Secure computation offloading method for cache-assisted ultra-dense heterogeneous MEC networks

Abstract

The present invention discloses a secure computing offloading method for a cache-assisted ultra-dense heterogeneous MEC network, the method comprising: obtaining network basic information of an ultra-dense multi-access mobile edge computing network, constructing a network system according to the network basic information, and constructing an optimization problem under the constraints of the network system; obtaining an initial solution according to the optimization problem, and defining the initial solution as an initial population, using an improved whale algorithm to search the initial population to obtain a target population, and outputting the position of the global optimal whale in the target population; performing secure computing efficiency optimization configuration according to the global optimal whale position. The present invention has multi-base station offloading, considers cache factors, meets the constraints of rate, mobile device energy consumption, latency, security vulnerability cost, and task and transmission power ratio, and can well achieve the goal of minimizing the local total energy consumption of all mobile devices.

Description

Technical Field

[0001] The present invention relates to the field of wireless communication technologies, and in particular to a secure computing offloading method for a cache-assisted ultra-dense heterogeneous MEC network. Background Art

[0002] The rapid development of 5G networks has greatly promoted the popularization of smart devices, which have extremely high requirements for computing performance and response time. These requirements are particularly evident in many fields, including biometric technology, augmented reality and virtual reality, natural language processing for artificial intelligence, interactive games, IoT applications for healthcare, and autonomous driving technology. However, due to the limitations of traditional devices in processing power and battery life, the experience of mobile devices has been affected.

[0003] To solve this problem, mobile edge computing (MEC) technology came into being. It deploys computing resources at the edge of wireless networks, allowing devices to transfer some tasks to edge servers, thereby reducing the burden on the devices themselves. At the same time, the cache function in mobile edge computing can also provide more reliable services for mobile devices.

[0004] The concept of caching in mobile edge computing (MEC) refers to temporarily storing data, content or services on devices at the edge of the network to reduce latency, increase access speed and reduce the burden on the core network. This caching mechanism can store data that mobile devices request to access frequently. When the mobile device requests the same data again, it can be quickly provided directly from the edge node without having to go through a long backhaul to the core network.

[0005] The role of cache is mainly reflected in reducing latency, reducing bandwidth consumption, improving system capacity, enhancing mobile device experience and saving energy. In the MEC environment, cache can not only store static content, but also store tasks and related data to reduce the latency of task offloading. When a mobile device issues a task offloading request, if the task has been cached on the edge cloud, the edge cloud can directly process the task and transmit the result to the mobile device without the mobile device having to offload the task again.

[0006] With the growing demand for reducing the end-to-end latency of 5G networks, people have begun to explore ultra-dense multi-access mobile edge computing networks. This network uses densely deployed small base stations to enhance signal coverage and allows mobile devices to establish connections with multiple base stations to make better use of the computing power of these base stations. However, this approach also brings new challenges. For example, the dense deployment of small base stations may cause serious interference problems. At the same time, in order to ensure the safe offloading of tasks, complex cryptographic techniques need to be introduced, which requires finding a balance between security and computing resource consumption. At the same time, the optimization of caching mechanisms is also an important direction in MEC research. Researchers are exploring how to manage edge caches more effectively to achieve joint optimization of task offloading and caching, so as to minimize energy consumption while ensuring service quality.

[0007] In addition, although the existing whale algorithm has a fast convergence speed, it is insufficient in local search capability, which may limit its effectiveness in practical applications. Summary of the invention

[0008] In view of the shortcomings of the prior art, the present invention provides a secure computing offloading method for cache-assisted ultra-dense heterogeneous MEC networks, which aims to solve the problem of increased system energy consumption caused by the lack of research on new problems arising from the consideration of cache issues during computing offloading in the prior art.

[0009] To achieve the above object, the present invention provides the following technical solution: a secure computing offloading method for cache-assisted ultra-dense heterogeneous MEC networks, comprising the following steps:

[0010] Step S1: Obtain network basic information of an ultra-dense multi-access mobile edge computing network and construct a network system, wherein the network system includes a communication model, a computing model, and a security model; construct an optimization problem under the constraints of the network system;

[0011] Step S2: Obtain an initial solution according to the optimization problem, and define the initial solution as the parent population. Use the improved whale algorithm to search the parent population to obtain the target population, and output the global optimal solution in the target population.

[0012] The process of using the improved whale algorithm to search the parent population to obtain the target population includes: first, the population is initialized and the historical optimal whale individual is determined; then, the partial matching crossover operation and mutation operation of the genetic algorithm are used to expand the whale population and increase the population diversity; then, in the shrinking and encircling stage, an adaptive nonlinear weight factor is used to update the position of the whale individual, and it is also used in the spiral bubble net attack and prey search stages; in the spiral bubble net attack stage, the inertia weight coefficient is used to adaptively adjust the spiral amplitude to avoid falling into the local optimum; in the prey search stage, the long tail mutation operation of the whale using the inverse cumulative distribution function of Cauchy is used to expand the search space of the improved whale algorithm, improve the global search ability of the whale, and avoid falling into the local optimum; then, shrinking and encircling, spiral bubble net attack, prey search and update of the historical optimal whale individual are repeated in sequence;

[0013] Step S3: Execute the security computing energy consumption optimization configuration according to the global optimal solution.

[0014] Further preferably, the ultra-dense multi-access mobile edge computing network in step S1 is an ultra-dense MAMEC network; basic network information of the ultra-dense MAMEC network is obtained, and a communication model, a computing model and a security model in the network system are further constructed through the basic network information, and an optimization problem is constructed based on the communication model, the computing model and the security model.

[0015] Further preferably, a communication model is constructed: the system frequency band is divided into two parts, which are used by the macro base station and the micro base station respectively, and the widths are and ; is the width of the system frequency band; is the frequency band division factor, with a value between 0 and 1. In the uplink NOMA (non-orthogonal multiple access) mode, any mobile device served by some micro base stations will receive co-channel interference from other mobile devices with worse channel gain than it. service, the received signals are decoded in descending order of channel gain; considering that mobile devices often transmit tasks one by one on a certain channel, assuming that each macro base station has sub-channels, which correspond to only one channel in reality, and any mobile device uses one of the sub-channels to transmit a task in a certain time slot.

[0016] Further optimization, build cache and offload model: any mobile device In each time slot, there is a time-sensitive task denoted as; each mobile device may be associated with a base station. With micro base station If the mobile device is associated with the micro base station, the mobile device and the micro base station association index , otherwise 0; if the mobile device If associated with a macro base station, the mobile device is associated with the macro base station indicator. , otherwise 0; when the mobile device Have a mission When processing is required, if it is not associated with any base station, it will directly perform local calculations; if there is an associated base station, it will first query whether the associated micro base station has cached the corresponding task. If so, the micro base station task cache indicator , otherwise it is 0; if the micro base station task cache indicator , directly perform tasks on the associated micro base station and feed back the results to the mobile device; if the associated micro base station is not cached, continue to query whether the macro base station has cached the corresponding task. If it is cached, the macro base station task cache indicator , otherwise it is 0; if the macro base station task cache indicator , then the task is performed on the macro base station and the result is fed back to the mobile device; if neither is cached, that is , then the task is offloaded step by step. If it is associated with the macro base station and the task is not cached, that is , then upload the task to the macro base station to execute the task.

[0017] Further optimization is to construct a security model: in order to ensure safe unloading, different encryption algorithms are used to encrypt and decrypt the unloaded tasks; when a mobile device is associated with a micro base station, a part of any task of the mobile device is unloaded to this base station after encryption; the base station first decrypts it, and then transmits part of it to the macro base station after encryption; the remaining part is handed over to the associated micro base station for execution; when a mobile device is associated with a micro base station, a part of any task of the mobile device is unloaded to the macro base station after encryption, and the task is executed by the macro base station after decryption.

[0018] Further preferably, the optimization problem is:

[0019] ;

[0020] Indicates the association index between the task and the base station, represents the security protection decision variable of the task, Indicates the subchannel selected by the mobile device, Indicates the index set of the amount of data offloaded from the mobile device to the base station, Represents a data volume index set that is offloaded from a micro base station to a macro base station; Choose a value small enough to avoid division by zero; For 10 constraints, Represents a mobile device Task execution time Cannot exceed its deadline ; Represents a mobile device Total damage cost cannot exceed its maximum acceptable cost ; and This means that a mobile device can only be associated with one base station; and Represents a mobile device Mission Only one encryption algorithm can be selected; and Indicates that a mobile device can only select one subchannel; Represents a mobile device Transmit power The lower bound and upper bound ; Indicates uninstall part and Must be greater than or equal to 0, but less than or equal to the mobile device Mission Data size ,at the same time Must be less than or equal to ; Indicates the association index between mobile devices and micro base stations. For mobile devices Medium Task Safety decision indicators, Represents a mobile device In subchannel The associated decision on is the set of subchannels, The amount of data offloaded from the micro base station to the macro base station, Offloading to micro base stations for mobile devices Mission The amount of data.

[0021] Further preferably, step S2 comprises:

[0022] Step S2.1: Initialize the maximum number of iterations of the improved whale algorithm GWOA , and the current number of iterations Set to 1;

[0023] Step S2.2: Optimize the parameters of the problem Encoded as ,in and It is a whale individual The base station index selected in; Represents a collection of virtual mobile devices, Is a mobile device In individual whales The cryptographic algorithm index selected in; Is a mobile device In individual whales The channel index selected in ; For mobile devices In individual whales The transmission power in From a mobile device Offloading to individual whales Data volume of micro-cells; From a mobile device or its associated small base station to the whale individual The amount of data from nearby macro base stations; , Is A population of whales.

[0024] Step S2.3: Evaluate the fitness of individual whales;

[0025] Step S2.4: Population initialization; to satisfy the constraints , the initial population is generated using the following rules; specifically, any individual whale All initialized to:

[0026] ;

[0027] in, They are the base station index selected by the initial whale individual and the initial mobile device. In individual whales The cryptographic algorithm index selected in the initialized mobile device In individual whales The channel index selected in the initialized mobile device In individual whales The transmission power in the initialized mobile device Offloading to individual whales Data volume of micro-base stations, initialized from mobile devices or its associated small base station to the whale individual The amount of data from nearby macro base stations; The returned I*K matrix corresponds to the linear index The row subscript and column subscripts ; From the collection Output a random element in Generate a and A random number between ; the fitness function is used to calculate the fitness values ​​of all individuals in the whale population, and the whale individual with the highest fitness value is taken as the historical optimal whale individual.

[0028] Step S2.5: Determine the current number of iterations Is it less than or equal to the maximum number of iterations? If the current number of iterations Less than or equal to the maximum number of iterations ; Then the whale population is searched for prey, surrounded by shrinkage and attacked by spiral bubble net, and the whale population is updated to obtain the target population; if the current number of iterations Greater than the maximum number of iterations , then output the codes of all whale individuals in the target population.

[0029] Further preferably, step S2.5 comprises:

[0030] Generate two random numbers between 0 and 1 and ;like , then execute step S2.50, and then execute step S2.51; if , then directly execute step S2.51;

[0031] Step S2.50: Generate a random number between 0 and 1 and compare this number with the mutation probability and the crossover probability Compare them respectively; if the mutation probability is met, the whale population is mutated; if the crossover probability is met, a partial matching crossover operation is performed on the whale population; in combination with practical problems, the mutation operation randomly exchanges elements in the gene sequence of the parent individuals; the partial matching crossover operation first randomly selects two parent individuals, recorded as parent 1 and parent 2; then randomly selects two or more crossover points in the gene sequence of the parent individuals; between the selected crossover points, the gene sequence of parent 1 is partially matched with the gene sequence of parent 2; starting from the gene sequence of parent 1, copy it to the first crossover point; further, starting from the first crossover point, try to match the gene sequence of parent 2 with the sequence of parent 1. If the match is successful, the genes of parent 2 are copied to the offspring; if not, the genes of parent 1 are copied to the offspring; continue this process until all crossover points have been checked; after the last crossover point, copy the remaining gene sequence of parent 1 to the offspring;

[0032] Step S2.51: In the whale algorithm, all whales are looking for prey, and the best historical agent is considered to be the target solution; the current best agent refers to the whale individual with the highest fitness value among all individuals in the current iteration, and the best historical agent is the whale individual with the highest fitness value among all individuals in the previous and current iterations; after executing the whale algorithm, the best historical agent is the best solution to the optimization problem; the humpback whale identifies the location of the prey and then completely surrounds the prey. In the behavior of surrounding the prey, all whales move towards the best historical agent during the iteration;

[0033] Step S2.52: The bubble net attack of the humpback whale includes shrinking and spiraling, and the probability of the two operations is equal; by performing these operations, the new position of any agent will be between its current position and the position of the best historical agent, and a local optimal solution to the optimization problem is obtained;

[0034] Step S2.53: When searching for prey in the traditional WOA, the whale is forced to move towards a random whale, and a mutation operation is performed using Cauchy's inverse cumulative distribution function to formulate the whale's prey search;

[0035] Step S2.54: Forcing the individual whale to search for prey again in the vicinity of the best agent in history;

[0036] Step S2.55: Set the current number of iterations The value of is increased by 1.

[0037] Compared with the existing technology, the present invention has the following beneficial effects:

[0038] (1) The present invention uses an improved whale algorithm (GWOA) to search; based on WOA, an adaptive nonlinear weight factor is used to update the position of individual whales in the shrinking and encircling stage, and is also used in the spiral bubble net attack and prey search stages; in the spiral bubble net attack stage, an inertia weight coefficient is used to adaptively adjust the spiral amplitude to avoid falling into a local optimum; in the prey search stage, the inverse cumulative distribution function of Cauchy is used to perform a long-tail mutation operation on the whale to expand the search space of the improved whale algorithm, forming IWOA. On the basis of IWOA, the mutation and crossover operations in the genetic algorithm are introduced to form GOWA, which increases species diversity and solves problems more accurately.

[0039] (2) The present invention constructs an optimization problem based on the network basic information of the ultra-dense multi-access mobile edge computing network, and constructs the optimization problem under the constraints of the network system; obtains an initial solution based on the optimization problem, and defines the initial solution as the parent population, uses the improved whale algorithm to search the parent population to obtain the target population, and outputs the position of the global optimal whale in the target population; performs a secure computing efficiency optimization configuration based on the global optimal whale position. The present invention considers the influence of base station cache factors to jointly optimize mobile device association, channel selection, security decision-making and power allocation to minimize the local total energy consumption of mobile devices.

[0040] (3) Under the constraints of computing resources and delay, security cost, task offloading ratio, mobile device transmission power ratio, minimum rate and maximum energy consumption, the method can well minimize the total local energy consumption of mobile devices. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is a flow chart of the method of the present invention.

[0042] Figure 2 The present invention discloses a schematic diagram of the effect of the number of mobile devices in the network on the total local energy consumption of the mobile devices.

[0043] Figure 3 The present invention discloses a schematic diagram of the effect of the number of network mobile devices on the value of the fitness function.

[0044] Figure 4 The present invention discloses the convergence diagram of the iterative process of the GWOA fitness function value of the method of the present invention. DETAILED DESCRIPTION

[0045] The present invention will be further described in detail below with reference to the accompanying drawings.

[0046] like Figure 1 As shown, a secure computing offloading method for cache-assisted ultra-dense heterogeneous MEC networks includes the following steps:

[0047] Step S1: obtaining basic network information of an ultra-dense multi-access mobile edge computing network and constructing a network system, wherein the network system includes a communication model, a computing model, and a security model; constructing an optimization problem under the constraints of the network system; the ultra-dense multi-access mobile edge computing network is an ultra-dense MAMEC network;

[0048] Step S2: Obtain an initial solution according to the optimization problem, and define the initial solution as the parent population. Use the improved whale algorithm to search the parent population to obtain the target population, and output the global optimal solution in the target population.

[0049] The process of using the improved whale algorithm to search the parent population to obtain the target population includes: first, the population is initialized and the historical optimal whale individual is determined; then, the partial matching crossover operation and mutation operation of the genetic algorithm are used to expand the whale population and increase the population diversity; then, in the shrinking and encircling stage, an adaptive nonlinear weight factor is used to update the position of the whale individual, and it is also used in the spiral bubble net attack and prey search stages; in the spiral bubble net attack stage, the inertia weight coefficient is used to adaptively adjust the spiral amplitude to avoid falling into the local optimum; in the prey search stage, the long tail mutation operation of the whale using the inverse cumulative distribution function of Cauchy is used to expand the search space of the improved whale algorithm, improve the global search ability of the whale, and avoid falling into the local optimum; then, shrinking and encircling, spiral bubble net attack, prey search and update of the historical optimal whale individual are repeated in sequence;

[0050] Step S3: Execute the security computing energy consumption optimization configuration according to the global optimal solution.

[0051] The ultra-dense multi-access mobile edge computing network described in step S1 is an ultra-dense MAMEC network; the specific process of step S1 is: obtaining the network basic information of the ultra-dense MAMEC network, further building a communication model, a computing model and a security model in the network system through the network basic information, and constructing an optimization problem based on the communication model, the computing model and the security model.

[0052] Furthermore, the specific process of obtaining the network basic information of the ultra-dense MAMEC network is as follows: The index set of mobile devices is ,in, Represents a mobile device in the network system. Represents the number of mobile devices in the network system; collection Represents the task collection of each mobile device, , where k represents the task of the mobile device and K represents the number of tasks; assuming that there is a macro base station (with is the index) and a micro base station set ,in ,in, Represents a micro base station in the network system, Indicates the number of micro base stations in the network system, Represents the set of all base stations; then, obtain the mobile device Mission ,in is the data volume of the task, The number of CPU cycles required to complete a 1-bit task, The maximum task processing time of the mobile device; obtain the total bandwidth of the network system , subchannel bandwidth ,mobile device Channel gain and noise power between the base station and the , then, get the mobile device Maximum computing resources Energy consumption coefficient of mobile devices , and micro base stations Maximum computing resources .

[0053] Furthermore, the specific process of building the communication model, cache and offload model, security model and calculation model in the network system through the basic network information is as follows:

[0054] Building a communication model: To reduce cross-layer interference, the system frequency band is divided into two parts, one for macro base stations and the other for micro base stations, with widths of and ; is the width of the system frequency band; is the frequency band division factor, which takes a value between 0 and 1. It is assumed that there is no intra-cluster interference for any task of the mobile device. In the uplink NOMA (non-orthogonal multiple access) mode, any mobile device served by some micro base stations will receive co-channel interference from other mobile devices with worse channel gain than it. service, the received signals are decoded in descending order of channel gain. Micro Base Station Associated mobile devices The uplink data rate can be given by:

[0055] ;

[0056] in, Subchannel Micro Base Station Associated mobile devices Uplink data rate, Is interfering with mobile devices and micro base stations In subchannel A group of mobile devices communicating on and For mobile devices Selected channels and microcell indicators; and For mobile devices Selected channels and microcell indicators; For mobile devices The transmission power; is the noise power; Micro base station The cluster to which it belongs; is the subchannel bandwidth, Is a mobile device and micro base stations The channel gain between Is a mobile device and micro base stations When Intra-cluster interference refers to the interference between different micro base stations in the same cluster. From a mobile device on the same channel The interference received. Intra-cluster interference is the interference of mobile devices on the same channel of the same base station (BS) in the same cluster. Received interference.

[0057] Since the mobile devices associated with the macro base station also use a different frequency band from the micro base station, there is no intra-layer and inter-layer interference. Considering that mobile devices often transmit tasks one by one on a certain channel, it is assumed that each macro base station (MBS) has subchannels, which correspond to only one channel in reality. That is, any mobile device can use one of the subchannels to transmit a task in a certain time slot, which means that such a mobile device uses the real channel to complete the task. Based on this, the subchannel Mobile devices associated with macro base stations Uplink data rate It can be given by the following formula:

[0058] ;

[0059] in, is the number of mobile devices associated with the macro base station, It is the indicator of the association between mobile devices and macro base stations. Is a mobile device and the channel gain between the macro base station, is the frequency band division factor, with a value between 0 and 1. is the width of the system frequency band.

[0060] Building a caching and offloading model: Assuming any mobile device In each time slot there is a time-sensitive task denoted as .

[0061] Each mobile device may be associated with a base station. With micro base station If the mobile device is associated with the micro base station, the mobile device and the micro base station association index , otherwise 0; if the mobile device If associated with a macro base station, the mobile device is associated with the macro base station indicator. , otherwise 0. Have a mission When processing is required, if it is not associated with any base station, it will directly perform local calculations; if there is an associated base station, it will first query whether the associated micro base station has cached the corresponding task. If so, the micro base station task cache indicator , otherwise it is 0. If the micro base station task cache indicator , directly perform tasks on the associated micro base station and feed back the results to the mobile device; if the associated micro base station is not cached, continue to query whether the macro base station has cached the corresponding task. If it is cached, the macro base station task cache indicator , otherwise it is 0; if the macro base station task cache indicator , then the task is performed on the macro base station and the result is fed back to the mobile device; if neither is cached, that is , then the task is offloaded step by step. If it is associated with the macro base station and the task is not cached, that is , then upload the task to the macro base station to execute the task.

[0062] Building a security model: To ensure secure offloading, different encryption algorithms are used to encrypt and decrypt offloaded tasks. When a mobile device is associated with a micro base station, a part of any task of the mobile device is offloaded to this base station after encryption. The base station first decrypts it, and then encrypts and transmits part of it to the macro base station. The remaining part is executed by the associated micro base station. When a mobile device is associated with a micro base station, a part of any task of the mobile device is offloaded to the macro base station after encryption, and the macro base station executes the task after decryption.

[0063] Different encryption algorithms correspond to different security levels. Assume that the security protection level set ,in Indicates the encryption algorithm The protection level, L is the number of encryption algorithms. In addition, the computing power for encrypting and decrypting 1 bit is and (CPU cycles / bit);

[0064] When mobile devices Mission Using encryption algorithm The probability of failure when safely offloading part of its tasks for:

[0065] ;

[0066] in, For mobile devices Mission Safety risk factor; For mobile devices Mission The desired level of security. Mission The security level of the encryption algorithm is greater than or equal to the expected security level. To the task The protection is successful. Otherwise, the encryption algorithm The mission cannot be protected with a certain probability.

[0067] mobile device Mission Cost of security breaches It can be given by the following formula:

[0068] ;

[0069] in, is a set of encryption algorithms, For the task financial loss from failure; For mobile devices Medium Task Security decision indicators, if you choose an encryption algorithm To handle mobile devices Tasks in ,but , otherwise 0.

[0070] Build a computational model:

[0071] Local Computing:

[0072] Local computing latency and energy consumption as follows:

[0073] ;

[0074] ;

[0075] in It is the computing power of the mobile device; It is an energy consumption coefficient that depends on the hardware architecture.

[0076] Edge computing:

[0077] When mobile devices Performing its tasks at the micro base station , the task processing time required for:

[0078] ;

[0079] When mobile devices Performing its tasks at the macro base station , the task processing time required for:

[0080] ;

[0081] in, The computing power allocated to mobile devices by micro base stations, The computing power allocated to mobile devices by macro base stations.

[0082] Step-by-step calculation of unloading:

[0083] (1) Local computing: When mobile devices With micro base station Where relevant, mobile devices Mission The amount of data processed locally is ,in Offloading to micro base stations for mobile devices Mission Therefore, processing mobile devices With micro base station Related tasks Local execution time used for:

[0084] ;

[0085] in, For mobile devices The computing power of Offload tasks from mobile devices to macro base stations The amount of data.

[0086] (2) Offloading to micro base stations: When mobile devices With micro base station When associated, any task needs to Follow these steps. First, of Partially encrypted from mobile devices Offloading to micro base stations Secondly, micro base stations Decryption , then execute Third, In After part of the encryption is completed, the micro base station Fourth, the macro base station performs Therefore, processing mobile devices Mission Time for:

[0087] ;

[0088] in, Represents a mobile device In subchannel The association decision on the mobile device is ,but , otherwise 0. is the wired backhaul rate between the micro base station and the macro base station; Micro Base Station Assign to mobile device Mission The computing power of Assigned to mobile devices by macro base stations Mission The computing power of is the amount of data offloaded from the micro base station to the macro base station; on the right side of the equal sign, the first four items are From a mobile device Upload to Micro Base Station The time spent calculating the Next , used from micro base stations Upload to nearby macro base stations and used to calculate the The last three items are used for micro base stations. Decryption time; in micro base stations encryption time; decryption on nearby macro base stations time.

[0089] According to the processing of mobile devices Mission The ratio of CPU cycles used to the total usage cycles of the relevant base station, The computing power of the mobile device is allocated to the computation and security operations of the task. With micro base station Related, micro base station Assign to mobile device Mission The computing power can be expressed as:

[0090] ;

[0091] in, Micro base station The total computing power of It is used to process CPU cycles; Is used to decrypt and encryption CPU cycles, , , Cache indicators for micro base station tasks, representing mobile devices Mission Whether it is cached in the micro base station, Cache indicators for macro base station tasks, representing mobile devices Mission Whether it is cached in the macro base station, For mobile devices Tasks in Safety decision indicators, It is used to process CPU cycles, Is used to decrypt and encryption CPU cycles, For mobile devices Indicators associated with micro base stations, For mobile devices Mission The amount of data, For mobile devices The number of CPU cycles required to complete a 1-bit task.

[0092] Due to the micro base station The associated mobile devices can further offload part of the tasks to the nearby macro base stations for processing, while the mobile devices associated with the macro base stations can directly upload the tasks to these base stations for execution, so the data processed on the macro base stations should include the following two parts. The CPU cycles for data offloaded to the mobile device's selected macro base station are determined by Given. Among them, , For calculating and decrypting micro base stations Offload to mobile device The number of CPU cycles for data on the selected macro base station, The amount of data offloaded from the micro base station to the macro base station. In addition, the CPU cycles used to calculate and decrypt the data offloaded from the mobile device to its selected macro base station are Given, where , For mobile devices The correlation index with the macro base station, For computing and decrypting from mobile devices The number of CPU cycles to offload data to its selected macro base station, Offload tasks from mobile devices to macro base stations The amount of data, under the above proportion calculation distribution, when the mobile device When associated with a macro base station, the macro base station is assigned to the mobile device The computational power is given by:

[0093] ;

[0094] (3) Offloading to macro base stations: When mobile devices When associated with a macro base station, any task The following steps need to be performed. First, In Partially encrypted from mobile devices Offload to the macro base station. Secondly, the macro base station decrypts , and then execute. Therefore, for processing mobile devices associated with macro base stations Mission The time is given by:

[0095] ;

[0096] The right side of the equal sign represents the Upload Time to the macro base station; calculated at the macro base station time; and decryption at the macro base station time.

[0097] It is assumed that all tasks of each mobile device are executed sequentially to meet practical implementation requirements. However, local execution and computation offloading can be performed in parallel for any task. The total time required for all tasks is given by:

[0098] ;

[0099] Then the local total energy consumption of all mobile devices is for:

[0100] ;

[0101] in, is the chip architecture energy coefficient, and the right side of the formula is the total computing energy consumption, encryption total energy consumption and upload total energy consumption of the mobile device.

[0102] Furthermore, the optimization problem is constructed based on the communication model, cache and offload model, security model and calculation model:

[0103] ;

[0104] Indicates the association index between the task and the base station, represents the security protection decision variable of the task, Indicates the subchannel selected by the mobile device, Indicates the index set of the amount of data offloaded from the mobile device to the base station, Represents a data volume index set that is offloaded from a micro base station to a macro base station; Choose a value small enough to avoid division by zero, such as 10 -20 ; For 10 constraints, Represents a mobile device Task execution time Cannot exceed its deadline ; Represents a mobile device The total damage cost cannot exceed its maximum acceptable cost ; and This means that a mobile device can only be associated with one base station; and Represents a mobile device Mission Only one encryption algorithm can be selected; and Indicates that a mobile device can only select one subchannel; Represents a mobile device Lower bound of transmit power and upper bound ; Indicates uninstall part and Must be greater than or equal to 0, but less than or equal to the mobile device Mission Data size ,at the same time Must be less than or equal to , is a set of subchannels.

[0105] Further, step S2 includes:

[0106] Step S2.1: Initialize the maximum number of iterations of the improved whale algorithm GWOA , and the current number of iterations Set to 1;

[0107] Step S2.2: Optimize the parameters of the problem Encoded as ,in and It is a whale individual The base station index selected in; Represents a collection of virtual mobile devices, Is a mobile device In individual whales The cryptographic algorithm index selected in; Is a mobile device In individual whales The channel index selected in ; For mobile devices In individual whales The transmission power in From a mobile device Offloading to individual whales Data volume of micro-cells; From a mobile device or its associated small base station to the whale individual The amount of data from nearby macro base stations. , Is A population of whales.

[0108] Step S2.3: In order to evaluate the fitness of individual whales, a fitness function needs to be designed reasonably. As can be seen from the optimization problem, the constraint and is nonlinear, mixed integer and coupled, which is difficult to satisfy in the whale's actions. In view of this, the constraint is introduced into the fitness function as a penalty term, and the penalty term can be explicitly used to prevent individuals from falling into the infeasible region. In this way, the established population can always find a feasible optimal solution.

[0109] In order to make the constraint and The energy consumed by all mobile devices is minimal, and individual whales can be defined The fitness function is:

[0110] ;

[0111] in Represents individual whales The fitness function value of Is a mobile device The delay constraint penalty factor is Is a mobile device The total cost constraint penalty factor of the security vulnerability; represents the total energy consumption of the task, Represents the total time of the task, Represents the maximum allowable delay of task execution, For mobile devices The total cost of security breaches, For mobile devices The maximum allowable total cost of a security breach.

[0112] Step S2.4: Population initialization; to satisfy the constraints , the initial population can be generated using the following rules. Specifically, any single whale individual Can be initialized as:

[0113] ;

[0114] in, They are the base station index selected by the initial whale individual and the initial mobile device. In individual whales The cryptographic algorithm index selected in the initialized mobile device In individual whales The channel index selected in the initialized mobile device In individual whales The transmission power in the initialized mobile device Offloading to individual whales Data volume of micro-base stations, initialized from mobile devices or its associated small base station to the whale individual The amount of data from nearby macro base stations; The returned I*K matrix corresponds to the linear index The row subscript and column subscripts . From the collection Output a random element in Generate a and A random number between . The fitness function is used to calculate the fitness values ​​of all individuals in the whale population, and the whale individual with the highest fitness value is taken as the historical optimal whale individual.

[0115] Step S2.5: Determine the current number of iterations Is it less than or equal to the maximum number of iterations? If the current number of iterations Less than or equal to the maximum number of iterations . Then the whale population is searched for prey, surrounded by shrinkage and attacked by spiral bubble net, and the whale population is updated to obtain the target population; if the current number of iterations is Greater than the maximum number of iterations , then output the codes of all whale individuals in the target population.

[0116] Further, step S2.5 includes:

[0117] Generate two random numbers between 0 and 1 and ;like , then execute step S2.50, and then execute step S2.51; if , then directly execute step S2.51.

[0118] Step S2.50: Generate a random number between 0 and 1 and compare this number with the mutation probability and the crossover probability Compare them respectively; if the mutation probability is met, the whale population is mutated; if the crossover probability is met, the whale population is partially matched and crossover is performed. Combined with practical problems, the mutation operation randomly exchanges elements in the gene sequence of the parent individuals; the partial matching crossover operation first randomly selects two parent individuals, recorded as parent 1 and parent 2. Then randomly select two or more crossover points in the gene sequence of the parent individuals. Between the selected crossover points, the gene sequence of parent 1 is partially matched with the gene sequence of parent 2. Starting from the gene sequence of parent 1, copy to the first crossover point. Further, starting from the first crossover point, try to match the gene sequence of parent 2 with the sequence of parent 1. If the match is successful, the gene of parent 2 is copied to the offspring; if it does not match, the gene of parent 1 is copied to the offspring. Continue this process until all crossover points have been checked. After the last crossover point, copy the remaining gene sequence of parent 1 to the offspring.

[0119] Step S2.51: In the whale algorithm, all whales are searching for prey, and the best historical agent is considered the target solution. It is worth noting that the current best agent refers to the whale individual with the highest fitness (function) value among all individuals in the current iteration, while the best historical agent is the whale individual with the highest fitness value among all individuals in the previous and current iterations. After executing the whale algorithm, the best historical agent is the best solution to the optimization problem. Humpback whales can identify the location of the prey and then completely surround the prey. In the act of surrounding the prey, all whales move towards the best historical agent during the iteration. For any individual , which surrounds its prey The behavior can be expressed as:

[0120] ;

[0121] in, The base station index selected by the mobile device in the current best whale individual; The cryptographic algorithm index selected by the mobile device among the current best whale individuals; The channel index selected by the mobile device in the current optimal whale individual; is the transmission power of the mobile device in the current optimal whale individual; is the amount of data offloaded from the mobile device to the micro base station in the current optimal whale individual; The amount of data offloaded from the mobile device or its associated micro base station to the macro base station in the current optimal whale individual. The base station index selected by the mobile device among the historical best whale individuals; Index of cryptographic algorithms selected by mobile devices from the historical best whale individuals; The channel index selected by the mobile device among the best whale individuals in history; is the transmission power of the mobile device in the historically optimal whale individual; The amount of data offloaded from mobile devices to micro base stations in the historically optimal whale individual; The amount of data offloaded from a mobile device or its associated micro base station to the macro base station in the historically optimal whale individual.

[0122] ;

[0123] in, and for A random number between is the iteration index; is the total number of iterations; and is an adaptive nonlinear weight used to update the position of individual whales; The value of A random number between .

[0124] Step S2.52: The bubble net attack of the humpback whale includes shrinking and spiraling, and the probability of the two operations is equal. By performing these operations, the new position of any agent will be between its current position and the position of the best historical agent, and a local optimal solution to the optimization problem can be obtained. In order to simulate the spiral movement of the whale, any individual Prey location The spiral equation between is as follows:

[0125] ;

[0126] in, The base station index selected by the mobile device in the current best whale individual; The cryptographic algorithm index selected by the mobile device among the current best whale individuals; The channel index selected by the mobile device in the current optimal whale individual; is the transmission power of the mobile device in the current optimal whale individual; is the amount of data offloaded from the mobile device to the micro base station in the current optimal whale individual; The amount of data offloaded from the mobile device or its associated micro base station to the macro base station in the current optimal whale individual. The base station index selected by the mobile device among the historical best whale individuals; Index of cryptographic algorithms selected by mobile devices from the historical best whale individuals; The channel index selected by the mobile device among the best whale individuals in history; is the transmission power of the mobile device in the historically optimal whale individual; The amount of data offloaded from mobile devices to micro base stations in the historically optimal whale individual; The amount of data offloaded from a mobile device or its associated micro base station to the macro base station in the historically optimal whale individual.

[0127] Used to adaptively adjust the spiral amplitude to avoid falling into the local optimum. Available from:

[0128] ;

[0129] in is a random number between 0 and 1; is the inertia weight coefficient; is the number of iterations, is the maximum number of iterations.

[0130] Step S2.53: When searching for prey in traditional WOA, the whale is forced to move towards the random whale. By doing so, the search space of the algorithm can be expanded. However, its global search capability depends largely on the selection of random whales and is prone to falling into local optimality. To solve this problem, since the whale's tail is very long, the Cauchy's inverse cumulative distribution function can be used for mutation operation to formulate the whale's prey search. Mathematically, the behavior of any individual m in searching for prey can be expressed as:

[0131] ;

[0132] in, The base station index selected by the mobile device in the current best whale individual; The cryptographic algorithm index selected by the mobile device among the current best whale individuals; The channel index selected by the mobile device in the current optimal whale individual; is the transmission power of the mobile device in the current optimal whale individual; is the amount of data offloaded from the mobile device to the micro base station in the current optimal whale individual; The amount of data offloaded from the mobile device or its associated micro base station to the macro base station in the current optimal whale individual. The base station index selected by the mobile device among the historical best whale individuals; Index of cryptographic algorithms selected by mobile devices from the historical best whale individuals; The channel index selected by the mobile device among the best whale individuals in history; is the transmission power of the mobile device in the historically optimal whale individual; The amount of data offloaded from mobile devices to micro base stations in the historically optimal whale individual; The amount of data that mobile devices offload from their micro base stations or their associated micro base stations to the macro base stations in the historically optimal whale individual. Used to adaptively adjust the size of mutations.

[0133] Step S2.54: In order to improve the convergence speed and avoid premature convergence to obtain a better solution, we further force the whale individuals to search for prey again in the vicinity of the historical best agent. Mathematically speaking, in the historical best agent In the vicinity of The new position of can be given by:

[0134] ;

[0135] in, The base station index selected by the mobile device in the current best whale individual; The cryptographic algorithm index selected by the mobile device among the current best whale individuals; The channel index selected by the mobile device in the current optimal whale individual; is the transmission power of the mobile device in the current optimal whale individual; is the amount of data offloaded from the mobile device to the micro base station in the current optimal whale individual; The amount of data offloaded from the mobile device or its associated micro base station to the macro base station in the current optimal whale individual. The base station index selected by the mobile device among the historical best whale individuals; Index of cryptographic algorithms selected by mobile devices from the historical best whale individuals; The channel index selected by the mobile device among the best whale individuals in history; is the transmission power of the mobile device in the historically optimal whale individual; The amount of data offloaded from mobile devices to micro base stations in the historically optimal whale individual; The amount of data offloaded from a mobile device or its associated micro base station to the macro base station in the historically optimal whale individual. is a random number between 0 and 1.

[0136] Step S2.55: Set the current number of iterations The value of is increased by 1.

[0137] Furthermore, the specific process of step S3 is to restore the position of the global optimal whale into the form of the original optimization parameter solution according to the index set of the optimal position of the improved whale algorithm in the iteration in step S2; and according to the obtained solution, perform task offloading, cryptographic algorithm selection, mobile device computing resource allocation, base station resource allocation and mobile device power control.

[0138] The effects of the implementation cases of the present invention can be further illustrated by simulation.

[0139] The simulation conditions are set as follows: base stations and mobile devices are randomly distributed in a macro cell with a radius of 500m; 31 base stations (30 small base stations and 1 macro base station) are considered, and each base station caches up to 3 tasks; 30 mobile devices, each with 3 subtasks; the system bandwidth is 20MHz, the maximum computing power of the base station is 20GHz, and the maximum computing power of the mobile device is 1GHz; 6 cryptographic algorithms, the cryptographic algorithms require [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, and the energy consumption of encrypting and decrypting one bit of data is [2.5296 5.0425 6.837 7.8528 8.7073 26.3643]*1e-7 joules respectively; the task data size of the mobile device is 200~500 KB, the maximum allowed delay of the task is 5~10s, requiring 50~100 CPU cycles; the weight coefficient in the optimization objective function is 0.5; the energy coefficient of the mobile device is 10 -24 ; The base station energy coefficient is 10 -26 ; Mission safety incurs costs due to protection failure, and the mission safety factor is {5,6}.

[0140] Figure 2 The present invention discloses a schematic diagram of the impact of the number of network mobile devices on the local total energy consumption of mobile devices. In a single macrocell network, the number of network mobile devices in a single macrocell increases from 20 to 30. The local total energy consumption of mobile devices of the three algorithms, GWOA, IWOA and WOA, increases with the increase in the number of network mobile devices. However, the local total energy consumption of mobile devices of the GWOA algorithm proposed in the present invention is always better than the fitness function values ​​of the other two algorithms.

[0141] Figure 3 The present invention discloses the effect of the number of network mobile devices on the fitness function value. In the network, when the number of network mobile devices increases from 20 to 30, the fitness function value decreases accordingly. However, the fitness function value obtained by the GWOA algorithm proposed in the present invention is always greater than the fitness function value obtained by the unimproved IWOA algorithm.

[0142] like Figure 4 As shown, the GWOA algorithm of the present invention will converge to a fixed value after the fitness function undergoes a certain number of iterations, so that a solution to the optimization problem can be obtained.

[0143] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A secure computing offloading method for cache-assisted ultra-dense heterogeneous MEC networks, characterized in that: The steps include: Step S1: Obtain network basic information of an ultra-dense multi-access mobile edge computing network and construct a network system, wherein the network system includes a communication model, a computing model, and a security model; construct an optimization problem under the constraints of the network system; Step S2: Obtain an initial solution according to the optimization problem, and define the initial solution as the parent population. Use the improved whale algorithm to search the parent population to obtain the target population, and output the global optimal solution in the target population. The process of using the improved whale algorithm to search the parent population to obtain the target population includes: first, the population is initialized and the historical optimal whale individual is determined; then, the partial matching crossover operation and mutation operation of the genetic algorithm are used to expand the whale population and increase the population diversity; then, in the shrinking and encircling stage, an adaptive nonlinear weight factor is used to update the position of the whale individual, and it is also used in the spiral bubble net attack and prey search stages; in the spiral bubble net attack stage, the inertia weight coefficient is used to adaptively adjust the spiral amplitude to avoid falling into the local optimum; in the prey search stage, the long tail mutation operation of the whale using the inverse cumulative distribution function of Cauchy is used to expand the search space of the improved whale algorithm, improve the global search ability of the whale, and avoid falling into the local optimum; then, shrinking and encircling, spiral bubble net attack, prey search and update of the historical optimal whale individual are repeated in sequence; Step S3: Execute the security computing energy consumption optimization configuration according to the global optimal solution; The ultra-dense multi-access mobile edge computing network described in step S1 is an ultra-dense MAMEC network; obtaining network basic information of the ultra-dense MAMEC network, further building a communication model, a computing model and a security model in the network system through the network basic information, and constructing an optimization problem based on the communication model, the computing model and the security model; The specific process of obtaining the basic network information of the ultra-dense MAMEC network is as follows: The index set of mobile devices is ,in, Represents a mobile device in the network system. Represents the number of mobile devices in the network system; collection Represents the task collection of each mobile device, , where k represents the task of the mobile device and K represents the number of tasks; assuming there is a macro base station and a micro base station set ,in ,in, Represents a micro base station in the network system, Indicates the number of micro base stations in the network system, represents the set of all base stations, 0 is the index of the macro base station; then, obtain the mobile device Mission ,in is the data volume of the task, The number of CPU cycles required to complete a 1-bit task, The maximum task processing time of the mobile device; obtain the total bandwidth of the network system , subchannel bandwidth ,mobile device Channel gain and noise power between the base station and the , then, get the mobile device Maximum computing resources Energy consumption coefficient of mobile devices , and micro base stations Maximum computing resources ; Build Cache and Offload Model: Any Mobile Device In each time slot, there is a time-sensitive task denoted as; if the mobile device With micro base station If the mobile device is associated with the micro base station, the mobile device and the micro base station association index , otherwise 0; if the mobile device If associated with a macro base station, the mobile device is associated with the macro base station indicator. , otherwise 0; when the mobile device Have a mission When processing is required, if it is not associated with any base station, it will directly perform local calculations; if there is an associated base station, it will first query whether the associated micro base station has cached the corresponding task. If so, the micro base station task cache indicator , otherwise it is 0; if the micro base station task cache indicator , directly perform tasks on the associated micro base station and feed back the results to the mobile device; if the associated micro base station is not cached, continue to query whether the macro base station has cached the corresponding task. If it is cached, the macro base station task cache indicator , otherwise it is 0; if the macro base station task cache indicator , then the task is performed on the macro base station and the result is fed back to the mobile device; if neither is cached, that is , then the task is offloaded step by step; if it is associated with the macro base station and the task is not cached, that is, , then upload the task to the macro base station to execute the task.

2. The secure computing offloading method for cache-assisted ultra-dense heterogeneous MEC networks according to claim 1 is characterized in that: Building a communication model: The system frequency band is divided into two parts, which are used by macro base stations and micro base stations respectively, with widths of and ; is the width of the system frequency band; is the frequency band division factor, which takes a value between 0 and 1. In the uplink NOMA mode, any mobile device served by the micro base station will receive co-channel interference from other mobile devices with worse channel gain than it. In service, the received signals are decoded in descending order of channel gain; each macro base station has sub-channels, which correspond to only one channel in reality. Any mobile device uses one of the sub-channels to transmit a task in a certain time slot.

3. The secure computing offloading method for cache-assisted ultra-dense heterogeneous MEC networks according to claim 1 is characterized in that: Building a security model: To ensure secure offloading, different encryption algorithms are used to encrypt and decrypt offloaded tasks; when a mobile device is associated with a micro base station, part of any task of the mobile device is offloaded to this base station after encryption; the base station first decrypts it, then encrypts and transmits part of it to the macro base station; the remaining part is executed by the associated micro base station; When a mobile device is associated with a micro base station, a portion of any task of the mobile device is offloaded to the macro base station after encryption, and the macro base station performs the task after decryption.

4. The secure computing offloading method for cache-assisted ultra-dense heterogeneous MEC networks according to claim 1 is characterized in that: The optimization problem is: ； Indicates the association index between the task and the base station, represents the security protection decision variable of the task, Indicates the subchannel selected by the mobile device, Indicates the index set of the amount of data offloaded from the mobile device to the base station, Represents a data volume index set that is offloaded from a micro base station to a macro base station; Choose a value small enough to avoid division by zero; For 10 constraints, Represents a mobile device Task execution time Cannot exceed its deadline ; Represents a mobile device Total damage cost cannot exceed its maximum acceptable cost ; and This means that a mobile device can only be associated with one base station; and Represents a mobile device Mission Only one encryption algorithm can be selected; and Indicates that a mobile device can only select one subchannel; Represents a mobile device Transmit power The lower bound and upper bound ; Indicates uninstall part and Must be greater than or equal to 0, but less than or equal to the mobile device Mission Data size ,at the same time Must be less than or equal to ; Indicates the association index between mobile devices and micro base stations. For mobile devices Medium Task Safety decision indicators, Represents a mobile device In subchannel The associated decision on is the set of subchannels, The amount of data offloaded from the micro base station to the macro base station, Offloading to micro base stations for mobile devices Mission The amount of data.

5. The secure computing offloading method for cache-assisted ultra-dense heterogeneous MEC networks according to claim 4 is characterized in that: Step S2 includes: Step S2.1: Initialize the maximum number of iterations of the improved whale algorithm GWOA , and the current number of iterations Set to 1; Step S2.2: Optimize the parameters of the problem Encoded as ,in and It is a whale individual The base station index selected in; Represents a collection of virtual mobile devices, Is a mobile device In individual whales The cryptographic algorithm index selected in; Is a mobile device In individual whales The channel index selected in ; For mobile devices In individual whales The transmission power in From a mobile device Offloading to individual whales Data volume of medium and micro base stations; From a mobile device or its associated small base station to the whale individual The amount of data from nearby macro base stations; , Is A population of whales. Step S2.3: Evaluate the fitness of individual whales; Step S2.4: Population initialization; to satisfy the constraints , the initial population is generated using the following rules; specifically, any individual whale All initialized to: ； in, They are the base station index selected by the initial whale individual and the initial mobile device. In individual whales The cryptographic algorithm index selected in the initialized mobile device In individual whales The channel index selected in the initialized mobile device In individual whales The transmission power in the initialized mobile device Offloading to individual whales Data volume of micro-base stations, initialized from mobile devices or its associated small base station to the whale individual The amount of data from nearby macro base stations; The returned I*K matrix corresponds to the linear index The row subscript and column subscripts ; From the collection Output a random element in Generate a and A random number between ; Use the fitness function to calculate the fitness values ​​of all individuals in the whale population, and take the whale individual with the highest fitness value as the historical optimal whale individual; Step S2.5: Determine the current number of iterations Is it less than or equal to the maximum number of iterations? ; If the current number of iterations Less than or equal to the maximum number of iterations ; Then the whale population is searched for prey, surrounded by shrinkage and attacked by spiral bubble net, and the whale population is updated to obtain the target population; if the current number of iterations Greater than the maximum number of iterations , then output the codes of all whale individuals in the target population.

6. The secure computing offloading method for cache-assisted ultra-dense heterogeneous MEC networks according to claim 5 is characterized in that: Step S2.5 includes: Generate two random numbers between 0 and 1 and ;like , then execute step S2.50, and then execute step S2.51; if , then directly execute step S2.51; Step S2.50: Generate a random number between 0 and 1 and compare this number with the mutation probability and the crossover probability Compare them respectively; if the mutation probability is met, the whale population is mutated; if the crossover probability is met, the whale population is partially matched and crossover is performed; in combination with practical problems, the mutation operation randomly exchanges elements of the gene sequence of the parent individuals; the partial matching crossover operation first randomly selects two parent individuals, recorded as parent 1 and parent 2; then randomly selects two or more crossover points in the gene sequence of the parent individuals; between the selected crossover points, the gene sequence of parent 1 is partially matched with the gene sequence of parent 2; starting from the gene sequence of parent 1, copy it to the first crossover point; further, starting from the first crossover point, try to match the gene sequence of parent 2 with the sequence of parent 1; if the match is successful, the gene of parent 2 is copied to the offspring; if it does not match, the gene of parent 1 is copied to the offspring; continue this process until all crossover points have been checked; after the last crossover point, copy the remaining gene sequence of parent 1 to the offspring; Step S2.51: In the whale algorithm, all whales are searching for prey, and the best historical agent is considered to be the target solution; the current best agent refers to the whale individual with the highest fitness value among all individuals in the current iteration, and the best historical agent is the whale individual with the highest fitness value among all individuals in the previous and current iterations; after executing the whale algorithm, the best historical agent is the best solution to the optimization problem; the humpback whale identifies the location of the prey and then completely surrounds the prey; in the behavior of surrounding the prey, all whales move towards the best historical agent during the iteration; Step S2.52: The bubble net attack of the humpback whale includes shrinking and spiraling, and the probability of the two operations is equal; by performing these operations, the new position of any agent will be between its current position and the position of the best historical agent, and a local optimal solution to the optimization problem is obtained; Step S2.53: When searching for prey, the whale is forced to move towards a random whale, using a mutation operation using Cauchy's inverse cumulative distribution function to formulate the whale's prey search; Step S2.54: Forcing the individual whale to search for prey again in the vicinity of the best agent in history; Step S2.55: Set the current number of iterations The value of is increased by 1.

7. The secure computing offloading method for cache-assisted ultra-dense heterogeneous MEC networks according to claim 6 is characterized in that: In step S2.52, in order to simulate the spiral motion of the whale, any individual Prey location The spiral equation between is as follows: ； in, The base station index selected by the mobile device in the current best whale individual; The cryptographic algorithm index selected by the mobile device among the current best whale individuals; The channel index selected by the mobile device in the current optimal whale individual; is the transmission power of the mobile device in the current optimal whale individual; is the amount of data offloaded from the mobile device to the micro base station in the current optimal whale individual; The amount of data offloaded from the mobile device or its associated micro base station to the macro base station in the current optimal whale individual; The base station index selected by the mobile device among the historical best whale individuals; Index of cryptographic algorithms selected by mobile devices from the historical best whale individuals; The channel index selected by the mobile device among the best whale individuals in history; is the transmission power of the mobile device in the historically optimal whale individual; The amount of data offloaded from mobile devices to micro base stations in the historically optimal whale individual; The amount of data offloaded from the mobile device or its associated micro base station to the macro base station in the historically optimal whale individual; Used to adaptively adjust the spiral amplitude to avoid falling into the local optimum; ； in is a random number between 0 and 1; is the inertia weight coefficient.

Images