, an edge node layer, and a cloud center layer; the terminal device layer consists of port operation equipment, which is responsible for collecting on-site data and generating computing tasks; the edge node layer consists of edge servers deployed on-site in the port, close to the terminal devices, providing low-latency data processing and task execution capabilities; the cloud center layer serves as the global control center, responsible for overall resource management, task scheduling, and collaborative decision-making.

[0010] Furthermore, the upper-level non-cooperative game model is defined as follows: In the formula, Indicates terminal device; Indicates terminal device The transmit power, whose strategy space is discretized as , For each terminal device Selectable transmit power; Indicates terminal device utility function , defined as the negative value of its own transmission energy consumption, i.e. The goal of any device is to maximize its own utility, which is equivalent to minimizing its own transmission power consumption; all terminal devices ; An initial transmit power is randomly assigned to each terminal device; in each iteration round In this process, each device updates its own power according to the following formula, assuming that the power of other devices remains constant: In the formula, the dynamic iteration step size is... The subgradient of the objective function and the step size parameter that decays with the number of iterations A joint decision.

[0011] Furthermore, in the lower-level auction game model, task unloading and resource allocation are modeled as a sealed-bid auction with the cloud center as the auctioneer and the edge nodes as bidders; the auctioneer is the cloud center; and the bidders are each edge node. The target is the computing task generated by the terminal device; the bidding strategy is for each edge node. For the task Calculate the optimal bid price it can offer. The price is its own utility function. The maximum point; by solving the first-order optimality condition. ,have to: In the formula, For parameters, For computing resources, The task computation is calculated as follows: After the cloud center collects sealed bids from all edge nodes, it determines the task allocation result based on the principle of maximizing the total system utility, and assigns the task to the combination of edge nodes that can bring the greatest system utility.

[0012] Furthermore, the upper-layer non-cooperative game model calculates the transmission rate based on the Rayleigh channel communication model that incorporates port-specific weather attenuation factors, in order to optimize the transmission power of the terminal equipment and affect the transmission delay and the overall system utility; The lower-level auction game model uses the transmission rate of the Rayleigh channel communication model that incorporates the port-specific weather attenuation factor to calculate task delay, thereby ensuring the communication reliability of resource allocation in maximizing the total system utility. The Rayleigh channel communication model that incorporates port-specific meteorological attenuation factors includes: A Rayleigh fading channel is used to simulate the wireless transmission environment of a port; terminal equipment. To edge nodes The channel gain of the access link is defined as: In the formula, Indicates the reference distance And the channel power gain when the transmit power is 1W, Indicates terminal device To edge nodes distance, Indicates the fog attenuation factor; meteorological attenuation factor Based on visibility With wavelength Calculations show that In the formula, Indicates wavelength. This indicates a visibility reference wavelength of 550nm. Indicates visibility, Represents the scattering coefficient. Visibility The decision is expressed as follows: ; Task Request The time required for data transfer and unloading is: Task Request The execution time at the edge node is: ;in, Represents edge nodes Scheduling to task requests Computing power; total latency is defined as: ; Based on Shannon's theorem, from the terminal device To edge nodes uplink speed Represented as: ; In the formula, Represents the bandwidth of the uplink system. Indicates from port equipment To edge nodes The transmission power, This represents the Gaussian white noise power of the uplink system. This represents the channel gain from port equipment m to edge node n.

[0013] Furthermore, the method of using an improved non-dominated sorting genetic algorithm to solve the two-level game optimization problem model in a two-level joint manner includes: Initialize the population, including upper-level power variables and lower-level allocation variables; The evaluation involves two objectives: the upper-level non-cooperative game model aims to minimize energy consumption, while the lower-level auction game model aims to maximize the total system utility. Non-dominated sorting and improved crowding calculation; Selection, crossover, and mutation generate offspring; Merging populations and selecting the best to form the next generation; Iterate until the termination condition is met, then output the Pareto front solution set.

[0014] The improved non-dominated sorting genetic algorithm incorporates the Rayleigh channel communication model with port-specific weather attenuation factor as a physical constraint in the two-layer joint solution, simulating the impact of the port environment on power and allocation decisions.

[0015] Furthermore, the improved congestion calculation formula is as follows: ; In the formula, The calculation is the Euclidean distance between individual i and other individuals j in the population in the decision variable space; the average of the reciprocals of this term reflects the uniqueness of individual i in the decision space; the greater the difference from the surrounding individuals, the smaller this value and the larger its reciprocal, thus obtaining a higher reward; It is a weighting parameter used to balance the crowding of the target space and the diversity of the decision space; The adaptive crossover probability formula is: ; In the formula, The crossover probability; It is the parent individual with higher fitness among the two parent individuals involved in the crossover; and These are the worst fitness and average fitness in the current population, respectively. Furthermore, the total utility of the port cloud-edge collaborative architecture communication system is: In the formula, W Total system utility; Utility function for edge nodes: ; For cloud center utility functions: In the formula, the cloud center utility function is the net benefit of the cloud center, which includes the sum of delay benefits and payment savings; where It is a delayed benefit. It is the cloud center's valuation of the task. This is the actual price paid to the edge nodes; It is a reward income. It is the processing cost; To save on payments.

[0016] In another aspect, the present invention also provides a transmission power and task allocation optimization system under a port cloud-edge collaborative architecture, comprising: The terminal device module is used for data acquisition, task generation, and power adjustment. The edge node module is used to receive and process tasks and participate in resource bidding; The cloud center module is used for global resource management, task scheduling, and auction coordination. The optimization algorithm module is used to execute the transmission power and task allocation optimization methods under the port cloud-edge collaborative architecture described above.

[0017] Furthermore, the system also includes a communication environment perception module, which is used to acquire port meteorological data in real time and dynamically update the meteorological attenuation factor in the communication rate model accordingly.

[0018] The technological advancements achieved by this invention due to the adoption of the above technical solutions are as follows: This invention fundamentally improves the accuracy and environmental adaptability of system modeling by constructing a three-layer cloud-edge collaborative architecture and a precise communication model that fits the complex meteorological environment of ports, laying a reliable foundation for resource optimization. Secondly, the innovative economic incentive model based on delay utility directly links task completion time, resource consumption, and economic benefits, effectively stimulating the participation of edge nodes and achieving an organic balance between system performance and cost-effectiveness. Finally, through a two-layer optimization framework of upper-layer non-cooperative game theory and lower-layer auction game theory, and using an improved i-NSGA-II algorithm for joint solution, collaborative global optimization of transmission power and computing resources is achieved. Ultimately, simulations demonstrate a significant improvement in the communication efficiency and system reliability of port remote control services. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Figure 1 This is a model diagram of a two-layer game system with a cloud-edge collaborative architecture in an embodiment of the present invention; Figure 2 This is a flowchart illustrating a method for optimizing transmission power and task allocation under a port cloud-edge collaborative architecture, as described in an embodiment of the present invention. Detailed Implementation It should be noted that the terms "comprising" and "having" and any variations thereof in the specification, claims and accompanying drawings of this invention are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or units that are not explicitly listed or that are inherent to such processes, methods, products or devices.

[0020] This invention provides a method for optimizing transmission power and task allocation under a port cloud-edge collaborative architecture. It constructs a three-layer collaborative architecture (edge, cloud, and terminal) to replace single-layer UAV relay optimization. A port meteorological attenuation factor is introduced into the communication model to improve environmental adaptability. A two-layer coupling mechanism of upper-layer power game and lower-layer resource auction is adopted to achieve global collaborative optimization. Terminal device transmission energy consumption is significantly reduced, and the overall system efficiency is greatly improved. The economically incentive-based resource allocation mechanism effectively avoids resource congestion and idleness. The system's adaptability to complex port meteorological environments and communication reliability are significantly enhanced.

[0021] like Figure 1 As shown, a method for optimizing transmission power and task allocation under a port cloud-edge collaborative architecture includes the following steps: Step 1: Construct a three-layer cloud-edge collaborative architecture that includes port terminal equipment, edge nodes, and cloud center.

[0022] Specifically, a collaborative architecture consisting of a port terminal equipment layer, an edge node layer, and a cloud center layer is constructed: [Definition] A collection of terminal layer devices. This refers to the set of edge nodes in the edge layer of the system. Assume a time slot generates a set of task requests using this set. express, In the formula, This represents a task request generated by each port device in a single time slot. , in Indicates the size of the task data in bytes; This indicates the amount of computing resources required to perform a single task, expressed in cycles per bit. and These are the ideal latency and the maximum tolerable latency threshold, respectively. Indicates the priority of different requests.

[0023] Step 2: Design the channel model for the three-layer cloud-edge collaborative architecture; Within the constructed three-layer cloud-edge collaborative architecture, a Rayleigh fading channel is employed to simulate the complex wireless transmission environment of a port. The Rayleigh channel communication model is used to calculate the transmission rate, thereby influencing task latency assessment and game-theoretic optimization decisions.

[0024] terminal equipment To edge nodes The channel gain of the access link is defined as: (1) In the formula, Indicates the reference distance And the channel power gain when the transmit power is 1W, Indicates terminal device To edge nodes distance, Indicates the fog attenuation factor; meteorological attenuation factor This factor is a key innovation, as its calculation incorporates visibility. With wavelength : (2) In the formula, Indicates wavelength. This indicates a visibility reference wavelength of 550nm. Indicates visibility, Represents the scattering coefficient. Visibility The decision is expressed as follows: (3) Total task request latency: Task request The time required for data transfer and unloading is: ; ask The execution time at the edge node is: In the formula Represents edge nodes Dispatch to request Its computing power.

[0025] Port equipment unloading request The total delay is defined as: .

[0026] Uplink transmission rate: Based on Shannon's theorem, from port equipment To edge nodes uplink speed It can be represented as: (4) In the formula, Represents the bandwidth of the uplink system. Indicates from port equipment To edge nodes The transmission power, This represents the Gaussian white noise power of the uplink system. Indicates from port equipment m To edge nodes n Channel gain.

[0027] Step 3: Design the overall utility of the three-layer cloud-edge collaborative architecture communication system; Specifically, based on the utility of latency, utility functions for edge nodes and the cloud center are designed. A reward mechanism is introduced to incentivize edge nodes to participate in task processing, and a cost function reflects the energy consumption caused by resource occupation. The latency utility function converts task completion time into economic benefits, ensuring low latency is prioritized. To incentivize edge node participation, a reward function is designed, whose benefits saturate as the bidding price increases. Simultaneously, a cost function is introduced to quantify the energy consumption and resource depletion generated by the task. Ultimately, the utility of edge nodes is the sum of these three factors, while the utility of the cloud center balances latency benefits and payment costs, thus balancing system performance and economic efficiency.

[0028] First, an incentive-compatible economic model is constructed. By designing a reasonable utility function, edge nodes are guided to actively participate in task processing, while resource costs are accurately quantified to achieve a balance between system performance and economic benefits. The utility function is then delayed. The core idea is to transform the physical indicator of task completion time into quantifiable economic benefits, thereby embedding quality of service (QoS) requirements into the optimization goal.

[0029] Design of delayed utility functions: (5) The average delay threshold The calculation formula is: (6) Incentivize low latency: at ideal latency The maximum benefit (normalized to 1) can be obtained by completing the task within the specified time.

[0030] Diminishing returns: delay in and During this period, revenue smoothly decreased with increasing latency, simulating the decay of user satisfaction.

[0031] Hard constraint: Exceeding the maximum tolerable delay The revenue is zero, ensuring that the system meets basic service quality requirements.

[0032] Edge Node Reward Mechanism Design: To incentivize edge nodes to contribute their computing resources, the cloud center designed a reward function. As a reward for the edge node processing tasks: (7) In the formula, It is an edge node To handle tasks The submitted bid price, This is the adjustment factor. The function exhibits an S-shaped growth. When the bid price... When the price is low, earnings grow faster; when... When prices are too high, returns tend to saturate. This design prevents edge nodes from demanding exorbitant prices, encourages them to submit reasonable bids, and ensures the system's economic efficiency.

[0033] Cost function This is used to quantify the actual losses incurred when edge nodes process tasks, mainly including computing power consumption, equipment depreciation, and maintenance costs.

[0034] (8) In the formula, This is the cost coefficient. This represents the computational load of the task.

[0035] This function shows that costs increase superlinearly with increasing bid prices and task complexity, prompting edge nodes to rationally assess their own resource status when bidding.

[0036] Based on the above components, the final utility functions of edge nodes and cloud centers are constructed to form a two-sided market model.

[0037] Edge node utility function : (9) In the formula, this function represents the net revenue of the edge node, which is the sum of latency revenue and reward revenue minus processing cost. The goal of the edge node is to maximize... This drives it to offer reasonable prices to balance revenue and costs while ensuring low latency.

[0038] Cloud center utility function : (10) In the formula, this function represents the net benefit of the cloud center, which is the sum of delay benefits and payment savings. It is the cloud center's valuation of the task (target payment price). This is the actual price paid to the edge nodes. The goal of the cloud center is to maximize... The goal is to obtain high-quality (low-latency) task processing services at the lowest possible price.

[0039] The utility of the port cloud-edge collaborative architecture communication system is the sum of the utility of both parties: (11) In the formula, W This represents the total utility of the system.

[0040] This economic incentive model provides a decision-making basis for subsequent two-level game optimization by transforming latency, rewards, and costs into quantifiable utility functions. Specifically, in the lower-level auction game, the bidding strategies of edge nodes are directly based on maximizing their utility functions, while the allocation rules of the cloud center use the total system utility W as the optimization objective, thereby ensuring the economic incentive compatibility of resource allocation. Simultaneously, upper-level power optimization affects transmission latency, which in turn feeds back into the latency utility function, achieving synergy between transmission and computation.

[0041] Step 4: Construct a two-layer game optimization problem model. The two-layer game optimization problem model includes: an upper-layer non-cooperative game model to optimize the transmission power of terminal equipment to minimize energy consumption, and a lower-layer auction game model to realize resource bidding and task allocation for edge nodes. The improved non-dominated sorting genetic algorithm i-NSGA-II is used to solve the two-layer problem together to achieve global optimization of power control and task allocation.

[0042] This step aims to construct a coupled, two-layer game theory framework to collaboratively optimize transmission power and task resource allocation, ultimately achieving the global objectives of minimizing system energy consumption and maximizing total utility. The upper-layer non-cooperative game theory model calculates the transmission rate based on the Rayleigh channel communication model to optimize transmission power and influence transmission delay and total system utility. The lower-layer auction game theory model uses the transmission rate from the Rayleigh channel communication model to calculate task delay, thereby ensuring communication reliability in resource allocation while maximizing total system utility.

[0043] (1) Upper-level optimization: Power control based on non-cooperative game theory: The upper-level game theory focuses on the distributed power competition among terminal devices, and is modeled as a non-cooperative game problem: (12) In the formula, all port terminal equipment ( Strategy set: per device Selectable transmit power Its strategy space is discretized as Each device utility function Defined as the negative of its own transmission energy consumption, i.e. The goal of a device is to maximize its own utility, which is equivalent to minimizing its own transmission energy consumption.

[0044] Game process and Nash equilibrium solution: Initialization: Randomly assign an initial transmit power to each device.

[0045] Iterative update: in each iteration round In this process, each device updates its own power according to the following formula, assuming that the power of other devices remains constant: (13) In the formula, the dynamic iteration step size The subgradient of the objective function and the step size parameter that decays with the number of iterations A joint decision.

[0046] (2) Lower-level optimization: Resource allocation based on auction game theory: In the lower level, task unloading and resource allocation are modeled as a sealed-bid auction with the cloud center as the auctioneer and the edge nodes as bidders. The auctioneer is the cloud center. The bidders are each edge node. The target is the computing task generated by the terminal device. The bidding strategy is per edge node. For the task It will calculate the best bid price it can offer. The price is its own utility function. The maximum point (usually related to the total utility of the system or its own benefit). This is achieved by solving the first-order optimality conditions. We can obtain: (14) In the formula, For parameters, For computing resources, This represents the computational load of the task.

[0047] Allocation Rules: After the cloud center collects sealed bids from all edge nodes, allocation is based on the principle of maximizing total system utility (i.e., maximizing...). To determine the task allocation result And assign the task to the edge node combination that brings the greatest system utility. (3) Joint solution: multi-objective optimization algorithm based on i-NSGA-II; the improvements of i-NSGA-II include: (1) The crowding degree calculation introduces the Euclidean distance term of the decision variable space to calculate the individual Other individuals The average reciprocal distance is used to enhance the uniqueness of the decision space, and the crowding of the target space and the diversity of the decision space are balanced by the weight parameter λ to improve the uniform distribution of the Pareto front; (2) the crossover probability is adaptively adjusted according to; It is the parent individual with higher fitness among the two parent individuals involved in the crossover; and Pc is dynamically calculated using the worst and average fitness values ​​in the current population, ensuring efficient exploration of diversity in the early stages and efficient convergence in the later stages. Furthermore, i-NSGA-II incorporates the Rayleigh channel communication model as a physical constraint in its two-layer joint solution, simulating the impact of the port environment on power and allocation decisions.

[0048] like Figure 2As shown, the two-layer joint solution process includes: initializing the population (including upper-layer power variables and lower-layer allocation variables); evaluating the two-layer objectives (upper-layer minimizing energy consumption, lower-layer maximizing total system utility); performing non-dominated sorting and improved crowding calculation; generating offspring through selection, crossover (adaptive), and mutation; merging the populations and performing elite selection to form the next generation; iterating to the termination condition and outputting the Pareto front solution set.

[0049] Improved congestion calculation formula: (15) In the formula, This calculates the Euclidean distance between individual i and other individuals j in the population across the decision variable space. The average of the reciprocals of this term reflects the uniqueness of individual i within the decision space. The greater the difference from surrounding individuals (the more unique the characteristics of the solution), the smaller this value, and the larger its reciprocal, resulting in a higher reward. It is a weighting parameter used to balance the crowding of the target space and the diversity of the decision space.

[0050] Adaptive crossover probability formula: (16) In the formula, is used to dynamically adjust the crossover probability. This is a manifestation of self-adaptation. It is the parent individual with higher fitness among the two individuals participating in the crossover (for minimization problems, the smaller the fitness value, the better). and These are the worst fitness and average fitness in the current population, respectively.

[0051] The advantages of this two-layer game theory scheme are: the upper-layer power control is distributed, reducing the computational burden on the central node; the lower-layer resource allocation is centralized, ensuring global efficiency. The game theory model naturally introduces an incentive mechanism, ensuring that the selfish behaviors of terminal devices and edge nodes ultimately lead to the achievement of the system's global goal. By jointly solving the two-layer game using the i-NSGA-II algorithm, the drawback of hierarchical optimization potentially getting trapped in local optima is overcome, achieving true synergistic optimization of power and resources, and significantly improving the overall system performance.

[0052] The cloud-edge collaborative architecture provides a computing framework that combines distributed and centralized approaches through a three-layer structure of terminals, edge computing, and cloud centers. A two-layer game theory model maps this framework through an upper-layer non-cooperative game (terminal power control) and a lower-layer auction game (edge ​​resource allocation), achieving incentive-compatible resource management. The combination of these two approaches forms a closed-loop collaboration, and the inclusion of a meteorological attenuation factor enhances adaptability to the complex port environment, improving system robustness and thus addressing issues of irrational resource allocation and high interference.

[0053] In summary, this invention fundamentally improves the accuracy and environmental adaptability of system modeling by constructing a three-layer cloud-edge collaborative architecture and a precise communication model that fits the complex meteorological environment of ports, laying a reliable foundation for resource optimization. Secondly, the innovative economic incentive model based on delay utility directly links task completion time, resource consumption, and economic benefits, effectively stimulating the participation of edge nodes and achieving an organic balance between system performance and cost-effectiveness. Finally, through a two-layer optimization framework of upper-layer non-cooperative game theory and lower-layer auction game theory, and using the improved i-NSGA-II algorithm for joint solution, collaborative global optimization of transmission power and computing resources is achieved. Ultimately, simulations demonstrate a significant improvement in the communication efficiency and system reliability of port remote control services.

[0054] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for optimizing transmission power and task allocation under a port cloud-edge collaborative architecture, characterized in that: Includes the following steps: Construct a three-layer cloud-edge collaborative architecture that includes port terminal equipment, edge nodes, and cloud center; The three-layer cloud-edge collaborative architecture is mapped to construct a two-layer game optimization problem model. The two-layer game optimization problem model includes: an upper-layer non-cooperative game model to optimize the transmission power of the terminal device to minimize energy consumption, and a lower-layer auction game model to realize resource bidding and task allocation of the edge nodes. An improved non-dominated sorting genetic algorithm is used to solve the two-layer game optimization problem model in a two-layer joint manner, so as to achieve global optimization of power control and task allocation. The improvement includes: introducing a decision space Euclidean distance term into the crowding calculation of the non-dominated sorting genetic algorithm, balancing the convergence of the target space and the diversity of the solution set through weight parameters; and dynamically adjusting the crossover rate according to the parent fitness and the overall level of the population.

2. The method for optimizing transmission power and task allocation under a port cloud-edge collaborative architecture according to claim 1, characterized in that, The three-layer cloud-edge collaborative architecture includes: a terminal device layer, an edge node layer, and a cloud center layer; the terminal device layer consists of port operation equipment, which is responsible for collecting on-site data and generating computing tasks; the edge node layer consists of edge servers deployed on-site in the port, close to the terminal devices, providing low-latency data processing and task execution capabilities; the cloud center layer serves as the global control center, responsible for overall resource management, task scheduling, and collaborative decision-making.

3. The method for optimizing transmission power and task allocation under a port cloud-edge collaborative architecture according to claim 1, characterized in that, The upper-level non-cooperative game model is defined as follows: In the formula, Indicates terminal device; Indicates terminal device The transmit power, whose strategy space is discretized as , For each terminal device Selectable transmit power; Indicates terminal device utility function , defined as the negative value of its own transmission energy consumption, i.e. The goal of any device is to maximize its own utility, which is equivalent to minimizing its own transmission power consumption; all terminal devices ; Randomly assign an initial transmit power to each terminal device; In each iteration round In this process, each device updates its own power according to the following formula, assuming that the power of other devices remains constant: In the formula, the dynamic iteration step size is... The subgradient of the objective function and the step size parameter that decays with the number of iterations A joint decision.

4. The method for optimizing transmission power and task allocation under a port cloud-edge collaborative architecture according to claim 1, characterized in that, In the lower-level auction game model, task unloading and resource allocation are modeled as a sealed first-price auction with the cloud center as the auctioneer and the edge nodes as bidders; the auctioneer is the cloud center; and the bidders are the various edge nodes. The subject matter is the computing task generated by the terminal device; The bidding strategy is for each edge node For the task Calculate the optimal bid price it can offer. The price is its own utility function. The maximum point; by solving the first-order optimality condition. ,have to: In the formula, For parameters, For computing resources, The task computation is calculated as follows: After the cloud center collects sealed bids from all edge nodes, it determines the task allocation result based on the principle of maximizing the total system utility, and assigns the task to the combination of edge nodes that can bring the greatest system utility.

5. A method for optimizing transmission power and task allocation under a port cloud-edge collaborative architecture according to claim 3 or 4, characterized in that, The upper-layer non-cooperative game model calculates the transmission rate based on the Rayleigh channel communication model that incorporates port-specific weather attenuation factors, in order to optimize the transmission power of the terminal equipment and affect transmission delay and overall system utility. The lower-level auction game model uses the transmission rate of the Rayleigh channel communication model that incorporates the port-specific weather attenuation factor to calculate task delay, thereby ensuring the communication reliability of resource allocation in maximizing the total system utility. The Rayleigh channel communication model that incorporates port-specific meteorological attenuation factors includes: A Rayleigh fading channel is used to simulate the wireless transmission environment of a port; terminal equipment. To edge nodes The channel gain of the access link is defined as: In the formula, Indicates the reference distance And the channel power gain when the transmit power is 1W, Indicates terminal device To edge nodes distance, Indicates the fog attenuation factor; meteorological attenuation factor Based on visibility With wavelength Calculations show that In the formula, Indicates wavelength. This indicates a visibility reference wavelength of 550nm. Indicates visibility, Represents the scattering coefficient. Visibility The decision is expressed as follows: ; Task Request The time required for data transfer and unloading is: Task Request The execution time at the edge node is: ;in, Represents edge nodes Scheduling to task requests Computing power; total latency is defined as: ; Based on Shannon's theorem, from the terminal device To edge nodes uplink speed Represented as: ； In the formula, Represents the bandwidth of the uplink system. Indicates from port equipment To edge nodes The transmission power, This represents the Gaussian white noise power of the uplink system. This represents the channel gain from port equipment m to edge node n.

6. The method for optimizing transmission power and task allocation under a port cloud-edge collaborative architecture according to claim 5, characterized in that, The method of using an improved non-dominated sorting genetic algorithm to solve the two-level game optimization problem model in a joint two-level solution includes: Initialize the population, including upper-level power variables and lower-level allocation variables; The evaluation involves two objectives: the upper-level non-cooperative game model aims to minimize energy consumption, while the lower-level auction game model aims to maximize the total system utility. Non-dominated sorting and improved crowding calculation; Selection, crossover, and mutation generate offspring; Merging populations and selecting the best to form the next generation; Iterate until the termination condition is met, then output the Pareto front solution set. The improved non-dominated sorting genetic algorithm incorporates the Rayleigh channel communication model with port-specific weather attenuation factor as a physical constraint in the two-layer joint solution, simulating the impact of the port environment on power and allocation decisions.

7. A method for optimizing transmission power and task allocation under a port cloud-edge collaborative architecture according to claim 1 or 6, characterized in that, The improved congestion calculation formula is as follows: ； In the formula, The calculation is the Euclidean distance between individual i and other individuals j in the population in the decision variable space; the average of the reciprocals of this term reflects the uniqueness of individual i in the decision space; the greater the difference from the surrounding individuals, the smaller this value and the larger its reciprocal, thus obtaining a higher reward; It is a weighting parameter used to balance the crowding of the target space and the diversity of the decision space; The adaptive crossover probability formula is: ; In the formula, The crossover probability; It is the parent individual with higher fitness among the two parent individuals involved in the crossover; and These are the worst fitness and average fitness in the current population, respectively.

8. The method for optimizing transmission power and task allocation under a port cloud-edge collaborative architecture according to claim 6, characterized in that, The overall utility of the port cloud-edge collaborative architecture communication system is: In the formula, W Total system utility; Utility function for edge nodes: ; For cloud center utility functions: In the formula, the cloud center utility function is the net benefit of the cloud center, which includes the sum of delay benefits and payment savings; where It is a delayed benefit. It is the cloud center's valuation of the task. This is the actual price paid to the edge nodes; It is a reward income. It is the processing cost; To save on payments.

9. A transmission power and task allocation optimization system under a port cloud-edge collaborative architecture, characterized in that, include: The terminal device module is used for data acquisition, task generation, and power adjustment. The edge node module is used to receive and process tasks and participate in resource bidding; The cloud center module is used for global resource management, task scheduling, and auction coordination. An optimization algorithm module is used to execute the transmission power and task allocation optimization method under the port cloud-edge collaborative architecture as described in any one of claims 1 to 8.

10. The transmission power and task allocation optimization system under a port cloud-edge collaborative architecture according to claim 9, characterized in that, The system also includes a communication environment perception module, which is used to acquire port meteorological data in real time and dynamically update the meteorological attenuation factor in the communication rate model accordingly.

Images