Method for cache and secure multi-hop partial offloading in ultra-dense millimeter wave networks

Abstract

The application discloses a kind of buffer and safe multi-hop partial offloading methods in super dense millimeter wave network, belong to wireless communication field, this method includes: obtaining the network basic information of super dense millimeter wave mobile edge computing network supporting non-orthogonal multiple access and orthogonal frequency division multiple access, network system is constructed according to network basic information, and optimization problem is constructed under the constraint of network system;Initial population is generated according to certain rule, and the position of the historical optimal individual in target population is output by searching initial population using improved gradient optimization algorithm;According to the position of historical optimal individual, execute security type calculation efficiency optimization configuration.The application has multi-base station offloading, considers buffer and millimeter wave factor, meets rate, mobile terminal energy consumption, delay, security vulnerability cost and task and transmission power proportion constraint, can well realize the goal of all mobile terminal local energy consumption minimization.

Description

Technical Field

[0001] This invention belongs to the field of wireless communication, specifically relating to a method for offloading buffers and secure multi-hop components in ultra-dense millimeter-wave networks. Background Technology

[0002] The iteration of communication technologies and the widespread adoption of terminals have spurred a large number of computationally intensive and latency-critical applications such as facial recognition, natural language processing, and interactive games, but have also highlighted energy consumption bottlenecks: local devices often cannot keep up due to limitations in chip computing power and other factors. Mobile edge computing has thus become an essential offloading solution; millimeter wave (mmWave) is seen as an effective means to further reduce latency and increase speed. However, millimeter wave coverage is inherently short, requiring ultra-dense base stations to fill the gaps, which significantly increases the overall network energy consumption; at the same time, the open interfaces on the edge side also amplify the risk of malicious intrusion. In addition, to reduce the air interface and backhaul overhead caused by repeated offloading, "task-level caching" is introduced on the edge side—not only caching static content, but also pre-storing popular reusable tasks and their intermediate data in the edge cloud. Once a mobile terminal's request is successful, the edge node executes it locally and sends back the result, without the terminal needing to upload the original task again, thereby further reducing latency, saving bandwidth, expanding system capacity, and reducing local terminal energy consumption. To address this, this paper proposes a "green, secure, multi-step, and cached" offloading strategy within the framework of ultra-dense millimeter wave and non-orthogonal multiple access (NOMA) + orthogonal frequency division multiple access (OFDMA): achieving joint optimization of task offloading and caching, thereby minimizing energy consumption while ensuring service quality.

[0003] To address the aforementioned issues, it is necessary to explore ways to reduce the local power consumption and signal interference of mobile terminals under multiple constraints such as latency and security costs, thereby improving the user experience. Traditional gradient optimization algorithms suffer from limited global exploration capabilities and are prone to getting trapped in local optima, exhibiting certain limitations in practical applications. Therefore, it is necessary to improve these algorithms to enhance their performance and applicability. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a method for offloading cached and secure multi-hop components in ultra-dense millimeter-wave networks, aiming to solve the problems discussed in the background section.

[0005] To achieve the above objectives, the present invention provides the following technical solution.

[0006] A method for offloading the caching and secure multi-hop components in an ultra-dense millimeter-wave network includes the following steps:

[0007] Step S1: Obtain the basic network information of an ultra-dense millimeter-wave mobile edge computing network that supports non-orthogonal multiple access and orthogonal frequency division multiple access, and construct the network system. The network system includes a communication model, a caching and offloading model, a security model, and a computing model. Construct an optimization problem under the constraints of the network system.

[0008] Step S2: Encode the feasible solutions to the optimization problem;

[0009] Step S3: Generate an initial population based on the encoding results according to certain rules, and iterate the initial population using an improved gradient optimization algorithm. During the iteration process, in addition to the gradient search rule operation and local escape operator operation in the original gradient optimization algorithm, a diversity-guided mutation operation is performed to update the population position information. Based on the position information, the best individual in the current population is obtained and compared with the initial best individual. The historical best individual is selected and the position information of the individual is output.

[0010] Step S4: Perform task calculation and unload resource configuration based on the location information of the selected best individual in the population history.

[0011] Furthermore, the ultra-dense millimeter-wave mobile edge computing network supporting non-orthogonal multiple access and orthogonal frequency division multiple access includes macro base stations, several micro base stations, and several mobile terminals. Each micro base station is equipped with a mobile edge computing server. All micro base stations are connected to the nearest macro base station via wired backhaul links. The number of micro base stations is greater than or equal to the number of mobile terminals. Each mobile terminal must complete the process sequentially under the dual constraints of security costs and time limits. For computationally intensive and latency-sensitive tasks, the mobile terminal communicates with some micro base stations using millimeter waves, while the remaining micro base stations communicate with the mobile terminal using macro waves; the mobile terminal and macro base stations communicate using macro waves.

[0012] Suppose there exists a macro base station in an ultra-dense millimeter-wave mobile edge computing network supporting non-orthogonal multiple access and orthogonal frequency division multiple access. There are [number] micro base stations, with macro base station number 0, and the micro base station index set is [set]. , and set The micro base station in the system communicates with the mobile terminal using millimeter waves, and enables... , This refers to a set of micro base stations that communicate with mobile terminals using millimeter waves. This indicates the number of millimeter-wave communications used between the micro base station and the mobile terminal. The set of all base stations, including micro base stations and macro base stations, is represented by the index set of mobile terminals. Each mobile terminal The task index set is , Indicates the number of mobile terminals. Indicates the index number of the mobile terminal. Indicates the index number of the task. Indicates the total number of tasks; mobile terminal The The task is , The amount of data for the task; The number of CPU cycles required to compute each bit of data, i.e., computational intensity; The required level of security for the mission; The maximum acceptable safety cost; The maximum tolerable processing delay for the task;

[0013] Each micro base station divides its available spectrum resources based on the principle of equal bandwidth partitioning. There are several orthogonal sub-channels, which are divided into two categories: millimeter-wave channels and macro-wave channels; the index set of orthogonal sub-channels is... , , For orthogonal subchannel indexes, It is the sum of millimeter-wave and macro-wave channels; Indicates the system bandwidth. Orthogonal subchannel bandwidth; association set Mobile terminals connected to micro base stations are millimeter-wave mobile terminals, while mobile terminals connected to other micro base stations are macro-wave mobile terminals. All of them use orthogonal multiple access to share the same orthogonal sub-channel of the micro base station. Mobile terminals connected to macro base stations share the frequency band resources of the macro base station.

[0014] Furthermore, the specific process of constructing the communication model is as follows:

[0015] Orthogonal subchannels Mobile terminals associated with macro base stations Uplink transmission rate Represented as:

[0016] ;

[0017] In the formula, Indicates the bandwidth of the macrowave channel; For mobile terminals Other mobile terminals with poorer channel gain that cause interference; For mobile terminals Other mobile terminals with worse channel gain that cause interference Decision-making related to macro base stations; Indicates mobile terminal The transmission power; It is the noise power of the macrowave channel; Indicates mobile terminal Channel gain between macro base stations;

[0018] Orthogonal subchannels Mobile terminals that are associated with micro base stations and use millimeter-wave channels for communication Uplink transmission rate Represented as:

[0019] ;

[0020] In the formula, Indicates the bandwidth of the millimeter-wave channel; For mobile terminals The transmission power; Indicates to mobile terminals The transmit power of other mobile terminals with poorer channel gain that cause interference. For mobile terminals With micro base stations Channel gain for millimeter-wave communication; Indicates to mobile terminals Other mobile terminals with worse channel gain that cause interference With micro base stations Channel gain for millimeter-wave communication; This represents the noise power of the millimeter-wave channel. Indicates in sub-channel Up to mobile terminals With micro base stations A set of indexes of mobile terminals that cause interference in communication between them; Indicates to mobile terminals Other mobile terminals with worse channel gain that cause interference With micro base stations Channel gain for communication between them For mobile terminals With micro base stations Channel gain for communication between them; and Mobile terminals The selected orthogonal sub-channels and the micro base station numbers, and These respectively represent the mobile terminal Other mobile terminals with worse channel gain that cause interference The selected orthogonal sub-channels and the micro base station numbers; The cluster to which the micro base station belongs;

[0021] When satisfied and At that time, the mobile terminal will be in the orthogonal sub-channels of different micro base stations in different clusters. The above received from Interference between base stations outside the cluster.

[0022] In orthogonal sub-channels Mobile terminals that are associated with micro base stations and use macrowave channels for communication The uplink transmission rate is expressed as:

[0023] ;

[0024] In the formula, For mobile terminals With micro base stations Channel gain for macrowave communication between them; Indicates to mobile terminals Other mobile terminals with worse channel gain that cause interference With micro base stations Channel gain for macrowave communication between them;

[0025] When satisfied and At that time, the mobile terminal will be in the sub-channel of different micro base stations within different clusters. The above received from Interference across base stations outside the cluster;

[0026] The noise power of the macrowave channel and the noise power of the millimeter-wave channel are respectively:

[0027] ;

[0028] in, Indicates mobile terminal Decision-making related to macro base stations;

[0029] mobile terminal With micro base stations In sub-channel The channel gain in the equation is given by the following formula:

[0030] ;

[0031] In the formula, Indicates mobile terminal With micro base stations In sub-channel Channel gain in; This represents the path loss during transmission in the channel; Represents the speed of light; This indicates the carrier frequency in the channel; the carrier frequency is 2GHz for macrowave channels and 70GHz for millimeter-wave channels. This indicates the antenna gain of the channel; the antenna gain of the macrowave channel is 0 dBi, the antenna gain value is 18 dBi when the millimeter wave channel is used as the main lobe gain, and the antenna gain value is 2 dBi when it is used as the side lobe gain. Represents the small-scale components of the channel; This represents the path loss of the channel.

[0032] Furthermore, the specific process of constructing the caching and unloading model is as follows:

[0033] A cache status indicator is set on the base station for each task. When the indicator value is 1, it means that the task has been cached on the base station, and when the value is 0, it means that it has not been cached. The cache status indicator is generated based on the historical access probability distribution of files by mobile terminals.

[0034] When a mobile terminal establishes an association with a micro base station, it first checks whether the task exists in the micro base station's storage space. If the task is already cached in the micro base station, the entire task is processed directly at the local base station, and the processing result is transmitted back to the mobile terminal via a wireless backhaul link.

[0035] If the associated micro base station does not cache the task, the mobile terminal will continue to check whether the macro base station has cached the task; if the task has been cached in the macro base station, the entire task will be processed on the macro base station, and the processing result will first be sent to the micro base station through the wired link, and then sent back to the mobile terminal through the wireless backhaul link.

[0036] When a mobile terminal establishes an association with a macro base station, it first checks whether the task exists in the macro base station's storage space. If the task is already cached on the macro base station, the entire task is processed directly on the macro base station, and the processing result is sent directly to the mobile terminal via the wireless backhaul link.

[0037] Furthermore, the optimization problem can be expressed as:

[0038] ;

[0039] In the formula, Let represent the objective function to be optimized. This represents the total energy consumption calculated locally. To optimize the unified representation of the problem-related set, we denote it as the variable set; Indicates mobile terminal and micro base stations The set of related decisions between them Indicates mobile terminal and micro base stations Related decisions; For mobile terminals The A task for cryptographic algorithms A set of security decision instructions, For mobile terminals The A task for cryptographic algorithms Security decision instructions A set of cryptographic algorithm indexes. The index number representing the cryptographic algorithm. Indicates the total number of cryptographic algorithms; For mobile terminals In orthogonal subchannels The set of associated indicators on, For mobile terminals In orthogonal subchannels The associated instructions on; Frequency band allocation factor; For mobile terminals Transmit power set, For mobile terminals Transmission power; For mobile terminals The collection of local computing capabilities, For mobile terminals Local computing power; For mobile terminals The The task is offloaded to the micro base station. The data set For mobile terminals The The task is offloaded to the micro base station. The amount of data; For mobile terminals The The task is transmitted via micro base station Decryption Then, the data set is unloaded to the macro base station again. For mobile terminals The The task is transmitted via micro base station Decryption Then, the amount of data unloaded to the macro base station again; For the task At the base station The set of cache status indicators on For the task At the base station The cache status indicator on the above; the set of optimization problems mentioned above is uniformly denoted as the variable set. ;

[0040] The meaning of constraints: It is a latency constraint, representing the total time required for all tasks to be completed on the mobile terminal. It must not exceed its maximum tolerable delay. ; It is a security constraint, representing the total security breach cost of a mobile terminal. It must not exceed its preset upper limit. ; and Together they constitute an association constraint, meaning that each mobile terminal can only be associated with one base station; and Together they constitute the encryption algorithm constraint, meaning that each task can only use one encryption algorithm; and Together they constitute the sub-channel constraint, meaning that each sub-channel can be allocated to at most one mobile terminal; and Together, they constitute the cache quantity constraint, meaning that each task can be cached at most in one base station; This is a total cache limit, meaning that the total number of cached tasks at a base station cannot exceed its maximum cache capacity. ; It is a computing power constraint, representing the local computing power of the mobile terminal. The upper limit is ; It is a frequency band allocation factor constraint, representing the frequency band allocation factor. The value range is (0, 1); It is a transmit power constraint, representing the transmit power of the mobile terminal. The upper limit is ; This is an unloading data volume constraint, indicating that the data volume of multiple unloading steps must meet the constraint. , For mobile terminals The Total amount of unloaded data for each task.

[0041] Furthermore, step S2 is represented as:

[0042] Optimization problem Each feasible solution is encoded as a vector individual, and the index set is... For any vector individual To compile all sets related to the optimization problem , , , , , , , and One-to-one mapping to multiple component sets: For vector individuals China Mobile Terminal Selected base station index set, For vector individuals China Mobile Terminal Selected base station index; For vector individuals China Mobile Terminal The set of selected cryptographic algorithm numbers, For vector individuals China Mobile Terminal The selected cryptographic algorithm number, A collection of mobile terminals containing different tasks; For vector individuals China Mobile Terminal The set of sub-channel indices assigned, For vector individuals China Mobile Terminal The assigned sub-channel index; For vector individuals Frequency band allocation factor; For vector individuals China Mobile Terminal The set of computing power, For vector individuals China Mobile Terminal Computational power; For vector individuals China Mobile Terminal Selected cache base station index set, For vector individuals China Mobile Terminal Selected cache base station index; For vector individuals China Mobile Terminal The set of transmit power, For vector individuals China Mobile Terminal The transmission power; For vector individuals China Mobile Terminal The set of data volumes offloaded to associated base stations For vector individuals China Mobile Terminal The amount of data offloaded to associated base stations; For vector individuals China Mobile Terminal The set of data forwarded from the associated micro base station to the macro base station. For vector individuals China Mobile Terminal The amount of data forwarded from the associated micro base station to the macro base station.

[0043] Furthermore, the specific operation of step S3 is as follows:

[0044] Step S3.1: Set the maximum number of iterations , This represents the current iteration number, starting from 1.

[0045] Step S3.2: Initialize the population individuals by generating the initial population according to the following rules;

[0046] ;

[0047] In the formula, Represents individual vectors Other mobile terminals The initial solution for the selected base station index. For vector individuals Other mobile terminals The initial solution for the assigned sub-channel index. For vector individuals The initial solution of the frequency band division factor, For vector individuals Other mobile terminals The initial solution based on computational power. For vector individuals Other mobile terminals The initial solution for the transmit power, For vector individuals Mobile terminals related to the task The initial solution of the selected cryptographic algorithm number, For vector individuals Mobile terminals related to the task The initial solution for the selected cache base station index. For vector individuals Mobile terminals related to the task The amount of data offloaded to associated base stations For vector individuals Mobile terminals related to the task The amount of data forwarded from the associated micro base station to the macro base station, Represents from discrete sets A uniformly and randomly selected element is chosen from the middle; Indicates the interval A random real number is generated within the range; This represents mapping individual linear index vectors to... Matrix row index Subscript ; This indicates that a random hold operation is performed on the input variable within its capacity space;

[0048] Step S3.3: Let the current position of the individual in the t-th iteration be... Gradient search rule operations are performed on the population to obtain the updated positions of individuals after the gradient search rule operations. The process is as follows:

[0049]

[0050] In the formula, , For a set containing 4 components The first set of encoded variables, For a set containing 4 components The second set of encoded variables, Let be the current individual position in the t-th iteration of the first set of encoded variables. Let be the current individual position in the t-th iteration of the second set of encoded variables. , The positions of individuals after being updated by gradient search rules are respectively the first set of encoded variables and the second set of encoded variables. , These represent the positions of the worst-performing and best-performing individuals in the history of the first set of encoded variables after the t-th iteration. , These are the positions of the worst and best historical individuals in the second set of encoded variables after the t-th iteration, respectively. This indicates a rounding function; Represents a standard normally distributed random number; Represents a random number between (0, 1); It is a small constant; For element-wise multiplication, To divide element by element, , These are the first adaptive iteration coefficient and the second adaptive iteration coefficient, respectively, and they are numerically equal. These are the directions of the third gradient search and the second gradient search, respectively.

[0051] Then, the idea of ​​Newton's method is introduced to update the population position. The position update rule is as follows:

[0052] ;

[0053] in: , To introduce the updated individual position using the Newton method through gradient search rule operations, , The positions of the individuals updated by Newton's method are introduced from the first and second sets of encoded variables respectively through gradient search rule operations. These are the first and third gradient coefficients introduced into Newton's method, respectively. The second and fourth gradient coefficients of Newton's method are introduced respectively.

[0054] Based on the individual's current position in the t-th iteration And the two newly generated updated positions and Next generation vector individuals The location is determined as follows:

[0055]

[0056] in, , These are the first and second sets of encoded variables after the (t+1)th iteration, respectively. This means randomly generating a real number within the interval (0,1);

[0057] Step S3.4: Perform compound operations on the population; sequentially execute the gradient search rule and local escape operator operations on the entire population; after each operation, perform boundary processing and greedy selection, and update the position information of the worst and best historical values. The specific process is as follows:

[0058] With probability The individual position after the (t+1)th iteration Perform the following operations:

[0059]

[0060] in, , These are the vector individuals of the first set of encoded variables after the t-th iteration. Location, For the random vector individuals of the first set of encoded variables after the t-th iteration The updated location These are the vector individuals of the second set of encoded variables after the t-th iteration. Location, For the random vector individuals of the second set of encoded variables after the t-th iteration The updated location For interval A random number that is uniformly distributed within the range; These are random numbers distributed according to a standard normal distribution. for Random numbers within.

[0061] Step S3.5: Perform diversity-guided mutation operations on the population. The specific process is as follows:

[0062] Diversity measure of multidimensional numerical problems Defined as:

[0063] ;

[0064] In the formula, They represent the sets related to the optimization problem, respectively. , , , , , The length of the diagonal of the feasible region; For vector individuals China Mobile Terminal The average value of the selected base station indexes. For vector individuals China Mobile Terminal Average number of selected cryptographic algorithms For vector individuals China Mobile Terminal The average of the assigned sub-channel indices, For vector individuals China Mobile Terminal The average computing power, For vector individuals China Mobile Terminal The average value of the selected cache base station indexes. For vector individuals China Mobile Terminal The average transmit power;

[0065] Performing mutation operations according to probability is...

[0066]

[0067] In the formula, All are constant coefficients greater than 0 and less than 1, representing the diversity threshold. These represent the first, second, and third mutation probabilities. It is the probability of performing the mutation operation. Three possible values; when the mutation probability of the diversity threshold meets the requirements, the mutation operation will be performed again on the individuals in the population.

[0068] Step S3.6: Determine whether the current iteration count has reached the maximum iteration count. If not, execute steps S3.3, S3.4, and S3.5 in sequence. After execution, increment the iteration count by one. If the maximum iteration count has been reached, output the current result.

[0069] Step S3.7: Calculate the fitness value of all individuals in the current population using the fitness function, select the individual with the highest fitness value as the current best individual and compare it with the initial best individual, take the individual with the larger fitness value as the historical best individual, and output the corresponding position information.

[0070] Furthermore, vector individuals The fitness function is defined as:

[0071] ;

[0072] In the formula, For vector individuals The set of encoded variables; The fitness function; and Mobile terminals The penalty coefficients corresponding to the time delay constraints and security constraints.

[0073] The present invention also provides a non-volatile computer storage medium storing computer-executable instructions that execute the cache and secure multi-hop offloading method in the ultra-dense millimeter-wave network.

[0074] The present invention also provides an electronic device, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to cause the at least one processor to perform the cache and secure multi-hop offloading method in the ultra-dense millimeter-wave network.

[0075] The beneficial effects of this invention are:

[0076] (1) This invention alleviates the energy consumption and terminal computing pressure of ultra-dense base stations by combining a multi-step offloading framework with NOMA technology, breaks through the bottleneck of millimeter wave coverage and interference by hybrid communication mode, improves the efficiency of resource optimization by improving gradient optimization algorithm, and fills the gap in security protection and multi-constraint balance by security model; thus comprehensively improving the user experience.

[0077] (2) This invention addresses the problems of vulnerability of edge computing communication to attacks and lack of coordinated consideration of latency and security costs by constructing an integrated model covering communication, computing offloading, and security. Through hierarchical encryption algorithms and quantitative security indicators, it achieves precise protection for different sensitive tasks; by incorporating multiple constraints such as latency, security risks, and device association, it ensures that while minimizing energy consumption, task latency and security costs are controlled within an acceptable range.

[0078] (3) This invention addresses the limitations of millimeter wave coverage and the prominent problem of network interference by improving the communication method of traditional single-band spectrum resources. It adopts a hybrid method of dual-band communication between mobile terminals and micro base stations (millimeter wave or macro wave communication) and macro base stations (macro wave communication). This further reduces co-channel interference, significantly improves communication speed, and ensures efficient data transmission for task offloading.

[0079] (4) This invention addresses the problem that traditional gradient optimization algorithms have limited global exploration capabilities and are prone to getting trapped in local optima. By operating the gradient search rules, the global search capability of the algorithm is enhanced. Furthermore, by operating the local escape operator, the risk of local optima is reduced, and the optimization accuracy and convergence speed are improved. In addition, the diversity-guided mutation is introduced to enhance the diversity of materials, so as to achieve multiple solutions and optimal solutions. It can efficiently solve nonlinear mixed integer multi-constraint optimization problems coupled in the network. Attached Figure Description

[0080] Figure 1 This is a flowchart of the method of the present invention.

[0081] Figure 2 This is a schematic diagram illustrating the impact of mobile terminal task data volume on total local energy consumption according to the present invention.

[0082] Figure 3 This is a schematic diagram showing the changes in the fitness values ​​of the traditional gradient optimization algorithm and the improved gradient optimization algorithm of this invention during the iteration process. Detailed Implementation

[0083] The method of the present invention will be further described below with reference to the accompanying drawings.

[0084] like Figure 1 As shown, a method for offloading the buffer and secure multi-hop components in an ultra-dense millimeter-wave network includes the following steps:

[0085] Step S1: Obtain the basic network information of an ultra-dense millimeter-wave mobile edge computing network that supports non-orthogonal multiple access and orthogonal frequency division multiple access, and construct the network system. The network system includes a communication model, a caching and offloading model, a security model, and a computing model. Construct an optimization problem under the constraints of the network system.

[0086] Step S2: Encode the feasible solutions to the optimization problem;

[0087] Step S3: Generate an initial population based on the encoding results according to certain rules, and iterate the initial population using an improved gradient optimization algorithm. During the iteration process, in addition to the gradient search rule operation and local escape operator operation in the original gradient optimization algorithm, a diversity-guided mutation operation is performed to update the population position information. Based on the position information, the best individual in the current population is obtained and compared with the initial best individual. The historical best individual is selected and the position information of the individual is output.

[0088] Step S4: Perform task calculation and unload resource configuration based on the location information of the selected best individual in the population history.

[0089] The ultra-dense millimeter-wave mobile edge computing network, supporting both non-orthogonal multiple access (NOAMI) and orthogonal frequency division multiple access (OFDMA), comprises macro base stations, several micro base stations, and several mobile terminals. Each micro base station is equipped with a mobile edge computing server. All micro base stations are connected to the nearest macro base station via wired backhaul links. The number of micro base stations is greater than or equal to the number of mobile terminals. Each mobile terminal must complete the following sequentially under the dual constraints of security costs and time limits. For computationally intensive and latency-sensitive tasks, the mobile terminal communicates with some micro base stations using millimeter waves, while the remaining micro base stations communicate with the mobile terminal using macro waves, adapting to different situations to provide the most suitable experience for users of the mobile terminal; the mobile terminal communicates with the macro base station using macro waves.

[0090] Suppose that in an ultra-dense millimeter-wave mobile edge computing network supporting non-orthogonal multiple access and orthogonal frequency division multiple access, there exists one macro base station and There are [number] micro base stations, with macro base station number 0, and the micro base station index set is [set]. , and set The micro base station in the system communicates with the mobile terminal using millimeter waves, and enables... , This refers to a set of micro base stations that communicate with mobile terminals using millimeter waves. This indicates the number of millimeter-wave communications used between the micro base station and the mobile terminal. The set of all base stations, including micro base stations and macro base stations, is represented by the index set of mobile terminals. Each mobile terminal The task index set is , Indicates the number of mobile terminals. Indicates the index number of the mobile terminal. Indicates the index number of the task. Indicates the total number of tasks; mobile terminal The The task is , The amount of data for the task; The number of CPU cycles required to compute each bit of data, i.e., computational intensity; The required level of security for the mission; The maximum acceptable safety cost; This represents the maximum tolerable processing delay for the task.

[0091] Each micro base station divides its available spectrum resources based on the principle of equal bandwidth partitioning. There are several orthogonal sub-channels, which are divided into two categories: millimeter-wave channels and macro-wave channels; the index set of orthogonal sub-channels is... , , For orthogonal subchannel indexes, It is the sum of millimeter-wave and macro-wave channels; Indicates the system bandwidth. Orthogonal subchannel bandwidth; association set Mobile terminals connected to micro base stations are millimeter-wave mobile terminals, while mobile terminals connected to other micro base stations are macro-wave mobile terminals. All of them use orthogonal multiple access to share the same orthogonal sub-channel of the micro base station. Mobile terminals connected to macro base stations share the frequency band resources of the macro base station.

[0092] The communication model construction process is as follows:

[0093] Orthogonal subchannels Mobile terminals associated with macro base stations Uplink transmission rate Represented as:

[0094] ;

[0095] In the formula, Indicates the bandwidth of the macrowave channel; For mobile terminals Other mobile terminals with poorer channel gain that cause interference; For mobile terminals Other mobile terminals with worse channel gain that cause interference Decision-making related to macro base stations; Indicates mobile terminal The transmission power; It is the noise power of the macrowave channel; Indicates mobile terminal Channel gain between macro base stations.

[0096] Orthogonal subchannels Mobile terminals that are associated with micro base stations and use millimeter-wave channels for communication Uplink transmission rate Represented as:

[0097] ;

[0098] In the formula, Indicates the bandwidth of the millimeter-wave channel; For mobile terminals The transmission power; Indicates to mobile terminals The transmit power of other mobile terminals with poorer channel gain that cause interference. For mobile terminals With micro base stations Channel gain for millimeter-wave communication; Indicates to mobile terminals Other mobile terminals with worse channel gain that cause interference With micro base stations Channel gain for millimeter-wave communication; This represents the noise power of the millimeter-wave channel. Indicates in sub-channel Up to mobile terminals With micro base stations A set of indexes of mobile terminals that cause interference in communication between them; Indicates to mobile terminals Other mobile terminals with worse channel gain that cause interference With micro base stations Channel gain for communication between them For mobile terminals With micro base stations Channel gain for communication between them; and Mobile terminals The selected orthogonal sub-channels and the micro base station numbers; This refers to the cluster to which the micro base station belongs.

[0099] When satisfied and At that time, the mobile terminal will be in the orthogonal sub-channels of different micro base stations in different clusters. The above received from Interference between base stations outside the cluster.

[0100] In orthogonal sub-channels Mobile terminals that are associated with micro base stations and use macrowave channels for communication The uplink transmission rate is expressed as:

[0101] ;

[0102] In the formula, For mobile terminals With micro base stations Channel gain for macrowave communication between them; Indicates to mobile terminals Other mobile terminals with worse channel gain that cause interference With micro base stations Channel gain for macrowave communication between them; This represents the noise power of the macrowave channel;

[0103] When satisfied and At that time, the mobile terminal will be in the sub-channel of different micro base stations within different clusters. The above received from Interference across base stations outside the cluster;

[0104] The noise power of the macrowave channel and the noise power of the millimeter-wave channel are respectively:

[0105] ;

[0106] in, Indicates mobile terminal Decisions related to macro base stations.

[0107] mobile terminal With micro base stations in sub-channels The channel gain in the equation is given by the following formula:

[0108] ;

[0109] In the formula, Indicates mobile terminal With micro base stations in sub-channels Channel gain in; This represents the path loss during transmission in the channel; Represents the speed of light; This indicates the carrier frequency in the channel; the carrier frequency is 2GHz for macrowave channels and 70GHz for millimeter-wave channels. This indicates the antenna gain of the channel; the antenna gain of the macrowave channel is 0 dBi, the antenna gain value is 18 dBi when the millimeter wave channel is used as the main lobe gain, and the antenna gain value is 2 dBi when it is used as the side lobe gain. Represents the small-scale components of the channel; This represents the path loss of the channel.

[0110] The process of building a caching and offloading model:

[0111] Indicates task In micro base stations The cache status indicator: a value of 1 indicates that the task has been cached locally at the base station. A value of 0 indicates that the file is not cached. In real-world scenarios, mobile terminals' file request behavior follows a specific probability distribution, which is obtained through statistical analysis of historical access data streams; therefore, the cache status indicator for a task can be generated by simulating a process that follows these distributions.

[0112] Specifically, when mobile terminals With micro base stations When establishing an association, the first step is to check if the task exists in the micro base station's storage space. If the task is already cached in the micro base station, then the condition is met. , To represent mobile terminals For the association decision with the micro base station, the entire task is processed directly at the local base station, and the processing result is transmitted back to the mobile terminal via the wireless backhaul link.

[0113] When mobile terminal When establishing a connection with a macro base station, the first step is to check if the task exists in the macro base station's storage space. If the task is already cached on the macro base station, then the condition is met. , Indicates task If the cache status indicator is on the macro base station, the entire task is processed directly on the macro base station, and the processing result is sent directly to the mobile terminal via the wireless backhaul link.

[0114] If the associated micro base station does not cache tasks mobile terminal The system will continue to check whether the macro base station has cached the task. If the task is already cached in the macro base station, the following conditions will be met. In this case, the entire task will be processed on the macro base station, and the processing result will first be sent to the micro base station via a wired link. Then, the data is transmitted back to the mobile terminal via a wireless backhaul link.

[0115] If mobile terminal It has been linked to a macro base station, but the task... It is not cached at this base station, which satisfies the requirement. Then, part of the task will be uploaded to the macro base station for execution.

[0116] If neither the associated micro base station nor the macro base station caches the task That is, satisfying Then the task will be gradually unloaded and calculated step by step.

[0117] The process of building a security model:

[0118] To ensure secure unloading, different cryptographic algorithms are used to encrypt and decrypt the unloading task. Since the data volume of the task's computation result is extremely small, its processing latency and energy consumption are negligible; therefore, the encryption, decryption, and transmission operations of the result are not a concern.

[0119] In uninstallation mode without cache assistance:

[0120] When mobile terminal When associated with a micro base station, it supports any task. First, a portion of the task is executed locally, and the remaining portion is encrypted and offloaded to the micro base station. Upon receiving the data, the micro base station decrypts it and then executes this portion of the task. Simultaneously, the micro base station encrypts another portion of the task and forwards it to the macro base station, which decrypts it and executes the received portion. When the mobile terminal... When directly associated with a macro base station, a portion of the task is first executed locally, while the remaining portion is encrypted and offloaded to the macro base station. The macro base station then decrypts the received task portion and executes it.

[0121] Assume there is a total A cryptographic algorithm, whose index set is Among them, let Indicates the first This algorithm has a protection level of [missing information]. When the task Algorithm When uninstalling; if The probability of encryption failure is then... ;like ,but .

[0122] in, It is a task The safety risk coefficient, The algorithm's expected safety level is set. The algorithm is only considered valid if its safety level is not lower than the task's expected safety level. Only in this way can we ensure absolute success; otherwise, the probability of encryption failure will be high. A security failure has occurred.

[0123] Therefore, the task The cost of security breach of contract is ,in The economic losses incurred should the mission fail. For mobile terminals The A task for cryptographic algorithms Security decision instructions: If an algorithm is selected Processing tasks ,but Otherwise, it is 0. Ultimately, the mobile terminal... The total cost of security breach is .

[0124] Computational model:

[0125] The computation offloading model includes two modes: cache-assisted computation and non-cache-assisted computation. Non-cache-assisted computation includes local computation, offloading to micro base station computation, and offloading to macro base station computation.

[0126] Cache-assisted calculation:

[0127] In cache-assisted computation mode, the task It is already cached in the associated micro base station or the macro base station connected to that micro base station, so it can be executed directly on the micro base station or macro base station. When the mobile terminal Associated micro base stations Cached tasks Although the associated micro base station did not cache the task However, when the macro base station has this cache, it processes the task. Time required for:

[0128] ;

[0129] in, micro base stations Assigned to task computing power Indicates that the macro base station is assigned to the task. The computing power; the first term on the right side of the equation represents the cached task. The computation time on the associated micro base station; the second term on the right side of the equation represents the computation task of the macro base station. The required time is during which the micro base station associated with the mobile terminal has not cached the task, but the macro base station has. In cache-assisted computing mode, local power consumption is negligible.

[0130] Non-cached auxiliary computation:

[0131] In non-cached assisted computation mode, the task Not cached in associated micro base stations It is also not cached on the macro base station connected to the micro base station. At this time, if the mobile terminal is associated with the micro base station, the task... The execution process is as follows: The mobile terminal first calculates the task locally. Part of the data volume ; the remaining data volume The encrypted data is sent to the associated micro base station; upon receiving the ciphertext, the micro base station decrypts it and calculates the amount of data in its sub-part. Then process the remaining unprocessed data. The data is encrypted and forwarded to the macro base station; the macro base station decrypts and processes the received data. It is a mobile terminal The The amount of data computed locally for each task; It is a mobile terminal The Total unloaded data volume for each task; It is a mobile terminal The The task is offloaded to the micro base station. The amount of data; It is a mobile terminal The The task is transmitted via micro base station The amount of data to be calculated; It is a mobile terminal The The task is transmitted via micro base station Decryption Then, the amount of data is unloaded to the macro base station again.

[0132] If the mobile terminal is directly associated with the macro base station, the execution process is simplified to: mobile terminal local computing task Part of the data volume ;The task Remaining data volume The encrypted message is sent to the macro base station; the macro base station decrypts the message upon receiving it and immediately performs calculations.

[0133] Local computing: In non-cached assisted computing mode, mobile terminal The amount of data processed locally is Therefore, mobile terminal processing tasks Local execution time for:

[0134] ;

[0135] in For mobile terminals Computational power; (Unit: CPU cycles / bit) indicates the use of cryptographic algorithms. The computational cost required to encrypt 1 bit of task data. The first item on the right is the computational cost for a mobile terminal in a scenario associated with a micro base station. And for those that are about to be uninstalled The second item is the time consumed for encryption; the third item is the computation time of the mobile terminal in the scenario of associated macro base station. And for those that are about to be uninstalled The time required for encryption.

[0136] During local computing, the mobile terminal Processing tasks Local energy consumption for:

[0137] ;

[0138] in, The energy consumption coefficient is related to the hardware architecture; (Unit: J / bit) indicates the cryptographic algorithm used. Energy consumption for encrypting or decrypting 1 bit of data; For mobile terminals in orthogonal sub-channels The association indicator on the screen indicates that orthogonal sub-channels are selected. but Otherwise, it is 0. The first item on the right is the calculation for the mobile terminal in the scenario of associating with a micro base station. And for those that are about to be uninstalled The second item is the energy consumption generated by encryption; the third item is the computational cost of the mobile terminal in a scenario associated with a macro base station. And for those that are about to be uninstalled The energy consumption generated by encryption; the third item is the energy consumption of mobile terminals. The fourth item is the power consumption of uploading to the macro base station; the fourth item is the power consumption of the mobile terminal. The energy consumption of uploading to the micro base station.

[0139] Unloading to micro base station: In non-cache-assisted unloading mode, when the mobile terminal When associated with a micro base station, the processing task Time required for:

[0140] ;

[0141] in, For sub-channel index set; This refers to the wired backhaul rate between the micro base station and the macro base station. (Unit: CPU cycles / bit) indicates the use of cryptographic algorithms. The computational power required to decrypt a 1-bit task. The task is assigned by the micro base station. computing power The task is assigned by the macro base station. The computing power. The first five items on the right represent, in order: the waiting time for the task to be uploaded to the micro base station; the time for the mobile terminal to upload data to the micro base station; and the time for data transmission. Time; Micro base station calculation Time; Micro base stations forward data to macro base stations Time; Macro base station calculation The time. The last three items represent: micro base station decryption. Time; Micro base station encryption Time; Macro base station decryption The time.

[0142] During the computation process offloaded to the micro base station, resource allocation is calculated proportionally; that is, the micro base station is responsible for the task. The allocated computing power accounts for the portion of the micro base station's processing capacity. The proportion of total CPU cycles. Specifically, when the mobile terminal With micro base stations When establishing an association, the micro base station Assigned to computing power for:

[0143] ;

[0144] in Indicates micro base station Total computing power Indicates to mobile terminals Other mobile terminals with worse channel gain that cause interference With micro base stations Related decisions, This indicates the mobile terminal currently performing the task. The proportion of computing resources occupied This indicates the proportion of computing resources occupied by all mobile terminals associated with this micro base station; and

[0145] ;

[0146] On the right side of the above equation, the first term represents the micro base station's cached data. Time calculation Required CPU cycles; the second item is only applicable when neither micro base stations nor macro base stations are cached. It exists at that time, specifically including: calculation CPU cycles; decryption on micro base stations CPU cycles; encryption on micro base stations CPU cycles.

[0147] Similarly, under the principle of proportional allocation, macro base stations are allocated to tasks. computing power for

[0148] ;

[0149] in, This represents the total computing power of the macro base station;

[0150] ;

[0151] Calculate mobile devices for macro base stations The One task (i.e., task) The required CPU cycles include: when cached At that time, calculate Required CPU cycles; when macro base station is not cached At that time, calculate and decrypt Required CPU cycles. This indicates that the associated macro base station is not cached. The CPU cycles required to compute task data include: if the macro base station has cached the data. ,calculate CPU cycles; if the macro base station is also not cached, calculate and decrypt. CPU cycles, Calculate the mobile terminal for macro base station Other mobile terminals with worse channel gain that cause interference No. The number of CPU cycles required for each task For mobile terminals not cached by the associated macro base station Other mobile terminals with worse channel gain that cause interference No. Each task requires CPU cycles to compute task data.

[0152] Unloading to macro base station: In non-cache-assisted unloading mode, when the mobile terminal When establishing an association with a macro base station, the processing task is... Time required for:

[0153] ;

[0154] The items on the right represent, in order: the mobile terminal will Transmission time uploaded to macro base station (macro base station uncached tasks); macro base station calculation Computation time; local latency; macro base station pair The time required for decryption.

[0155] To align with actual deployment, tasks on each mobile terminal are completed sequentially according to a queue, but the local processing and offloading portions of the same task can be executed in parallel. Therefore, mobile terminals... The total time to complete all tasks is expressed as:

[0156] .

[0157] The optimization problem can be expressed as:

[0158] ;

[0159] To optimize computing and communication resources, reduce transmission overhead, ensure secure offloading, and extend terminal standby time, this paper conducts joint optimization with the objective of minimizing local energy consumption, under the joint constraints of proportional computing resource allocation, processing latency constraints, and security breach cost thresholds; where, Let represent the objective function to be optimized. This represents the total energy consumption calculated locally. To optimize the unified representation of the problem-related set, we denote it as the variable set; Indicates mobile terminal and micro base stations The set of related decisions between them Indicates mobile terminal and micro base stations Related decisions; For mobile terminals The A task for cryptographic algorithms A set of security decision instructions, For mobile terminals The A task for cryptographic algorithms Security decision instructions A set of cryptographic algorithm indexes. The index number representing the cryptographic algorithm. Indicates the total number of cryptographic algorithms; For mobile terminals In orthogonal subchannels The set of associated indicators on, For mobile terminals In orthogonal subchannels The associated instructions on; Frequency band allocation factor; For mobile terminals Transmit power set, For mobile terminals Transmission power; For mobile terminals The collection of local computing capabilities, For mobile terminals Local computing power; For mobile terminals The The task is offloaded to the micro base station. The data set For mobile terminals The The task is offloaded to the micro base station. The amount of data; For mobile terminals The The task is transmitted via micro base station Decryption Then, the data set is unloaded to the macro base station again. For mobile terminals The The task is transmitted via micro base station Decryption Then, the amount of data unloaded to the macro base station again; For the task At the base station The set of cache status indicators on For the task At the base station The cache status indicator on;

[0160] The meaning of constraints: It is a latency constraint, representing the total time required for all tasks to be completed on the mobile terminal. It must not exceed its maximum tolerable delay. ; It is a security constraint, representing the total security breach cost of a mobile terminal. It must not exceed its preset upper limit. ; and Together they constitute an association constraint, meaning that each mobile terminal can only be associated with one base station; and Together they constitute the encryption algorithm constraint, meaning that each task can only use one encryption algorithm; and Together they constitute the sub-channel constraint, meaning that each sub-channel can be allocated to at most one mobile terminal; and Together, they constitute the cache quantity constraint, meaning that each task can be cached at most in one base station; This is a total cache limit, meaning that the total number of cached tasks at a base station cannot exceed its maximum cache capacity. ; It is a computing power constraint, representing the local computing power of the mobile terminal. The upper limit is ; It is a frequency band allocation factor constraint, representing the frequency band allocation factor. The value range is (0, 1); It is a transmit power constraint, representing the transmit power of the mobile terminal. The upper limit is ; This is an unloading data volume constraint, indicating that the data volume of multiple unloading steps must meet the constraint. , For mobile terminals The Total amount of unloaded data for each task.

[0161] Step S2 can be represented as:

[0162] Optimization problem Each feasible solution is encoded as a vector individual, and the index set is... For any vector individual To compile all sets related to the optimization problem , , , , , , , and One-to-one mapping to multiple component sets: For vector individuals China Mobile Terminal Selected base station index set, For vector individuals China Mobile Terminal Selected base station index; For vector individuals China Mobile Terminal The set of selected cryptographic algorithm numbers, For vector individuals China Mobile Terminal The selected cryptographic algorithm number, A collection of mobile terminals containing different tasks; For vector individuals China Mobile Terminal The set of sub-channel indices assigned, For vector individuals China Mobile Terminal The assigned sub-channel index; For vector individuals Frequency band allocation factor; For vector individuals China Mobile Terminal The set of computing power, For vector individuals China Mobile Terminal Computational power; For vector individuals China Mobile Terminal Selected cache base station index set, For vector individuals China Mobile Terminal Selected cache base station index; For vector individuals China Mobile Terminal The set of transmit power, For vector individuals China Mobile Terminal The transmission power; For vector individuals China Mobile Terminal The set of data volumes offloaded to associated base stations For vector individuals China Mobile Terminal The amount of data offloaded to associated base stations; For vector individuals China Mobile Terminal The set of data forwarded from the associated micro base station to the macro base station. For vector individuals China Mobile Terminal The amount of data forwarded from the associated micro base station to the macro base station.

[0163] To optimize the fitness of individual evaluation vectors, designing a reasonable fitness function is crucial. As can be seen from the optimization problem, constraints... , and Its nonlinear, mixed-integer, and coupled characteristics make it difficult for individuals to be directly satisfied during the search process. Therefore, [the following is a more detailed explanation:] and Embedding the fitness function as a penalty term forces the population away from infeasible regions while preserving its ability to converge toward feasible optimal solutions. For To address the amplitude problem and accelerate optimization, a novel strategy for random retention within the capacity space is proposed: for any base station After obtaining the cache decision instructions, a portion of the instructions are randomly retained to meet the cache limit, while the remaining instructions are set to 0. This random retention mechanism enhances the exploration capability of the improved gradient optimization algorithm, helping to discover better solutions. This achieves the goal of minimizing the total energy consumption of all mobile terminals while also considering... and ,individual The fitness function is formally defined as:

[0164] ;

[0165] in For the first A set of encoded variables for each vector individual; The fitness function; and Mobile terminals The penalty coefficients corresponding to the time delay constraints and security constraints.

[0166] The specific operation of step S3 is as follows:

[0167] Step S3.1: Set the maximum number of iterations , This represents the current iteration number, starting from 1.

[0168] Step S3.2: Initialize the population. The initial population can be generated according to the following rules:

[0169] ;

[0170] In the formula, Represents individual vectors Other mobile terminals The initial solution for the selected base station index. For vector individuals Other mobile terminals The initial solution for the assigned sub-channel index. For vector individuals The initial solution of the frequency band division factor, For vector individuals Other mobile terminals The initial solution based on computational power. For vector individuals Other mobile terminals The initial solution for the transmit power, For vector individuals Mobile terminals related to the task The initial solution of the selected cryptographic algorithm number, For vector individuals Mobile terminals related to the task The initial solution for the selected cache base station index. For vector individuals Mobile terminals related to the task The amount of data offloaded to associated base stations For vector individuals Mobile terminals related to the task The amount of data forwarded from the associated micro base station to the macro base station, Represents from discrete sets A uniformly and randomly selected element is chosen from the middle; Indicates the interval A random real number is generated within the range; This represents mapping individual linear index vectors to... Matrix row index Subscript ; This indicates that a random hold operation is performed on the input variable within its capacity space.

[0171] Step S3.3: Perform gradient search rule operations on the population; gradient search rule operations further introduce stochastic characteristics into the optimization process to enhance exploration capabilities and avoid getting trapped in local optima. Its movement direction mechanism adaptively determines the search direction, thereby accelerating the convergence rate of the improved gradient optimization algorithm; the current position of the individual in the t-th iteration is updated according to the following rules. The updated location of an individual. The specific rules are as follows:

[0172] ;

[0173] In the formula, , For a set containing 4 components The first set of encoded variables, For a set containing 4 components The second set of encoded variables, Let be the current individual position in the t-th iteration of the first set of encoded variables. Let be the current individual position in the t-th iteration of the second set of encoded variables. , The positions of individuals after being updated by gradient search rules are respectively the first set of encoded variables and the second set of encoded variables. , These represent the positions of the worst-performing and best-performing individuals in the history of the first set of encoded variables after the t-th iteration. , These are the positions of the worst and best historical individuals in the second set of encoded variables after the t-th iteration, respectively. This indicates a rounding function; Represents a standard normally distributed random number; Represents a random number between (0, 1); It is a small constant; For element-wise multiplication, To divide element by element, These are the first adaptive iteration coefficient and the second adaptive iteration coefficient, respectively, and they are numerically equal. These are the directions of the third gradient search and the second gradient search, respectively. , , , , ;in,

[0174] ;

[0175] This represents the maximum number of iterations. These are gradient coefficients. It is an adaptive gradient bound. It is the gradient rule constant for the second operation. It is the gradient rule constant for the third operation. and Parameters The minimum and maximum values. For the vector individual of the t-th iteration of the first set of encoded variables Location, The vector individual in the t-th iteration of the second encoded variable set Location, Four distinct and different Vector individuals; Indicates generation and A random vector of the same dimension, each element in Uniform sampling within the interior; Represents generation and A random vector of the same dimension, each element in Uniform sampling within the area.

[0176] To further improve the algorithm, the idea of ​​Newton's method is introduced to update the population position, and an alternative position update rule is proposed. This is different from using only a single vector. Unlike other rules, the new rules incorporate the vector arithmetic mean. and This aims to optimize the search dynamics within the solution space, thereby enhancing overall optimization performance. Accordingly, the current position is updated according to the following correction rules. You can get a new location. :

[0177]

[0178]

[0179] In the formula, To introduce the updated individual position using the Newton method through gradient search rule operations, , The positions of the individuals updated by Newton's method are introduced from the first and second sets of encoded variables respectively through gradient search rule operations. These are the first and third gradient coefficients introduced into Newton's method, respectively. These are the second and fourth gradient coefficients introduced by Newton's method, respectively. This is the set of gradient search direction coefficients. These are the sets of search direction coefficients for the first and second gradients, respectively, when Newton's method is introduced.

[0180] Based on the individual's current position in the t-th iteration And the two newly generated candidate positions and The next generation of individuals The solution (location) is determined as follows:

[0181] The specific formula is as follows:

[0182] ;

[0183] In the formula, , These are the first and second sets of encoded variables after the (t+1)th iteration, respectively. This means randomly generating a real number within the interval (0,1).

[0184] Step S3.4: Perform the local escape operator operation on the population again. In traditional gradient optimization algorithms, the local escape operator operation is introduced to improve its efficiency in solving complex problems. Unlike traditional designs, this invention causes the entire population to sequentially execute the gradient search rule and the local escape operator operation; after each operation, boundary conditions are handled and a greedy selection is performed, and the historical worst and best position information is updated. This modification ensures the high efficiency of the algorithm. The local escape operator operation can update the individual positions after the (t+1)th iteration. It makes significant adjustments; it integrates multiple location components—including the historical best location and two previously generated candidate locations. and Two randomly selected locations , And a new random position —Constructing an enhanced solution .

[0185] Specifically, in terms of probability The individual position after the (t+1)th iteration Perform the following enhancement operations:

[0186] ;

[0187] ;

[0188] in, For the vector individual other than vector individual m in the t-th iteration Current location For the vector individual other than vector individual m in the t-th iteration Current location For the random vector individual after the t-th iteration The updated location For the random vector individuals of the first set of encoded variables after the t-th iteration The updated location For the random vector individuals of the second set of encoded variables after the t-th iteration The updated location For interval A random number that is uniformly distributed within the range; These are random numbers distributed according to a standard normal distribution. for The random number within is given by the following formula:

[0189] ;

[0190] in For about variables binary functions, like ,but ;otherwise .

[0191] Step S3.5: Perform diversity-guided mutation operations on the population to introduce a diversity measure, used to study the alternation between exploitative and exploratory behaviors. This is a diversity measure for multidimensional numerical problems. It can be defined as:

[0192] ;

[0193] In the formula, They represent the sets related to the optimization problem, respectively. , , , , , The length of the diagonal of the feasible region; where:

[0194] ;

[0195] In the formula, For vector individuals China Mobile Terminal The average value of the selected base station indexes. For vector individuals China Mobile Terminal Average number of selected cryptographic algorithms For vector individuals China Mobile Terminal The average of the assigned sub-channel indices, For vector individuals China Mobile Terminal The average computing power, For vector individuals China Mobile Terminal The average value of the selected cache base station indexes. For vector individuals China Mobile Terminal The average transmit power.

[0196] The mutation operation is performed with a certain probability, that is:

[0197] ;

[0198] In the formula, All are constant coefficients greater than 0 and less than 1, representing the diversity threshold. These represent the first, second, and third mutation probabilities. It is the probability of performing this mutation operation. There are 3 possible values; therefore, when the mutation probability of the diversity threshold meets the requirements, a mutation operation will be performed on the individuals in the population again.

[0199] Step S3.6: Determine whether the current iteration count has reached the maximum iteration count. If not, execute the above operations sequentially, and increment the iteration count after execution. If the maximum iteration count has been reached, output the current result.

[0200] Step S3.7: Calculate the fitness value of all individuals in the current population using the fitness function, select the individual with the highest fitness value as the current best individual and compare it with the initial best individual, take the individual with the larger fitness value as the historical best individual, and output the corresponding position information.

[0201] The effects of the embodiments of this invention can be further illustrated through simulation.

[0202] The simulation conditions are set as follows: base stations and mobile terminals are randomly distributed within a macrocell with a radius of 500m; there are 31 base stations (30 micro base stations and 1 macro base station), each base station can cache a maximum of 3 tasks; there are 30 mobile terminals, each with 3 subtasks; the system bandwidth is 20MHz, the maximum computing power of the base stations is 20GHz, and the maximum computing power of the mobile terminals is 1GHz; there are 6 cryptographic algorithms, each requiring [100 200 250 300 350 1050] CPU cycles to encrypt one bit of data, and [90 280 350 300 400 1700] CPU cycles to decrypt one bit of data, with energy consumption of [2.5296 5.0425 6.837 7.8528 8.7073 26.3643]*1e -7The task data size of the mobile terminal is 200~500KB, the maximum allowable latency of the task is 5~10s, and it requires 50~100 CPU cycles; the weight coefficient in the optimization objective function is 0.5; the energy coefficient of the mobile terminal is 10⁻²⁴; the energy coefficient of the base station is 10⁻²⁶; the task security incurs costs due to protection failure, and the task security coefficient is {5,6}. The simulation results are compared as follows.

[0203] CLC algorithm (Completely Local Computing): All tasks of all mobile terminals are computed locally.

[0204] The SSHO algorithm (Strongest Single-Hopping Offloading) offloads all mobile terminal tasks to the base station with the highest channel gain at a randomly selected frequency.

[0205] GBO algorithm (Gradient-Based Optimizer): also known as the traditional gradient optimization algorithm.

[0206] IGBO algorithm (Improved Gradient-Based Optimizer): an improved gradient optimization algorithm.

[0207] like Figure 2 As shown, when the mobile terminal task data volume increases from 100 kb to 500 kb, the local total energy consumption of the CLC algorithm, SSHO algorithm, GBO algorithm, and IGBO algorithm all increase with the increase in task data volume. However, the local total energy consumption of the IGBO algorithm proposed in this invention is always the lowest among the algorithms except for SSHO; this is consistent with the actual situation.

[0208] like Figure 3 As shown, the fitness values ​​of both the GBO and IGBO algorithms increase with the number of iterations, and eventually converge to a certain value, indicating that both algorithms have convergence. However, the fitness value of the IGBO algorithm is always higher than that of the GBO algorithm, indicating that the IGBO algorithm performs better than the GBO algorithm.

[0209] Another embodiment of the present invention also provides a non-volatile computer storage medium storing computer-executable instructions that execute the cache and secure multi-hop offloading method in the ultra-dense millimeter-wave network.

[0210] Another embodiment of the present invention provides an electronic device, including: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to cause the at least one processor to perform the cache and secure multi-hop offloading method in the ultra-dense millimeter-wave network.

[0211] 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 method for offloading buffers and secure multi-hop components in ultra-dense millimeter-wave networks, characterized in that, Includes the following steps: Step S1: Obtain the basic network information of an ultra-dense millimeter-wave mobile edge computing network supporting non-orthogonal multiple access and orthogonal frequency division multiple access, and construct the network system. The network system includes a communication model, a cache and offload model, a security model, and a computation model. Construct an optimization problem under the constraints of the network system. The specific process of constructing the cache and offload model is as follows: A cache status indicator is set on the base station for each task. When the indicator value is 1, it means that the task has been cached on the base station, and when the value is 0, it means that it has not been cached. The cache status indicator is generated based on the historical access probability distribution of files by mobile terminals. When a mobile terminal establishes an association with a micro base station, it first checks whether the task exists in the micro base station's storage space. If the task is already cached in the micro base station, the entire task is processed directly at the local base station, and the processing result is transmitted back to the mobile terminal via a wireless backhaul link. If the associated micro base station does not cache the task, the mobile terminal will continue to check whether the macro base station has cached the task; if the task has been cached in the macro base station, the entire task will be processed on the macro base station, and the processing result will first be sent to the micro base station through the wired link, and then sent back to the mobile terminal through the wireless backhaul link. When a mobile terminal establishes an association with a macro base station, it first checks whether the task exists in the macro base station's storage space. If the task is already cached on the macro base station, the entire task is processed directly on the macro base station, and the processing result is sent directly to the mobile terminal via the wireless backhaul link. Step S2: Encode the feasible solutions to the optimization problem; Step S3: Generate an initial population based on the encoding results according to certain rules, and iterate the initial population using an improved gradient optimization algorithm. During the iteration process, in addition to the gradient search rule operation and local escape operator operation in the original gradient optimization algorithm, a diversity-guided mutation operation is performed to update the population position information. Based on the position information, the best individual in the current population is obtained and compared with the initial best individual. The historical best individual is selected and the position information of the individual is output. Step S4: Perform task calculation and unload resource configuration based on the location information of the selected best individual in the population history.

2. The method for offloading buffer and secure multi-hop components in an ultra-dense millimeter-wave network according to claim 1, characterized in that, An ultra-dense millimeter-wave mobile edge computing network supporting both non-orthogonal multiple access (NOAMI) and orthogonal frequency division multiple access (OFDMA) comprises macro base stations, several micro base stations, and several mobile terminals. Each micro base station is equipped with a mobile edge computing server. All micro base stations are connected to the nearest macro base station via wired backhaul links. The number of micro base stations is greater than or equal to the number of mobile terminals. Each mobile terminal must complete its tasks sequentially under the dual constraints of security and time limits. For computationally intensive and latency-sensitive tasks, the mobile terminal communicates with some micro base stations using millimeter waves, while the remaining micro base stations communicate with the mobile terminal using macro waves; the mobile terminal and macro base stations communicate using macro waves. Suppose there exists a macro base station in an ultra-dense millimeter-wave mobile edge computing network supporting non-orthogonal multiple access and orthogonal frequency division multiple access. There are [number] micro base stations, with macro base station number 0, and the micro base station index set is [set]. , and set The micro base station in the system communicates with the mobile terminal using millimeter waves, and enables... , This refers to a set of micro base stations that communicate with mobile terminals using millimeter waves. This indicates the number of millimeter-wave communications used between the micro base station and the mobile terminal. The set of all base stations, including micro base stations and macro base stations, is represented by the index set of mobile terminals. Each mobile terminal The task index set is , Indicates the number of mobile terminals. Indicates the index number of the mobile terminal. Indicates the index number of the task. Indicates the total number of tasks; mobile terminal The The task is , The amount of data for the task; The number of CPU cycles required to compute each bit of data, i.e., computational intensity; The required level of security for the mission; The maximum acceptable safety cost; The maximum tolerable processing delay for the task; Each micro base station divides its spectrum resources based on the principle of equal bandwidth allocation. There are several orthogonal sub-channels, which are divided into two categories: millimeter-wave channels and macro-wave channels; the index set of orthogonal sub-channels is... , , For orthogonal subchannel indexes, It is the sum of millimeter-wave and macro-wave channels; Indicates the system bandwidth. Orthogonal subchannel bandwidth; association set Mobile terminals connected to micro base stations are millimeter-wave mobile terminals, while mobile terminals connected to other micro base stations are macro-wave mobile terminals. All of them use orthogonal multiple access to share the same orthogonal sub-channel of the micro base station. Mobile terminals connected to macro base stations share the frequency band resources of the macro base station.

3. The method for offloading buffer and secure multi-hop components in an ultra-dense millimeter-wave network according to claim 2, characterized in that, The specific process of constructing the communication model is as follows: Orthogonal subchannels Mobile terminals associated with macro base stations Uplink transmission rate Represented as: ； In the formula, Indicates the bandwidth of the macrowave channel; For mobile terminals Other mobile terminals with poorer channel gain that cause interference; For mobile terminals Other mobile terminals with worse channel gain that cause interference Decision-making related to macro base stations; Indicates mobile terminal The transmission power; It is the noise power of the macrowave channel; Indicates mobile terminal Channel gain between macro base stations; Orthogonal subchannels Mobile terminals that are associated with micro base stations and use millimeter-wave channels for communication Uplink transmission rate Represented as: ； In the formula, Indicates the bandwidth of the millimeter-wave channel; For mobile terminals The transmission power; Indicates to mobile terminals The transmit power of other mobile terminals with poorer channel gain that cause interference. For mobile terminals With micro base stations Channel gain for millimeter-wave communication; Indicates to mobile terminals Other mobile terminals with worse channel gain that cause interference With micro base stations Channel gain for millimeter-wave communication; This represents the noise power of the millimeter-wave channel. Indicates in sub-channel Up to mobile terminals With micro base stations A set of indexes of mobile terminals that cause interference in communication between them; Indicates to mobile terminals Other mobile terminals with worse channel gain that cause interference With micro base stations Channel gain for communication between them For mobile terminals With micro base stations Channel gain for communication between them; and Mobile terminals The selected orthogonal sub-channels and the micro base station numbers, and These respectively represent the mobile terminal Other mobile terminals with worse channel gain that cause interference The selected orthogonal sub-channels and the micro base station numbers; The cluster to which the micro base station belongs; When satisfied and At that time, the mobile terminal will be in the orthogonal sub-channels of different micro base stations in different clusters. The above received from Interference across base stations outside the cluster; Orthogonal subchannels Mobile terminals that are associated with micro base stations and use macrowave channels for communication The uplink transmission rate is expressed as: ； In the formula, For mobile terminals With micro base stations Channel gain for macrowave communication between them; Indicates to mobile terminals Other mobile terminals with worse channel gain that cause interference With micro base stations Channel gain for macrowave communication between them; When satisfied and At that time, the mobile terminal will be in the sub-channel of different micro base stations within different clusters. The above received from Interference across base stations outside the cluster; The noise power of the macrowave channel and the noise power of the millimeter-wave channel are respectively: ； in, Indicates mobile terminal Decision-making related to macro base stations; mobile terminal With micro base stations In sub-channel The channel gain in the equation is given by the following formula: ； In the formula, Indicates mobile terminal With micro base stations In sub-channel Channel gain in; This represents the path loss during transmission in the channel; Represents the speed of light; This indicates the carrier frequency in the channel; the carrier frequency is 2GHz for macrowave channels and 70GHz for millimeter-wave channels. This indicates the antenna gain of the channel; the antenna gain of the macrowave channel is 0 dBi, the antenna gain value is 18 dBi when the millimeter wave channel is used as the main lobe gain, and the antenna gain value is 2 dBi when it is used as the side lobe gain. Represents the small-scale components of the channel; This represents the path loss of the channel.

4. The method for offloading buffer and secure multi-hop components in an ultra-dense millimeter-wave network according to claim 1, characterized in that, The optimization problem is represented as: ； In the formula, Let represent the objective function to be optimized. This represents the total energy consumption calculated locally. To optimize the unified representation of the problem-related set, we denote it as the variable set; Indicates mobile terminal and micro base stations The set of related decisions between them Indicates mobile terminal and micro base stations Related decisions; For mobile terminals The A task for cryptographic algorithms A set of security decision instructions, For mobile terminals The A task for cryptographic algorithms Security decision instructions A set of cryptographic algorithm indexes. The index number representing the cryptographic algorithm. Indicates the total number of cryptographic algorithms; For mobile terminals In orthogonal subchannels The set of associated indicators on, For mobile terminals In orthogonal subchannels The associated instructions on; Frequency band allocation factor; For mobile terminals Transmit power set, For mobile terminals Transmission power; For mobile terminals The collection of local computing capabilities, For mobile terminals Local computing power; For mobile terminals The The task is offloaded to the micro base station. The data set For mobile terminals The The task is offloaded to the micro base station. The amount of data; For mobile terminals The The task is transmitted via micro base station Decryption Then, the data set is unloaded to the macro base station again. For mobile terminals The The task is transmitted via micro base station Decryption Then, the amount of data unloaded to the macro base station again; For the task At the base station The set of cache status indicators on For the task At the base station The cache status indicator on; The meaning of constraints: It is a latency constraint, representing the total time required for all tasks to be completed on the mobile terminal. It must not exceed its maximum tolerable delay. ; It is a security constraint, representing the total security breach cost of a mobile terminal. It must not exceed its preset upper limit. ; and Together they constitute an association constraint, meaning that each mobile terminal can only be associated with one base station; and Together they constitute the encryption algorithm constraint, meaning that each task can only use one encryption algorithm; and Together they constitute the sub-channel constraint, meaning that each sub-channel can be allocated to at most one mobile terminal; and Together, they constitute the cache quantity constraint, meaning that each task can be cached at most in one base station; This is a total cache limit, meaning that the total number of cached tasks at a base station cannot exceed its maximum cache capacity. ; It is a computing power constraint, representing the local computing power of the mobile terminal. The upper limit is ; It is a frequency band allocation factor constraint, representing the frequency band allocation factor. The value range is (0, 1); It is a transmit power constraint, representing the transmit power of the mobile terminal. The upper limit is ; This is an unloading data volume constraint, indicating that the data volume of multiple unloading steps must meet the constraint. , For mobile terminals The Total unloaded data volume for each task.

5. The method for offloading buffer and secure multi-hop components in an ultra-dense millimeter-wave network according to claim 4, characterized in that, Step S2 is represented as: Optimization problem Each feasible solution is encoded as a vector individual, and the index set is... For any vector individual To compile all sets related to the optimization problem , , , , , , , and One-to-one mapping to multiple component sets: For vector individuals China Mobile Terminal Selected base station index set, For vector individuals China Mobile Terminal Selected base station index; For vector individuals China Mobile Terminal The set of selected cryptographic algorithm numbers, For vector individuals China Mobile Terminal The selected cryptographic algorithm number, A collection of mobile terminals containing different tasks; For vector individuals China Mobile Terminal The set of sub-channel indexes assigned, For vector individuals China Mobile Terminal The assigned sub-channel index; For vector individuals Frequency band allocation factor; For vector individuals China Mobile Terminal The set of computing power, For vector individuals China Mobile Terminal Computational power; For vector individuals China Mobile Terminal Selected cache base station index set, For vector individuals China Mobile Terminal Selected cache base station index; For vector individuals China Mobile Terminal The set of transmit power, For vector individuals China Mobile Terminal The transmission power; For vector individuals China Mobile Terminal The set of data volumes offloaded to associated base stations For vector individuals China Mobile Terminal The amount of data offloaded to associated base stations; For vector individuals China Mobile Terminal The set of data forwarded from the associated micro base station to the macro base station. For vector individuals China Mobile Terminal The amount of data forwarded from the associated micro base station to the macro base station.

6. The method for offloading buffer and secure multi-hop components in an ultra-dense millimeter-wave network according to claim 5, characterized in that, The specific operation of step S3 is as follows: Step S3.1: Set the maximum number of iterations , This represents the current iteration number, starting from 1. Step S3.2: Initialize the population individuals by generating the initial population according to the following rules; ； In the formula, Represents individual vectors Other mobile terminals The initial solution for the selected base station index. For vector individuals Other mobile terminals The initial solution for the assigned sub-channel index. For vector individuals The initial solution of the frequency band division factor, For vector individuals Other mobile terminals The initial solution based on computational power. For vector individuals Other mobile terminals The initial solution for the transmit power, For vector individuals Mobile terminals related to the task The initial solution of the selected cryptographic algorithm number, For vector individuals Mobile terminals related to the task The initial solution for the selected cache base station index. For vector individuals Mobile terminals related to the task The amount of data offloaded to associated base stations For vector individuals Mobile terminals related to the task The amount of data forwarded from the associated micro base station to the macro base station, Represents from discrete sets A uniformly and randomly selected element is chosen from the middle; Indicates the interval A random real number is generated within the range; This represents mapping individual linear index vectors to... Matrix row index Subscript ; This indicates that a random hold operation is performed on the input variable within its capacity space; Step S3.3: Let the current position of the individual in the t-th iteration be... Gradient search rule operations are performed on the population to obtain the updated positions of individuals after the gradient search rule operations. The process is as follows: ； In the formula, , For a set containing 4 components The first set of encoded variables, For a set containing 4 components The second set of encoded variables, Let be the current individual position in the t-th iteration of the first set of encoded variables. Let be the current individual position in the t-th iteration of the second set of encoded variables. , The positions of individuals after being updated by gradient search rules are respectively the first set of encoded variables and the second set of encoded variables. , These represent the positions of the worst-performing and best-performing individuals in the history of the first set of encoded variables after the t-th iteration. , These are the positions of the worst and best historical individuals in the second set of encoded variables after the t-th iteration, respectively. This indicates a rounding function; Represents a standard normally distributed random number; Represents a random number between (0, 1); It is a small constant; For element-wise multiplication, To divide element by element, , These are the first adaptive iteration coefficient and the second adaptive iteration coefficient, respectively, and they are numerically equal. These are the directions of the third gradient search and the second gradient search, respectively. Then, the idea of ​​Newton's method is introduced to update the population position. The position update rule is as follows: ； in: , To introduce the updated individual position using the Newton method through gradient search rule operations, , The positions of the individuals updated by Newton's method are introduced from the first and second sets of encoded variables respectively through gradient search rule operations. These are the first and third gradient coefficients introduced into Newton's method, respectively. The second and fourth gradient coefficients of Newton's method are introduced respectively; Based on the individual's current position in the t-th iteration And the two newly generated updated positions and Next generation vector individuals The location is determined as follows: ； in, , These are the first and second sets of encoded variables after the (t+1)th iteration, respectively. This means randomly generating a real number within the interval (0,1); Step S3.4: Perform compound operations on the population; sequentially execute the gradient search rule and local escape operator operations on the entire population; after each operation, perform boundary processing and greedy selection, and update the position information of the worst and best historical values. The specific process is as follows: With probability The individual position after the (t+1)th iteration Perform the following operations: ； ； in, , These are the vector individuals of the first set of encoded variables after the t-th iteration. Location, For the random vector individuals of the first set of encoded variables after the t-th iteration The updated location These are the vector individuals of the second set of encoded variables after the t-th iteration. Location, For the random vector individuals of the second set of encoded variables after the t-th iteration The updated location For interval A random number that is uniformly distributed within the range; These are random numbers distributed according to a standard normal distribution. for Random numbers within; Step S3.5: Perform diversity-guided mutation operations on the population. The specific process is as follows: Diversity measure of multidimensional numerical problems Defined as: ； In the formula, They represent the sets related to the optimization problem, respectively. , , , , , The length of the diagonal of the feasible region; For vector individuals China Mobile Terminal The average value of the selected base station indexes. For vector individuals China Mobile Terminal Average number of selected cryptographic algorithms For vector individuals China Mobile Terminal The average of the assigned sub-channel indices, For vector individuals China Mobile Terminal The average computing power, For vector individuals China Mobile Terminal The average value of the selected cache base station indexes. For vector individuals China Mobile Terminal The average transmit power; Performing mutation operations according to probability is... ； In the formula, All are constant coefficients greater than 0 and less than 1, representing the diversity threshold. These represent the first, second, and third mutation probabilities. It is the probability of performing the mutation operation. Three possible values; when the mutation probability of the diversity threshold meets the requirements, the mutation operation will be performed again on the individuals in the population. Step S3.6: Determine if the current iteration count has reached the maximum iteration count. If not, execute steps S3.3, S3.4, and S3.5 sequentially. After execution, increment the iteration count by one. If the maximum iteration count has been reached, output the current result. Step S3.7: Calculate the fitness value of all individuals in the current population using the fitness function, select the individual with the highest fitness value as the current best individual and compare it with the initial best individual, take the individual with the larger fitness value as the historical best individual, and output the corresponding position information.

7. A method for offloading buffer and secure multi-hop components in an ultra-dense millimeter-wave network according to claim 6, characterized in that, Vector Individual The fitness function is defined as: ； In the formula, For vector individuals The set of encoded variables; The fitness function; and Mobile terminals The penalty coefficients corresponding to the time delay constraints and security constraints.

8. A non-volatile computer storage medium storing computer-executable instructions, characterized in that, The computer can execute instructions to perform the cache and secure multi-hop offloading method in ultra-dense millimeter-wave networks as described in any one of claims 1-7.

9. An electronic device, comprising: At least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, characterized in that the instructions are executed by the at least one processor to cause the at least one processor to perform the cache and secure multi-hop offloading method in an ultra-dense millimeter-wave network according to any one of claims 1-8.

Images