A multi-stage security task offloading optimization method for low-altitude intelligent network

Abstract

The application discloses a multi-stage security task offloading optimization method for low-altitude intelligent networks, and belongs to the technical field of low-altitude intelligent network edge computing. It aims to solve the problems of exploration and development imbalance, insufficient optimization precision and lack of security guarantee of the traditional task offloading method in the low-altitude dynamic heterogeneous network environment. The method comprises the following steps: constructing a network architecture containing communication, calculation and security models based on low-altitude intelligent network device basic information; configuring a multi-constraint optimization problem; after generating an initial population, searching for a global optimal solution by using an improved escape optimization algorithm; and performing calculation energy consumption optimization configuration according to the optimal solution. Through the combination of algorithm innovation and security constraint design, the task offloading efficiency, resource configuration rationality and data transmission security are synergistically improved, and the low-altitude intelligent network dynamic heterogeneous characteristics are adapted.

Description

Technical Field

[0001] This invention relates to the fields of intelligent optimization and edge computing technology, specifically a multi-stage safety task offloading optimization method for low-altitude intelligent networks. Background Technology

[0002] With the continuous development of Low Altitude Intelligent Networks (LAIN), the computing tasks generated by terminal devices are characterized by high frequency, diversity, and high real-time requirements. Due to the limited computing power of the devices themselves, task offloading mechanisms are often used to migrate computationally intensive tasks to edge servers or base stations for execution, in order to reduce local load and energy consumption and improve response speed.

[0003] Mobile edge computing (MEC) provides the technological foundation for task offloading, but in complex wireless environments, such as non-line-of-sight transmission and low-altitude scenarios with severe obstruction and interference, communication quality cannot be guaranteed, directly affecting the stability and reliability of offloading tasks. Therefore, a collaborative architecture integrating intelligent reflective surfaces (IRS) and unmanned aerial vehicles (UAVs) is gradually becoming the mainstream solution. IRS can significantly improve wireless link quality by adjusting the phase angle of its reflective units; UAVs have flexible deployment characteristics and can be equipped with IRS to achieve dynamic aerial relay, assisting in channel enhancement and resource scheduling.

[0004] Under this architecture, the task offloading problem exhibits characteristics of multivariability, high dimensionality, nonlinearity, and strong coupling. The optimization objective is typically to minimize system energy consumption, while simultaneously satisfying multiple conditions such as computational latency, security constraints, channel allocation, and power limitations. Traditional heuristic optimization methods, such as particle swarm optimization and whale optimization, while possessing certain search capabilities, often suffer from problems such as getting trapped in local optima and insufficient global search ability, making them difficult to adapt to complex scheduling scenarios.

[0005] Escape behavior mechanism, as an intelligent optimization idea derived from population risk avoidance strategy, can dynamically adjust the exploration intensity during population evolution by simulating the behavioral shifts of individuals in different states (such as calm, consistency, and panic), effectively improving the search depth and diversity of the algorithm, and has good global optimization potential.

[0006] Furthermore, while the existing escape optimization algorithm, the whale algorithm, has a fast convergence speed, it is insufficient in global search capabilities, which may limit its effectiveness in practical applications. Summary of the Invention

[0007] To address the shortcomings of existing technologies, this invention provides a multi-stage safety task offloading optimization method for low-altitude intelligent networks, aiming to solve the problems in the background technology.

[0008] To achieve the above objectives, the present invention provides the following technical solution: a multi-stage security task offloading optimization method for low-altitude intelligent networks, comprising the following steps:

[0009] Step S1: Construct a network architecture based on the basic device information of the low-altitude intelligent network in the edge computing system. The network architecture includes a communication model, a computing model, and a security model; configure the optimization problem under multiple constraints in the network architecture.

[0010] Step S2: Generate an initial population based on the optimization problem, and use an improved escape optimization algorithm to search the initial population to obtain the target population, and output the global optimal solution in the target population; the improved escape optimization algorithm introduces crossover and mutation operations and a constraint penalty function containing task execution delay and security service cost on the basis of the original escape optimization algorithm; the crossover and mutation operation is used to improve population diversity and escape local optima; the constraint penalty function is used to calculate fitness and update the population ranking; specifically: the improved escape optimization algorithm is used for optimization search, first initialize the population and iteration-related parameters, calculate fitness and rank the population; in each iteration, the strategy is adjusted according to the iteration progress: when the iteration progress No more than At that time, calculate the fear index. For the number of iterations, To maximize the number of iterations, reorder and group the population into calm, conform, and panic groups, and perform update operations for the calm, conform, and panic groups; when the iteration progress is... arrive During this period, the development phase begins, where the population is adjusted and optimized through local exploration; when the iteration progress exceeds... When the crossover and mutation phase is initiated, an adaptive single-point crossover and mutation operation is performed during the crossover and mutation phase. After each update, the fitness is evaluated and the best individual is updated until the set number of iterations is completed. This updates the information of the best individual in the current population and outputs the position information encoding of all individuals in the population.

[0011] Step S3: Calculate the energy consumption optimization configuration based on the global optimal solution.

[0012] Furthermore, the specific process for obtaining basic equipment information for the low-altitude intelligent network is as follows: [The process involves obtaining...] The index set of each device is ,in, Indicates the index of the device. Indicates the total number of devices; The index set of micro base stations is The index set of macro base stations is The index set of all base stations is ,in, Indicates the index of the micro base station. This indicates the number of micro base stations, and s represents the index of the base station in the UAV-coordinated intelligent reflective surface-assisted mobile edge computing system. Indicates the total number of base stations; The index set of intelligent reflective surfaces coordinated by drones is ,in, The index representing the intelligent reflective surface of drone collaboration. This indicates the number of intelligent reflective surfaces used in drone collaboration; a single intelligent reflective surface used in drone collaboration contains... One reflective element, The set of indices for each reflective element is ,in, Indicates the index of the reflective element in the intelligent reflective surface of the drone collaboration. This represents the number of reflective elements in a drone-coordinated intelligent reflective surface; the... The reflection coefficient matrix of the reflective element in the intelligent reflective surface coordinated by multiple drones is denoted as follows: ,in, Indicates the first The first reflective element in the... The reflectivity of a smart reflective surface coordinated by multiple drones Next, the physical locations of all micro base stations are obtained, and the data is used based on the physical locations of all micro base stations. The clustering algorithm divides all micro base stations into groups. There are clusters; in each cluster there are Each sub-channel is used by the micro base stations, devices, and drones in the cluster to form a collaborative intelligent reflector. When a device transmits to the intelligent reflector formed by the base station and the drone, the sub-channel selected is the same. The set of indices of the sub-channels is denoted as ,in, Indicates the index of the sub-channel; the task passes through the first... The device was passed by the first The sub-channel is transmitted to the first The channel gain of the intelligent reflector coordinated by multiple drones is denoted as... The task passed the first A drone-integrated intelligent reflector via the first The sub-channel is transmitted to the first The channel gain of each micro base station is denoted as The task passed the first The micro base station transmits back to the first The channel gain of each device is denoted as The power of Gaussian white noise is denoted as .

[0013] Furthermore, consider the optimization problem under multiple constraints in the configuration network architecture:

[0014] ;

[0015] In the formula, The goal of the energy minimization optimization problem is to minimize the total computational energy consumption of all device tasks in a drone-cooperative intelligent reflector-assisted mobile edge computing system. ; An indexed state set indicating whether the device has decided to establish a connection with the base station. , For the first The device and the first The index of the status indicators for establishing connections between base stations. This indicates the cryptographic algorithm selected when the device first transmits the task to the drone-coordinated intelligent reflective surface, and then transmits the task to the base station through the drone-coordinated intelligent reflective surface. The set of index states, , Indicates the first The tasks transmitted by each device sequentially pass through the intelligent reflective surface coordinated by the drone and the first... The selected base station One cryptographic algorithm; This represents the indexed state set of state variables for selecting sub-channels as connection nodes when the device transmits tasks to the UAV-coordinated intelligent reflector and when the UAV-coordinated intelligent reflector transmits tasks to the base station. , Indicates the first The device will transmit the task to the first... When selecting a smart reflector in collaboration with multiple drones, choose the first one. Each sub-channel serves as a state variable for the connected nodes. Indicates the first A smart reflective surface, coordinated by multiple drones, transmits the mission to the first... When selecting the first base station Each sub-channel serves as a state variable for the connection node; This represents the set of indexes representing the allocation of transmit power for all devices. , Indicates the first The transmission power of each device Indicates the maximum transmission power; This is a set of indexes indicating whether the device is associated with a smart reflective surface selected for drone collaboration. , Indicates the first The device is associated with the first Decision-making index of intelligent reflective surface in collaboration with drones; Indicates the first Task processing latency of individual devices Indicates the first Maximum execution time for each device; This represents the reflection coefficient matrix of the reflective elements in the intelligent reflective surface used in drone collaboration. , Indicates the first The first reflective element in the... The reflectivity of a smart reflective surface coordinated by multiple drones; A set of indices representing the bit size of the task data offloaded from the device to the base station. , Indicates the first The device was offloaded to the micro base station. Task data bit size, Total amount of task data; This indicates task data that represents the offloading of tasks from micro base stations to macro base stations. A set of indices of bit size, , Indicates the first The tasks that each device needs to handle come from the micro base station. Task data unloaded to macro base station The size of the bits; Indicates the first The task data that each device needs to process bit size, Indicates the first The device transmits to the first Task data of each base station The size of the bits; Indicates the first The security cost of each device Indicates the first The maximum cost constrained by each device This represents the lower boundary value for data unloading; These represent the first through the fourteenth constraints, respectively.

[0016] Furthermore, the specific process of step S2 is as follows:

[0017] Step S2.1: Initialize the maximum number of iterations for the improved escape optimization algorithm and the current iteration number Set to 1;

[0018] Step S2.2: Let the number of individuals in the population be... Any individual in the population uses If expressed as such, then the population is For any given individual, encode it using the parameter variables of the optimization problem; index the state set. Encoding into population , Represents an individual The Middle Indexed state index of the status indicators for establishing a connection between a device and a base station; indexed state set Encoding into population , Represents an individual The Middle A set of cryptographic algorithms selected by each device; an index state set Encoding into population , Represents an individual The Middle The index of the sub-channel selected by each device; index set Encoding into population , Represents an individual The Middle Transmit power of individual devices; index set Encoding into population , Represents an individual The Middle The index of the drone collaborative intelligent reflector selected by each device; index set Encoding into population , Indicates the first The reflection coefficient of the reflective element in the collaborative intelligent reflective surface of the drone associated with each device; index set Encoding into population , Represents an individual The middle task starts from the first The amount of tasks offloaded from a device to a micro base station; index set Encoding into population , Represents an individual The Middle The amount of data that each device needs to process from micro base stations to macro base stations;

[0019] Step S2.3: Initialize the population and establish the individuals in the population. Fitness function:

[0020] ;

[0021] In the formula, For the first Penalty factor for the maximum execution time of a task on a single device; For the first The penalty factor for the maximum security cost of each device; Represents the fitness function; This represents the total energy consumption of all devices;

[0022] Step S2.4: Use the fitness function to calculate the fitness of all individuals in the population. The fitness value of each individual, and the individual with the highest fitness value. As the best individual in the overall historical context;

[0023] Step S2.5: First, determine the iteration progress. Is it less than or equal to? If the iteration progress Less than or equal to Then execute the update operations for the calm group, conformity group, and panic group; if If so, then proceed to the development phase to update the individual; if Then, proceed to the crossover and mutation phase to perform adaptive single-point crossover and mutation operations; if the current iteration number is... If so, then the codes of all individuals will be output.

[0024] Furthermore, the update process for the calming group is as follows:

[0025] ;

[0026] ;

[0027] ;

[0028] In the formula, To increase with the current iteration number The probability of panic due to change; This indicates a randomly selected position within the calming group; This indicates a randomly generated location within the range of the calming group. Parameters indicating adjustments to an individual's position; Indicates the first Individual during round iteration The Middle An index of the status indicators for a device establishing a connection with a base station; Indicates the first Individual during round iteration The Middle An index of the status indicators for a device establishing a connection with a base station; Indicates the central location of the calming group within the population; Represents the first binary variable; This is the floor function; Indicates the first weighting adjustment factor;

[0029] Similarly, , , , , , , replace Update.

[0030] Furthermore, the update process for the conformity group is as follows:

[0031] ;

[0032] ;

[0033] In the formula, Represents the second binary variable, and ; This indicates a position randomly selected from the group. This indicates a position randomly generated from the range of the crowd; This indicates the individuals selected from the calming group; This represents the second weighting adjustment factor;

[0034] Similarly, , , , , , , replace Update.

[0035] Furthermore, the update process for the panic group is as follows:

[0036] ;

[0037] ;

[0038] In the formula, This indicates a randomly selected location within the panic group. This indicates a randomly generated location within the panic group's area; This indicates that in the current iteration, one of the individuals selected from all individuals has the best performance (or the best state).

[0039] Similarly, , , , , , , replace Update.

[0040] Furthermore, when the iteration progress enters Then, the individual enters the crossover mutation phase to update itself. The update process of the crossover mutation phase is as follows:

[0041] ;

[0042] ;

[0043] In the formula, For crossover probability, The mutation probability, , , and These are four random parameters that take values ​​in the range (0, 1). Represents an individual fitness value; This represents the optimal fitness function value in the population. This represents the minimum fitness function value in the population; This represents the average fitness function value in the population. This represents the maximum fitness function value in the population;

[0044] The mutation operation performed on an individual is represented as:

[0045] ;

[0046] In the formula, and These are two random parameters whose values ​​range from (0, 1). Controlling the mutation magnitude in individuals, It controls the individual's search direction;

[0047] Similarly, , , , , , , replace Update.

[0048] An electronic device includes a processor, a memory, and a bus, wherein the processor and the memory are connected via the bus, wherein the memory is used to store a set of program code, and the processor is used to call the program code stored in the memory to execute a multi-stage security task offloading optimization method for low-altitude intelligent networks.

[0049] A non-volatile computer storage medium storing computer-executable instructions that execute a multi-stage security task offloading optimization method for low-altitude intelligent networks.

[0050] Compared with existing technologies, the present invention has the following advantages:

[0051] (1) This invention combines a multi-stage group update strategy (layered update of calm group, conformist group, and panic group), an optimization mechanism for the development stage that adapts to the iterative progress, and a dynamic cross-mutation probability adjustment scheme based on individual fitness with the security constraint design for task offloading in low-altitude intelligent networks. This achieves a synergistic improvement in task offloading efficiency, resource allocation rationality, and data transmission security. It effectively solves the problems of imbalance between exploration and development, insufficient optimization accuracy, and lack of security guarantee in the low-altitude dynamic network environment of traditional methods, and adapts to the dynamic and heterogeneous characteristics of low-altitude intelligent networks.

[0052] (2) The present invention adopts a two-layer algorithm architecture of phased group update and development stage precise optimization. Through the differentiated update logic of different stages (group collaboration, optimal guidance, local fine optimization), it balances the global exploration capability and local development efficiency of the algorithm, avoids the algorithm from getting stuck in local optima, significantly improves the optimization accuracy and convergence speed of the task unloading scheme, and can quickly respond to dynamic changes such as device movement and resource fluctuations in low-altitude intelligent networks.

[0053] (3) This invention ensures population diversity and stability of excellent solutions by dynamically adjusting the crossover probability and mutation probability (adaptively optimized according to individual fitness), and further improves the optimization quality of the offloading scheme; at the same time, it combines the security requirements of the low-altitude intelligent network to design a task offloading mechanism, which effectively ensures data security during task transmission and processing, and improves the resource utilization of edge nodes while meeting the security and compliance requirements of the low-altitude intelligent network for task offloading. Attached Figure Description

[0054] Figure 1 This is a flowchart of the method of the present invention.

[0055] Figure 2 This is a schematic diagram illustrating the effect of the number of devices on the fitness value according to the present invention.

[0056] Figure 3 This is a schematic diagram illustrating the relationship between the number of devices and total energy consumption as disclosed in this invention.

[0057] Figure 4 This is a schematic diagram illustrating the impact of the device's maximum computing power on the device's total energy consumption, as disclosed in this invention. Detailed Implementation

[0058] like Figure 1As shown, the present invention provides a technical solution: a multi-stage security task offloading optimization method for low-altitude intelligent networks, comprising the following steps:

[0059] Step S1: Construct a network architecture based on the basic device information of the low-altitude intelligent network in the edge computing system. The network architecture includes a communication model, a computing model, and a security model; configure the optimization problem under multiple constraints in the network architecture.

[0060] Step S2: Generate an initial population based on the optimization problem, and use an improved escape optimization algorithm to search the initial population to obtain the target population, and output the global optimal solution in the target population; the improved escape optimization algorithm introduces crossover and mutation operations and a constraint penalty function containing task execution delay and security service cost on the basis of the original escape optimization algorithm; the crossover and mutation operation is used to improve population diversity and escape local optima; the constraint penalty function is used to calculate fitness and update the population ranking;

[0061] Specifically: An improved escape optimization algorithm is used for optimization search. First, the population and iteration-related parameters are initialized, fitness is calculated, and the population is ranked. In each iteration, the strategy is adjusted according to the iteration progress: when the iteration progress... No more than At that time, calculate the fear index. For the number of iterations, To maximize the number of iterations, reorder and group the population into calm, conform, and panic groups, and perform update operations for the calm, conform, and panic groups; when the iteration progress is... arrive During this period, the development phase begins, where the population is adjusted and optimized through local exploration; when the iteration progress exceeds... When the crossover and mutation phase is initiated, an adaptive single-point crossover and mutation operation is performed. After each update, the fitness is evaluated and the best individual is updated until the set number of iterations is completed. This updates the information of the best individual in the current population and outputs the position information encoding of all individuals in the population.

[0062] Step S3: Calculate the energy consumption optimization configuration based on the global optimal solution.

[0063] The specific process for obtaining basic equipment information for the low-altitude intelligent network is as follows: [The process involves obtaining...] The index set of each device is ,in, Indicates the index of the device. Indicates the total number of devices; The index set of micro base stations is The index set of macro base stations is The index set of all base stations is ,in, Indicates the index of the micro base station. This indicates the number of micro base stations, and s represents the index of the base station in the UAV-coordinated intelligent reflective surface-assisted mobile edge computing system. Indicates the total number of base stations; The index set of intelligent reflective surfaces coordinated by drones is ,in, The index representing the intelligent reflective surface of drone collaboration. This indicates the number of intelligent reflective surfaces used in drone collaboration; a single intelligent reflective surface used in drone collaboration contains... One reflective element, The set of indices for each reflective element is ,in, Indicates the index of the reflective element in the intelligent reflective surface of the drone collaboration. This represents the number of reflective elements in a drone-coordinated intelligent reflective surface; the... The reflection coefficient matrix of the reflective element in the intelligent reflective surface coordinated by multiple drones is denoted as follows: ,in, Indicates the first The first reflective element in the... The reflectivity of a smart reflective surface coordinated by multiple drones Next, the physical locations of all micro base stations are obtained, and the data is used based on the physical locations of all micro base stations. The clustering algorithm divides all micro base stations into groups. There are clusters; in each cluster there are Each sub-channel is used by the micro base stations, devices, and drones in the cluster to form a collaborative intelligent reflector. When a device transmits to the intelligent reflector formed by the base station and the drone, the sub-channel selected is the same. The set of indices of the sub-channels is denoted as ,in, Indicates the index of the sub-channel; the task passes through the first... The device was passed by the first The sub-channel is transmitted to the first The channel gain of the intelligent reflector coordinated by multiple drones is denoted as... The task passed the first A drone-integrated intelligent reflector via the first The sub-channel is transmitted to the first The channel gain of each micro base station is denoted as The task passed the first The micro base station transmits back to the first The channel gain of each device is denoted as The power of Gaussian white noise is denoted as .

[0064] Among these steps, the communication model is constructed as follows: First, the total bandwidth of the network architecture is divided into... and They are used separately for micro base stations and macro base stations, among which It is the frequency band division factor, and The total bandwidth of the network architecture is The sub-channel bandwidth is The number of sub-channels used by each cluster is... The total number of clusters is denoted as , This is a floor function; when a device is associated with a micro base station, it uses (non-orthogonal multiple access) NOMA technology to round down the floor function. The unloaded portion of each device task is simultaneously sent to multiple micro base stations; subsequently, the uplink transmission rate is calculated, i.e.: Calculate the first On the sub-channel, the ... The device will send the task to the first... Uplink NOMA transmission rate of individual micro base stations ;in, This indicates the bandwidth used by the micro base station. Indicates the first The device is associated with the first The decision index coefficient of a drone-coordinated intelligent reflective surface, among which and For the first Each device must select a drone-coordinated smart reflective surface for association. Indicates the first The device's own transmission power Indicates the first The transmission power of the device itself; This indicates that during the uplink NOMA transmission process, except for the first... The device and the first Other devices besides the intelligent reflective surface coordinated by the drone. and the A drone-integrated intelligent reflective surface The device and the first Interference generated by a smart reflective surface coordinated by multiple drones Except for the first An index beyond the intelligent reflective surface in collaboration with drones. This indicates that the task is performed via a micro base station. Back to device Channel gain, This represents Gaussian white noise. Indicates the first in the cluster The task of the device has been completed. The sub-channels are transmitted to the first sub-channel via uplink NOMA. Channel gain of a smart reflective surface in collaboration with a drone; This indicates that during the uplink NOMA transmission process, other... The first device will send the task to another third device. Channel gain interference caused by a smart reflective surface coordinated by a drone; Indicates that the task has passed the first The intelligent reflective surface, coordinated by a drone, passes through the first... The sub-channel to the first One micro base station; Indicating drone-assisted smart reflective surfaces A set of indices for the reflection coefficients of the reflective elements. , and These represent the path loss factor and the Rice fading factor, respectively. Indicates the first in the cluster The device to the first The distance of a smart reflective surface coordinated by multiple drones; , indicating the first The device to the first A collaborative intelligent reflective surface line-of-sight link component for multiple drones; among which... For frequency, Represents the imaginary unit. and These all represent the serial numbers of the reflective elements; the drone-coordinated intelligent reflective surface has... Several reflective elements, numbered from 1 to... . Indicates the first Channel coefficient vectors of each reflective element in a smart reflective surface coordinated by multiple UAVs; This represents the phase change that occurs when the signal travels that distance, where Indicates the first in the cluster The device to the first The cosine value of a smart reflective surface in collaboration with drones; , These represent the x-coordinate of the device and the x-coordinate of the intelligent reflective surface in collaboration with the drone, respectively. Indicates the first The device to the first A drone-coordinated intelligent reflective surface non-line-of-sight link component.

[0065] Among them, constructing a security model involves: first, calculating the first... The device selection is number one. A cryptographic algorithm to protect the transmission task The probability of failure when the expected security level is not achieved. Otherwise, for ;in, Indicates the first The task of protecting transmission for individual devices The safety risk coefficient, Indicates the desired protection level for the task, and will the first The protection level of a cryptographic algorithm is denoted as . According to the formula Calculate the first Security costs of individual devices ,in, Indicates a transmission task Financial losses in the event of failure; For the first The device and the first The index of the status indicators for establishing connections between base stations; Indicates the first The device is associated with the first The decision index of a smart reflective surface in collaboration with drones.

[0066] A computational model was constructed based on the paper "Multi-stage offloading and resource allocation with security in ultra-dense NOMA-IoT networks" by Zhou Tianqing and Fu Yanyan published in the journal IEEE Internet of Things Journal in 2023.

[0067] Complete computing device The total energy consumption of all computing tasks can be expressed by the formula:

[0068] ;

[0069] The first term on the right side of the equals sign represents the energy consumption generated when the task is computed locally on the device; the second term on the right side of the equals sign represents the energy consumption generated during encrypted transmission of the task; and the last term represents the energy consumption generated during the uplink transmission of task data to the base station. Indicates equipment Calculation task data The number of CPU cycles used per bit This indicates the bit size of the task data computed locally. Indicates the first The task data that each device needs to process bit size, Indicates the first The device was offloaded to the micro base station. Task data bit size, Indicates the first Each device processes the allocated computing power locally. For the first The device and the first The index of the status indicators for establishing connections between base stations. Indicates the first The tasks transmitted by each device sequentially pass through the intelligent reflective surface coordinated by the drone and the first... The selected base station A cryptographic algorithm, It represents the computational power of the cryptographic algorithm for encryption processing. Indicates the first Each device is associated with a base station. Selecting sub-channels during calculation As a state variable of the uplink transmission connection node for the task Indicates the first The device will transmit the task to the first... When selecting a smart reflector in collaboration with multiple drones, choose the first one. Each sub-channel serves as a state variable for the connected nodes. Indicates the first A smart reflective surface, coordinated by multiple drones, transmits the mission to the first... When selecting the first base station Each sub-channel serves as a state variable for the connected nodes. Indicates the first The transmission power of each device Indicates the first On the sub-channel, the ... The device will send the task to the first... The uplink NOMA transmission rate of each base station.

[0070] Among them, the optimization problem under multiple constraints in configuring network architecture:

[0071] ;

[0072] In the formula, The goal of the energy minimization optimization problem is to minimize the total computational energy consumption of all device tasks in a drone-cooperative intelligent reflector-assisted mobile edge computing system. ; An indexed state set indicating whether the device has decided to establish a connection with the base station. , For the first The device and the first The index of the status indicators for establishing connections between base stations. This indicates the cryptographic algorithm selected when the device first transmits the task to the drone-coordinated intelligent reflective surface, and then transmits the task to the base station through the drone-coordinated intelligent reflective surface. The set of index states, , Indicates the first The tasks transmitted by each device sequentially pass through the intelligent reflective surface coordinated by the drone and the first... The selected base station One cryptographic algorithm; This represents the indexed state set of state variables for selecting sub-channels as connection nodes when the device transmits tasks to the UAV-coordinated intelligent reflector and when the UAV-coordinated intelligent reflector transmits tasks to the base station. , Indicates the first The device will transmit the task to the first... When selecting a smart reflector in collaboration with multiple drones, choose the first one. Each sub-channel serves as a state variable for the connected nodes. Indicates the first A smart reflective surface, coordinated by multiple drones, transmits the mission to the first... When selecting the first base station Each sub-channel serves as a state variable for the connection node; This represents the set of indexes representing the allocation of transmit power for all devices. , Indicates the first The transmission power of each device Indicates the maximum transmission power; This is a set of indexes indicating whether the device is associated with a smart reflective surface selected for drone collaboration. , Indicates the first The device is associated with the first Decision-making index of intelligent reflective surface in collaboration with drones; Indicates the first Task processing latency of individual devices Indicates the first Maximum execution time for each device; This represents the reflection coefficient matrix of the reflective elements in the intelligent reflective surface used in drone collaboration. , Indicates the first The first reflective element in the... The reflectivity of a smart reflective surface coordinated by multiple drones; A set of indices representing the bit size of the task data offloaded from the device to the base station. , Indicates the first The device was offloaded to the micro base station. Task data bit size, Total amount of task data; This indicates task data that represents the offloading of tasks from micro base stations to macro base stations. A set of indices of bit size, , Indicates the first The tasks that each device needs to handle come from the micro base station. Task data unloaded to macro base station The size of the bits; Indicates the first The task data that each device needs to process bit size, Indicates the first The device transmits to the first Task data of each base station The size of the bits; Indicates the first The security cost of each device Indicates the first The maximum cost constrained by each device This represents the lower boundary value for data unloading; These represent the first through the fourteenth constraints, respectively.

[0073] The specific process of step S2 is as follows:

[0074] Step S2.1: Initialize the maximum number of iterations for the improved escape optimization algorithm and the current iteration number Set to 1.

[0075] Step S2.2: Let the number of individuals in the population be... Any individual in the population uses If expressed as such, then the population is For any given individual, encode it using the parameter variables of the optimization problem; index the state set. Encoding into population , Represents an individual The Middle Indexed state index of the status indicators for establishing a connection between a device and a base station; indexed state set Encoding into population , Represents an individual The Middle A set of cryptographic algorithms selected by each device; an index state set Encoding into population , Represents an individual The Middle The index of the sub-channel selected by each device; index set Encoding into population , Represents an individual The Middle Transmit power of individual devices; index set Encoding into population , Represents an individual The Middle The index of the drone collaborative intelligent reflector selected by each device; index set Encoding into population , Indicates the first The reflection coefficient of the reflective element in the collaborative intelligent reflective surface of the drone associated with each device; index set Encoding into population , Represents an individual The middle task starts from the first The amount of tasks offloaded from a device to a micro base station; index set Encoding into population , Represents an individual The Middle The amount of data that each device needs to handle, which is offloaded from the micro base station to the macro base station.

[0076] Step S2.3: Initialize the population and establish the individuals in the population. Fitness function:

[0077] ;

[0078] In the formula, For the first Penalty factor for the maximum execution time of a task on a single device; For the first The penalty factor for the maximum security cost of each device; Represents the fitness function; This represents the total energy consumption of all devices.

[0079] Step S2.4: Use the fitness function to calculate the fitness of all individuals in the population. The fitness value of each individual, and the individual with the highest fitness value. As the best individual in the overall historical context.

[0080] Step S2.5: First, determine the iteration progress. Is it less than or equal to? If the iteration progress Less than or equal to Then execute the update operations for the calm group, conformity group, and panic group; if If so, then proceed to the development phase to update the individual; if Then, proceed to the crossover and mutation phase to perform adaptive single-point crossover and mutation operations; if the current iteration number is... If so, then the codes of all individuals will be output.

[0081] The update process for the Calm Group is as follows:

[0082] ;

[0083] ;

[0084] ;

[0085] In the formula, To increase with the current iteration number The probability of panic due to change; This indicates a randomly selected position within the calming group; This indicates a randomly generated location within the range of the calming group. Parameters indicating adjustments to an individual's position; Indicates the first Individual during round iteration The Middle An index of the status indicators for a device establishing a connection with a base station; Indicates the first Individual during round iteration The Middle An index of the status indicators for a device establishing a connection with a base station; Indicates the central location of the calming group within the population; Represents the first binary variable; This is the floor function; Indicates the first weighting adjustment factor;

[0086] Similarly, , , , , , , replace Update.

[0087] The update process for the conformist group is as follows:

[0088] ;

[0089] ;

[0090] In the formula, Represents the second binary variable, and ; This indicates a position randomly selected from the group. This indicates a position randomly generated from the range of the crowd; This indicates the individuals selected from the calming group; This represents the second weighting adjustment factor;

[0091] Similarly, , , , , , , replace Update.

[0092] The update process for the panic group is as follows:

[0093] ;

[0094] ;

[0095] In the formula, This indicates a randomly selected location within the panic group. This indicates a randomly generated location within the panic group's area; This indicates that in the current iteration, one of the individuals selected from all individuals has the best performance (or the best state).

[0096] Similarly, , , , , , , replace Update.

[0097] Among them, when the iteration progress enters If so, the individual will enter the development phase for updating; the update process in the development phase is as follows:

[0098] ;

[0099] In the formula, This refers to individuals randomly selected from the calming group.

[0100] Among them, when the iteration progress enters Then, the individual enters the crossover mutation phase to update itself. The update process of the crossover mutation phase is as follows:

[0101] ;

[0102] ;

[0103] In the formula, For crossover probability, The mutation probability, , , and These are four random parameters that take values ​​in the range (0, 1). Represents an individual fitness value; This represents the optimal fitness function value in the population. This represents the minimum fitness function value in the population; This represents the average fitness function value in the population. This represents the maximum fitness function value in the population;

[0104] The mutation operation performed on an individual is represented as:

[0105] ;

[0106] In the formula, and These are two random parameters whose values ​​range from (0, 1). Controlling the mutation magnitude in individuals, It controls the individual's search direction;

[0107] Similarly, , , , , , , replace Update.

[0108] An electronic device includes a processor, a memory, and a bus, wherein the processor and the memory are connected via the bus, wherein the memory is used to store a set of program code, and the processor is used to call the program code stored in the memory to execute a multi-stage security task offloading optimization method for low-altitude intelligent networks.

[0109] A non-volatile computer storage medium storing computer-executable instructions that execute a multi-stage security task offloading optimization method for low-altitude intelligent networks.

[0110] The effects of the embodiments of this invention can be further illustrated through simulation.

[0111] Figure 2The changes in fitness values ​​for different algorithms are shown under varying numbers of devices. As the number of devices increases from 10 to 40, the fitness values ​​of four algorithms—IESC (improved escape optimization algorithm), IWOA (improved whale algorithm), ESC (escape optimization algorithm), and IESC w / o IRS (improved escape optimization algorithm without IRS)—all show a decreasing trend. Among them, the IESC algorithm consistently has the highest fitness value, indicating its optimal performance. ESC and IWOA are next, while IESC w / o IRS performs the worst.

[0112] Figure 3 The data shows the changes in total energy consumption of different algorithms under varying numbers of devices. As the number of devices increases from 10 to 40, the total energy consumption of the four algorithms—IESC, IWOA, ESC, and IESC w / o IRS—all show an upward trend. Among them, the IESC algorithm consistently exhibits the lowest total energy consumption, demonstrating the best performance. ESC and IWOA are next, while IESC w / o IRS has the highest energy consumption.

[0113] Figure 4 The data shows the changes in total energy consumption of different algorithms under different device maximum computing speeds. As the device maximum computing speed increases from 0.5 to 2, the total energy consumption of the four algorithms—IESC, IWOA, ESC, and IESC w / o IRS—all show an upward trend. Among them, the IESC algorithm consistently has the lowest total energy consumption, demonstrating the best performance. ESC and IWOA are next, while IESC w / o IRS has the highest energy consumption.

[0114] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A multi-stage safety task offloading optimization method for low-altitude intelligent networks, characterized in that, Includes the following steps: Step S1: Construct a network architecture based on the basic device information of the low-altitude intelligent network in the edge computing system. The network architecture includes a communication model, a computing model, and a security model. Configure the optimization problem under multiple constraints in the network architecture. The basic device information of the low-altitude intelligent network includes devices, macro base stations, micro base stations, UAV collaborative intelligent reflectors, the total number of reflective elements contained in a single intelligent reflector, and the corresponding index set and identifier. Step S2: Generate an initial population based on the optimization problem, use an improved escape optimization algorithm to search the initial population to obtain the target population, and output the global optimal solution in the target population; Step S3: Calculate energy consumption optimization configuration based on the global optimal solution; The specific process of step S2 is as follows: Step S2.1: Initialize the maximum number of iterations for the improved escape optimization algorithm and the current iteration number Set to 1; Step S2.2: Let the number of individuals in the population be... Any individual in the population uses If expressed as such, then the population is ; For any given individual, it is encoded using the parameter variables of the optimization problem; The parameter variables of the optimization problem include: The index set of state indicators indicating whether the device has decided to establish a connection with the base station. , For the first The device and the first The index of the status indicators for establishing connections between base stations; The device first transmits the task to the drone-coordinated intelligent reflective surface, and then transmits the task to the base station via the drone-coordinated intelligent reflective surface using the selected cryptographic algorithm. index state set , Indicates the first The tasks transmitted by each device sequentially pass through the intelligent reflective surface coordinated by the drone and the first... The selected base station One cryptographic algorithm; The device transmits tasks to the UAV-coordinated intelligent reflector, and the UAV-coordinated intelligent reflector transmits tasks to the base station. This involves selecting a sub-channel as the index state set of the connection node's state variables. , Indicates the first The device will transmit the task to the first... When selecting a smart reflector in collaboration with multiple drones, choose the first one. Each sub-channel serves as a state variable for the connected nodes. Indicates the first A smart reflective surface, coordinated by multiple drones, transmits the mission to the first... When selecting the first base station Each sub-channel serves as a state variable for the connection node; Set of indexes for the transmission power allocation of all devices , Indicates the first The transmission power of each device Indicates the maximum transmission power; Whether the device is associated with the set of indexes of the intelligent reflective surface for drone collaboration , Indicates the first The device is associated with the first Decision-making index of intelligent reflective surface in collaboration with drones; Indicates the first Task processing latency of individual devices Indicates the first Maximum execution time for each device; Reflection coefficient matrix of reflective elements in a drone-coordinated intelligent reflective surface , Indicates the first The first reflective element in the... The reflectivity of a smart reflective surface coordinated by multiple drones; A set of indexes for the bit size of the task data offloaded from the device to the base station. , Indicates the first The device was offloaded to the micro base station. Task data bit size, Total amount of task data; Task data offloaded from micro base stations to macro base stations A set of indexes of bit size , Indicates the first The tasks that each device needs to handle come from the micro base station. Task data unloaded to macro base station The size of the bits; index state set Encoding into population , Represents an individual The Middle Indexed state index of the status indicators for establishing a connection between a device and a base station; indexed state set Encoding into population , Represents an individual The Middle A set of cryptographic algorithms selected by each device; an index state set Encoding into population , Represents an individual The Middle The index of the sub-channel selected by each device; index set Encoding into population , Represents an individual The Middle Transmit power of individual devices; index set Encoding into population , Represents an individual The Middle The index of the drone collaborative intelligent reflector selected by each device; index set Encoding into population , Indicates the first The reflection coefficient of the reflective element in the collaborative intelligent reflective surface of the drone associated with each device; index set Encoding into population , Represents an individual The middle task starts from the first The amount of tasks offloaded from a device to a micro base station; index set Encoding into population , Represents an individual The Middle The amount of data that each device needs to process from micro base stations to macro base stations; Step S2.3: Initialize the population and establish the individuals in the population. Fitness function: ； In the formula, For the first Penalty factor for the maximum execution time of a task on a single device; For the first The penalty factor for the maximum security cost of each device; Represents the fitness function; This represents the total energy consumption of all devices; Step S2.4: Use the fitness function to calculate the fitness of all individuals in the population. The fitness value of each individual, and the individual with the highest fitness value. As the best individual in the overall historical context; Step S2.5: First, determine the iteration progress. Is it less than or equal to? If the iteration progress Less than or equal to Then execute the update operations for the calm group, conformity group, and panic group; if If so, then proceed to the development phase to update the individual; if Then, proceed to the crossover and mutation phase to perform adaptive single-point crossover and mutation operations; if the current iteration number is... If so, then the codes of all individuals will be output.

2. The multi-stage security task offloading optimization method for low-altitude intelligent networks according to claim 1, characterized in that: The specific process for obtaining basic equipment information for the Low Altitude Intelligent Network is as follows: [Identifying...] The index set of each device is ,in, Indicates the index of the device. Indicates the total number of devices; The index set of micro base stations is The index set of macro base stations is The index set of all base stations is ,in, Indicates the index of the micro base station. This indicates the number of micro base stations, and s represents the index of the base station in the UAV-coordinated intelligent reflective surface-assisted mobile edge computing system. Indicates the total number of base stations; The index set of intelligent reflective surfaces coordinated by drones is ,in, The index representing the intelligent reflective surface of drone collaboration. This indicates the number of intelligent reflective surfaces used in drone collaboration; a single intelligent reflective surface used in drone collaboration contains... One reflective element, The set of indices for each reflective element is ,in, Indicates the index of the reflective element in the intelligent reflective surface of the drone collaboration. This represents the number of reflective elements in a drone-coordinated intelligent reflective surface; the... The reflection coefficient matrix of the reflective element in the intelligent reflective surface coordinated by multiple drones is denoted as follows: ,in, Indicates the first The first reflective element in the... The reflectivity of a smart reflective surface coordinated by multiple drones Next, the physical locations of all micro base stations are obtained, and the data is used based on the physical locations of all micro base stations. The clustering algorithm divides all micro base stations into groups. There are clusters; in each cluster there are Each sub-channel is used by the micro base stations, devices, and drones in the cluster to form a collaborative intelligent reflector. When a device transmits to the intelligent reflector formed by the base station and the drone, the sub-channel selected is the same. The set of indices of the sub-channels is denoted as ,in, Indicates the index of the sub-channel; the task passes through the first... The device was passed by the first The sub-channel is transmitted to the first The channel gain of the intelligent reflector coordinated by multiple drones is denoted as... The task passed the first A drone-integrated intelligent reflector via the first The sub-channel is transmitted to the first The channel gain of each micro base station is denoted as The task passed the first The micro base station transmits back to the first The channel gain of each device is denoted as The power of Gaussian white noise is denoted as .

3. The multi-stage security task offloading optimization method for low-altitude intelligent networks according to claim 2, characterized in that: Optimization issues under multiple constraints in configuring network architecture: ； In the formula, The goal of the energy minimization optimization problem is to minimize the total computational energy consumption of all device tasks in a drone-cooperative intelligent reflector-assisted mobile edge computing system. ; Indicates the first The task data that each device needs to process bit size, Indicates the first The device transmits to the first Task data of each base station The size of the bits; Indicates the first The security cost of each device Indicates the first The maximum cost constrained by each device This represents the lower boundary value for data unloading; These represent the first through the fourteenth constraints, respectively.

4. The multi-stage security task offloading optimization method for low-altitude intelligent networks according to claim 3, characterized in that: The update process for the Calm Group is as follows: ； ； ； In the formula, To increase with the current iteration number The probability of panic due to change; This indicates a randomly selected position within the calming group; This indicates a randomly generated location within the range of the calming group. Parameters indicating adjustments to an individual's position; Indicates the first Individual during round iteration The Middle An index of the status indicators for a device establishing a connection with a base station; Indicates the first Individual during round iteration The Middle An index of the status indicators for a device establishing a connection with a base station; Indicates the central location of the calming group within the population; Represents the first binary variable; This is the rounding function; Indicates the first weighting adjustment factor; Similarly, , , , , , , replace Update.

5. The multi-stage security task offloading optimization method for low-altitude intelligent networks according to claim 4, characterized in that: The update process for the conformist group is as follows: ； ； In the formula, Represents the second binary variable, and ; This indicates a position randomly selected from the group. This indicates a position randomly generated from the range of the crowd; This indicates the individuals selected from the calming group; This represents the second weighting adjustment factor; Similarly, , , , , , , replace Update.

6. The multi-stage security task offloading optimization method for low-altitude intelligent networks according to claim 5, characterized in that: The update process for the panic group is as follows: ； ； In the formula, This indicates a randomly selected location within the panic group. This indicates a randomly generated location within the panic group's area; This indicates that in the current iteration, one of the best-performing individuals selected from all individuals; Similarly, , , , , , , replace Update.

7. The multi-stage security task offloading optimization method for low-altitude intelligent networks according to claim 6, characterized in that: When the iteration progress enters Then, the individual enters the crossover mutation phase to update itself. The update process of the crossover mutation phase is as follows: ； ； In the formula, For crossover probability, The mutation probability, , , and These are four random parameters that take values ​​in the range (0, 1). Represents an individual fitness value; This represents the optimal fitness function value in the population. This represents the minimum fitness function value in the population; This represents the average fitness function value in the population. This represents the maximum fitness function value in the population; The mutation operation performed on an individual is represented as: ； In the formula, and These are two random parameters whose values ​​range from (0, 1). Controlling the mutation magnitude in individuals, It controls the individual's search direction; Similarly, , , , , , , replace Update.

8. An electronic device, characterized in that, The system includes a processor, a memory, and a bus, wherein the processor and the memory are connected via the bus, the memory is used to store a set of program code, and the processor is used to call the program code stored in the memory to execute a multi-stage safety task offloading optimization method for low-altitude intelligent networks as described in any one of claims 1-7.

9. A non-volatile computer storage medium storing computer-executable instructions, characterized in that, The computer can execute instructions to perform a multi-stage security task offloading optimization method for low-altitude intelligent networks as described in any one of claims 1-7.

Images