Energy-efficient collaborative offloading approach for ultra-dense NOMA-MEC networks

Abstract

The present invention discloses an energy-efficient collaborative unloading method for ultra-dense NOMA-MEC networks, which includes: constructing a network architecture based on network basic information, configuring optimization problems under multiple constraints in the network architecture; obtaining a set of initial solutions according to the optimization problem, and defining it as an original population after mapping, processing the original population to obtain a target population, and outputting the encoding information of all red-billed blue magpie individuals; using this information to initialize eel-grouper predator proxy individuals, using the eel-grouper algorithm to search, obtaining the best predator proxy individual information of the population after iterative updating, and performing task unloading resource configuration according to this position information. The present invention has the ability to perform joint computing task unloading configuration, and can achieve the goal of minimizing task unloading energy consumption while meeting security vulnerability costs, transmission power ratios, and latency constraints.

Description

Technical Field

[0001] The present invention relates to the field of wireless communication technologies, and in particular to an energy-efficient collaborative offloading method for an ultra-dense NOMA-MEC network. Background Art

[0002] With the rapid development of communication technology and the popularization of the Internet, new computing-intensive tasks such as virtual reality and autonomous driving are constantly emerging, but at the same time they are accompanied by serious energy consumption problems. Traditional terminal devices are limited by their own basic configuration and computing efficiency and cannot fully meet the task operation requirements. Mobile edge computing (MEC) is considered to be an effective solution strategy. It configures network computing resources at the edge of the communication network and offloads computing tasks to the edge server when the terminal device requests it, effectively meeting the task requirements.

[0003] In order to further provide network communication services close to users and provide better communication quality, we choose to build a mobile edge computing scenario under a mobile ultra-dense network. In this scenario, multiple small base stations are ultra-densely deployed around the macro base station to improve the communication coverage. Each user is associated with multiple base stations when requesting communication, ensuring continuous and efficient use of the computing resources of the ultra-dense base stations. However, this is also accompanied by some serious problems. For example, the ultra-dense deployment of small base stations will lead to increased energy consumption, and the association of users with multiple base stations will lead to complex signal interference. How to give users a better experience with limited spectrum resources is an urgent problem to be solved.

[0004] In order to solve the above problems, it is necessary to study how to calculate resource consumption and reduce interference under realistic conditions to improve user experience and reduce energy consumption. In addition, the traditional red-billed blue magpie algorithm has problems such as fast convergence speed and easy to fall into local optimality. It has certain limitations when applied to actual scenarios and needs to be improved. Summary of the invention

[0005] In order to make up for the shortcomings of the prior art, the present invention provides an energy-efficient collaborative offloading method for ultra-dense NOMA-MEC networks, so as to solve the problem of insufficient global exploration capability of the algorithm due to the new shortcomings in the research on computation offloading in the prior art.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: an energy-efficient collaborative unloading method for ultra-dense NOMA-MEC networks, comprising the following steps:

[0007] Step S1: Obtain network basic information of the ultra-dense NOMA-MEC network to build a network architecture, wherein the network architecture includes a communication model, a computing model, and a security model; configure the optimization problem under multiple constraints in the network architecture;

[0008] Step S2: Generate an initial solution set according to the optimization problem, define it as the original population after chaotic mapping, find the best initialized individual, use the improved red-billed blue magpie algorithm to search and iteratively update the original population to obtain the target population, and output the position information coding of all red-billed blue magpie individuals in the population;

[0009] Step S3: Encode the position of the initial eel-grouper predator agent using the position information of all red-billed blue magpies in the target population, select the optimal initialization individual, and use the eel-grouper algorithm to perform further search, update the position of the eel-grouper predator agent, select the best individual between the current population best individual and the initialized best individual as the historical best individual, and output the historical best position information of the optimal eel-grouper predator agent in the population;

[0010] Step S4: Energy-efficient task offloading resource allocation is performed based on the historical optimal position information of the optimal eel-grouper predator agent individuals in the population.

[0011] Furthermore, the detailed process of step S2 is as follows: the population is initialized according to the optimization problem. The initialization process includes generating a set of random candidate solutions within the constraints of the optimization problem and treating them as the initial solution set, defining them as the original population after chaotic mapping, and finding the initial best individual of the population. After the initialization, the algorithm iteration process is performed: in the stage of searching for prey, the red-billed blue magpies form different types of teams with different numbers, and hunt in the form of teams, and set the balance coefficient; when catching small prey, they form groups to carry out activities, and the number of red-billed blue magpies in the group is is to randomly select 2 from all individuals 5; they act in groups when catching large prey. The number of red-billed blue magpies in the group It is to randomly select more than 10 individuals from all individuals; after forming a team, use the dynamic change coefficient in the exploration of food and cluster search operation Adjust position; use random numbers when catching prey Assign different attack strategies to different teams, using standard normal distribution data Dynamically adjust the fitness value; in the food storage stage, the individual selects the optimal value based on the comparison of the fitness values ​​before and after, and determines whether to perform the Levy flight operation based on the random number; then, the prey search, prey capture, and food storage operations are repeated in the iterative process to update the information of the optimal individual in the current population, and output the location information code of all individuals in the red-billed blue magpie population.

[0012] In step S3, a team consisting of an eel and a grouper is regarded as a predator agent. The position information of all red-billed blue magpies in the target population is used to encode and initialize the position information of the predator agent in the eel-grouper population. The optimal individual is found based on the fitness. Then the algorithm is iterated. The process is as follows: In the process of grouper tracking prey, the adaptive factor Find the right position and the best fitness to avoid falling into the local optimum; use the coefficients in the signal transmission stage Different actions are distinguished when groupers send signals to eels to share the best position information, so that eels also continuously update their positions; in the cooperative attack phase, groupers and eels capture prey in different ways, and use different weight coefficients to update the positions of predator agents according to the positions of groupers and eels, and find the current optimal individual. After the iteration, the current optimal individual in the population is selected and regarded as the historical best individual with the better individual among the initialized optimal individuals, and the relevant information parameters are output.

[0013] In summary, according to the energy-efficient collaborative offloading method for ultra-dense NOMA-MEC networks described above, the improved red-billed blue magpie eel grouper algorithm RBMEGO is used to search and iteratively update the original population to obtain the target population, and the information coding of all red-billed blue magpie individuals in the population is output. According to this coding, the eel-grouper algorithm is used for further search. The red-billed blue magpie algorithm IRBMO and the eel-grouper algorithm EGO are combined to form the improved red-billed blue magpie eel grouper algorithm RBMEGO, which can effectively solve the problems of algorithm convergence too fast and easy to fall into local optimality. Under the condition of meeting the security vulnerability cost, transmission power ratio and delay constraints, the goal of minimizing the energy consumption of task offloading can be achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more intuitively demonstrate the effect of the solution proposed in the present invention, the following will briefly describe and introduce the drawings obtained by using the examples, and the drawings obtained are partial parameter indicators.

[0015] Figure 1 It is a schematic diagram of the present invention showing the influence of user mobile device density on energy consumption.

[0016] Figure 2 It is a schematic diagram of the present invention showing the influence of the density of user mobile equipment on the fitness value. DETAILED DESCRIPTION

[0017] The present invention is further explained in detail below with reference to the examples.

[0018] An energy-efficient collaborative offloading method for ultra-dense NOMA-MEC networks includes the following steps:

[0019] Step S1: Obtain the basic network information of the ultra-dense NOMA-MEC network to build a network architecture, which includes a communication model, a computing model, and a security model; configure the optimization problem under multiple constraints in the network architecture, that is, perform resource allocation and user association under the condition of meeting the security vulnerability cost, transmission power ratio, and delay constraints, so as to achieve the goal of minimizing the energy consumption of task offloading;

[0020] Step S2: Generate an initial solution set according to the optimization problem, define it as the original population after chaotic mapping, find the best initialized individual, use the improved red-billed blue magpie algorithm (IRBMO) to search and iteratively update the original population to obtain the target population, and output the location information coding of all red-billed blue magpie individuals in the population;

[0021] Step S3: Encode the position of the initial eel-grouper predator agent using the position information of all red-billed blue magpies in the target population, select the optimal initialization individual, and use the eel-grouper algorithm (EGO) for further search to update the position of the eel-grouper predator agent, select the best individual between the current population best individual and the initialized best individual as the historical best individual, and output the historical best position information of the optimal eel-grouper predator agent in the population;

[0022] Step S4: Energy-efficient task offloading resource allocation is performed based on the historical optimal position information of the optimal eel-grouper predator agent individuals in the population.

[0023] The applicable scenario of the present invention is the ultra-dense NOMA-MEC network. The basic network information is obtained to establish a network architecture. The network architecture includes a communication model, a security model, and a computing model. The optimization problem is constructed under the framework of this network architecture. The ultimate goal is to minimize the energy consumption of task offloading under multiple constraints.

[0024] Specifically, step S1.1: In the ultra-dense NOMA-MEC network, each small base station (SBS) is configured Server, all small base stations are connected to the nearby macro base station (MBS), the number of small base stations is not less than the number of user mobile devices (IMDs), and there is a macro base station and small base stations, the index of the macro base station is defined as 0, and the small base station index set is , is the number of small base stations, then the index set of all base stations is ; Each user mobile device The task index set is , is the number of tasks, from arrive The security protection level index set is , Indicates the encryption algorithm The security level is L, and L is the number of encryption algorithms. Assume that each user mobile device has an independent task that needs to be executed within the security vulnerability cost and the specified deadline. According to the different location distribution of small base stations, K-means clustering technology is used to divide them into clusters, each of which has Orthogonal sub-channels (frequency bands) are provided to the small base stations in the cluster for use by user mobile devices associated with these small base stations. The same subchannel can be used in NOMA mode. Subchannels in the set From 1 to to index.

[0025] Step S1.2: Construct a communication model based on basic network information. With small base stations The uplink signal to interference and noise ratio (SINR) when associated is as follows:

[0026] ;

[0027] in, Mobile device for users With small base stations Uplink signal-to-interference-and-noise ratio during association, is the index in the small base station set, Mobile device for users The transmission power, Move the device for another user The transmission power, Is the user's mobile device and small base stations The channel gain between Move the device for another user With another small base station The interference channel gain after correlation, Mobile device for users The collection index of is the interference power, represents the noise power; under the above allocation mechanism, the user's mobile device arrive Uplink transmission rate as follows:

[0028] ;

[0029] in, Is the user's mobile device and small base stations The associated index between When , it indicates that the user's mobile device is associated with the small base station, otherwise, it indicates that they are not associated; Represents small base stations The number of associated user mobile devices, is the frequency band division factor, is the frequency bandwidth; is the number of small base stations, Represents small base stations The bandwidth of any user's mobile device is associated with When associated with a macro base station, the user's mobile device Uplink signal-to-interference-noise ratio to the macro base station As shown below:

[0030] ;

[0031] in Represents the user's mobile device and the channel gain between the macro base station.

[0032] Likewise, users' mobile devices Uplink data rate to macro base station Given by:

[0033] ;

[0034] in, Represents the user's mobile device The index of the decision to associate with the macro base station, represents the number of user mobile devices associated with the macro base station; Represents the bandwidth of any user mobile device associated with the macro base station.

[0035] Step S1.3: Further build a computational model based on the communication model, and the user moves the device Mission Recorded as ,in, Represents the user's mobile device Mission The amount of data; is the number of CPU cycles used to calculate a 1-bit task; Is the user's mobile device The maximum delay of Mobile device for users Mission The expected level of security.

[0036] The computation offloading process can be divided into the following cases:

[0037] (1) Local computing: When users move their devices When associated with a base station (BS), a user moves the mobile device Mission The amount of data processed is ,in For the task From the user's mobile device The amount of data offloaded to the base station, Mobile device for users Mission The amount of data, is the index of the base station. At this time, it is used to process the user mobile device associated with the base station Mission Local execution time It is expressed by the following formula:

[0038] ;

[0039] in Mobile device for users The computing power allocated to the task; To encrypt the calculation time, is a set of encryption algorithms, is the number of CPU cycles used to calculate a 1-bit task, Mobile device for users Mission Encryption Algorithm decision to use, and are the computing time and encryption time of the task respectively. At this time, the energy consumption for processing the remaining data is:

[0040] ;

[0041] in, Processing tasks locally Energy consumption required, is the effective switched capacitance related to the chip architecture.

[0042] (2) Offloading to small base stations: When users move their devices When associated with a small cell, this is a two-step offloading task , then the task needs to be Perform the following steps.

[0043] First, of Some content is encrypted and sent from the user's mobile device Offloading to small base stations , small base station Decryption Post-execution , Tasks for moving equipment to users The amount of data, For small base stations offload the workload to the nearby macro base station. of After partial encryption, from small base station Offload to a nearby macro base station, which then executes the decryption Therefore, for user mobile devices and When associated, process the task Remote time It can be derived from the following formula:

[0044] ;

[0045] in, Represents the user's mobile device In subchannel The associated decision on The time spent decrypting the task, Subchannel Small Base Station Associated user mobile device The mission uplink data rate, is the subchannel set index; if Select subchannel but ,otherwise . for and The wired backhaul rate between It is a small base station Assign to user mobile devices Mission computing power; They are From the user's mobile device The latency of uploading to the small base station is calculated on the small base station. The delay will The latency of offloading from a small base station to a nearby macro base station, and the latency of calculating on the macro base station delay. They are used for decryption on small base stations. The latency is high, and encryption is performed in small base stations. The latency and decryption of the nearby macro base station delay.

[0046] (3) Offloading to the macro base station: This is a one-step offloading task. , used to handle the energy consumption of the unloading process for:

[0047] ;

[0048] in, Mobile device for users The transmission power, To offload tasks to the energy consumption of macro base stations, is the energy consumption executed on the macro base station, To decipher the energy consumption, is the amount of data uploaded to the macro base station, Subchannel The uplink transmission rate from the user mobile device to the macro base station, is the partial energy consumption of the macro base station, Energy consumption for encryption and decryption. Used in user mobile devices When associated with a macro base station, processing tasks Remote time It can be derived from the following formula:

[0049] ;

[0050] in, It is the macro base station assigned to the user's mobile device Mission computing power.

[0051] Complete user mobile device Total time for all tasks It can be given by the following formula:

[0052] ;

[0053] Then, the total local energy consumption of all user mobile devices can be given by:

[0054] ;

[0055] in is the energy coefficient of the chip architecture, Mobile device for users The three items on the right side of the above formula are the total computing energy consumption, encryption total energy consumption and upload total energy consumption of the user's mobile device.

[0056] Step S1.4: additionally construct a security model based on the communication model, and different encryption algorithms correspond to different security levels. arrive Index security level , Indicates the encryption algorithm The time consumed by the encryption and decryption processes is (CPU cycles / bit) and (CPU cycles / bit), and assuming the corresponding energy consumption is also the same, that is, (MJ / bit), , When a user moves a device Mission Using encryption algorithm The cost of vulnerabilities caused by security uninstallation for:

[0057] ;

[0058] in, For the task The security risk level; is the expected security level. If the security level of the task is greater than or equal to the expected security level, the encryption algorithm is considered Successfully protected the mission , otherwise it means that the task is not protected.

[0059] User mobile device Mission Cost of security breaches It can be given by the following formula:

[0060] ;

[0061] in, For the task Financial loss in case of failure, Encryption algorithm Protecting user mobile devices Mission The probability of failure when Is the user's mobile device Mission Encryption Algorithm The usage decision index, if an encryption algorithm is selected To handle user mobile devices Tasks in but ,otherwise Then, the user moves the device Total security breach cost as follows:

[0062] ;

[0063] Step S1.5: Construct an optimization problem in the above system architecture composed of multiple models: All tasks are executed in sequence, and the next task will be executed only after the previous task is completed. Execution and unloading are carried out simultaneously to complete the user mobile device The total energy consumption of all tasks As follows:

[0064] ;

[0065] in represents the total energy consumption, represents local energy consumption, represents the multi-step unloading energy consumption, is the total energy consumption offloaded to the small base station, is the total energy consumption offloaded to the macro base station.

[0066] Furthermore, the expression of the optimization problem generated in step S1 is as follows:

[0067] ;

[0068] Among them, the index set of task offloading association decision is Is the user's mobile device and small base stations The associated index between , indicating that the user's mobile device is associated with the small base station, otherwise it is not associated. The index set of task safety decision is Mobile device for users Mission Encryption Algorithm The usage decision of , the index set of subchannel decision is , Mobile device for users In subchannel The associated decision on is a subchannel set; the index of the uplink transmission power of the user's mobile device is , Mobile device for users The transmission power of the mobile device is the index set of the task data volume unloaded from the user to the base station. , the index of the task data volume offloaded from the small base station to the macro base station is ; Represents the total running time of the task, represents the maximum constraint delay of the task, Encryption protection cost for tasks, total user mobile devices, with a maximum of The user's mobile device index set is , each user has There are at most tasks, the task index set is ,have There are encryption algorithms with different robustness, arranged from small to large, from 1 to Indicates the level, the encryption algorithm index is ; Minimum uplink transmit power Arbitrarily choose a value small enough to prevent division by zero, such as , Select a base station to associate with the user's mobile device. is a collection of base stations; Indicates that the task execution time of the user's mobile device cannot exceed its maximum constraint delay ; Indicates that the total encryption protection cost of the user's mobile device cannot exceed its maximum constraint cost ; Represents the association decision when a user's mobile device associates with a base station , otherwise 0, This indicates that a user mobile device can only be associated with one base station at a time; Indicates the selection of encryption algorithm Time Decision , otherwise 0, Represents the tasks of the user's mobile device Only one encryption algorithm can be selected; Indicates that a user's mobile device can only select one sub-channel. Indicates that when the subchannel When selected , otherwise 0; The lower limit of the transmission power of the user's mobile device is given and upper boundary value ; Indicates the amount of uninstalled data and All are greater than or equal to 0, but less than or equal to the tasks performed by the user's mobile device Data size ,in addition Must be less than or equal to .

[0069] From the above information, it can be seen that the optimization problem is formed by combining the communication model, security model, and transmission model in the network architecture. Under multiple constraints such as task delay, transmission power, and security cost, it comprehensively optimizes multiple parameter indicators such as energy consumption, time support ratio, and delay to achieve collaborative task offloading with lower energy consumption.

[0070] The detailed process of step S2 is as follows: the population is initialized according to the optimization problem. The initialization process includes generating a set of random candidate solutions within the constraints of the optimization problem and treating them as the initial solution set. After the chaotic mapping, it is defined as the original population, and the initial best individual of the population is found. After the initialization, the algorithm iteration process is performed: in the stage of searching for prey, the red-billed blue magpies form different types of teams with different numbers, and hunt in the form of teams. The balance coefficient is set to 0.5. When catching small prey, small groups are formed to carry out activities. The number of red-billed blue magpies in the group is is to randomly select 2 from all individuals 5; they act in groups when catching large prey. The number of red-billed blue magpies in the group Usually, more than 10 individuals are randomly selected from all individuals. After forming a team, the dynamic change coefficient is used in the exploration of food and cluster search operations. Adjust position; use random numbers when catching prey Assign different attack strategies to different teams, using standard normal distribution data Dynamically adjust the fitness value to avoid falling into the local optimum; in the food storage stage, the comparison of the fitness values ​​before and after is conducive to the individual selection of the optimal value, and at the same time, determine whether to perform the Levy flight operation based on the random number. This operation can improve the diversity of the population and avoid falling into the local optimum; then, the search for prey, capture of prey, and food storage operations are repeated in the iterative process to update the information of the current population's optimal individual, and output the location information code of all individuals in the red-billed blue magpie population.

[0071] Further, the operation steps of step S2 are as follows:

[0072] Step S2.1: Initialize the maximum number of iterations of the improved red-billed blue magpie algorithm IRBMO , and the initial number of iterations Set to 1, Set to 5000;

[0073] Step S2.2: Assume that the total number of red-billed blue magpies in the population is , where the original population of the improved red-billed blue magpie algorithm RBMO Any individual of the red-billed blue magpie , the objective function solved by each red-billed blue magpie individual after encoding is the fitness value, and the objective function is used to calculate the fitness values ​​of all red-billed blue magpie individuals in the population and find the best individual. Replace with , where the base station index set , Represents the red-billed blue magpie individual and the user's mobile device Base station index when associated, encryption algorithm level set , Represents the red-billed blue magpie individual and the user's mobile device The associated encryption algorithm level, signal index collection , Represents the user's mobile device in the Red-billed Blue Magpie individual Selected signal index, uplink transmit power set , Represents the user's mobile device in the Red-billed Blue Magpie individual Uplink transmission power, data volume set , Represents the user's mobile device in the Red-billed Blue Magpie individual The amount of data and task volume offloaded to the associated small base station , Represents the user's mobile device in the Red-billed Blue Magpie individual Or the task volume offloaded from the small base station associated with it to the macro base station, the red-billed blue magpie individual in the red-billed blue magpie population is generated according to the following formula The fitness function of is:

[0074] ;

[0075] in is the fitness function value, is the task delay penalty factor, is the encryption protection cost penalty factor.

[0076] The fitness function mentioned above is used to calculate the fitness values ​​of all red-billed blue magpie individuals in the original population, and the individual with the largest fitness value is taken as the best red-billed blue magpie individual in history.

[0077] Step S2.3: Determine the current number of iterations Whether the maximum number of iterations has been exceeded , if the current iteration number Less than or equal to the maximum number of iterations , then search for prey, capture prey, and store food for the red-billed blue magpie population; if the current number of iterations is Greater than the maximum number of iterations Then directly output the information coding of all red-billed blue magpies individuals.

[0078] Step S2.3.1: The position of each individual red-billed blue magpie is considered as a valid solution. After multiple operations, the positions of several individuals in the solution space are continuously updated until the historical optimal solution of the population is output. .

[0079] The red-billed blue magpie is a predator with diverse hunting behaviors. It shows social and cooperative tendencies when hunting. In the RBMO algorithm, it continuously updates its position during the iteration process to better hunt. By selecting the best individual at the initialization and the current best individual, the best individual in history can be selected faster, and the global convergence of the algorithm can be appropriately accelerated. The best individual in history refers to the red-billed blue magpie individual with the largest fitness value in all iterations. Mathematically, the red-billed blue magpie individual Searching for prey in small groups The process of position update can be expressed as:

[0080] ;

[0081] in Respectively represent The transmission power selected by the user's mobile device during the round iteration, the amount of task data unloaded by the user to the small base station, the amount of task data unloaded by the user to the small base station, the base station index selected by the user's mobile device, the encryption algorithm index selected by the user's mobile device, and the channel index selected by the user's mobile device; similarly Respectively Iterate the transmission power selected by the user's mobile device, the amount of task data unloaded by the user to the small base station, the amount of task data unloaded by the user to the small base station, the base station index selected by the user's mobile device, the encryption algorithm index selected by the user's mobile device, and the channel index selected by the user's mobile device; They are respectively the transmission power selected by the user's mobile device, the amount of task data offloaded by the user to the small base station, the amount of task data offloaded by the user to the small base station, the base station index selected by the user's mobile device, the encryption algorithm index selected by the user's mobile device, and the channel index selected by the user's mobile device; is the iteration coefficient. This parameter can be used to continuously optimize the individual position information of the red-billed blue magpie according to the number of iterations. They respectively represent the transmission power selected by the user mobile device randomly selected in the population, the amount of task data unloaded by the user to the small base station, the amount of task data unloaded by the user to the small base station, the base station index selected by the user mobile device, the encryption algorithm index selected by the user mobile device, and the location information of the channel index selected by the user mobile device, Express Round down, Representation output A random number between Indicates the location of the prey. Indicates the number of individuals in the red-billed blue magpie group, and its value is between Integer between ;

[0082] When the red-billed blue magpie The process of updating position parameters when forming a cluster to search for prey can be expressed as:

[0083] ;

[0084] It indicates the number of individuals of red-billed blue magpies that form small groups to hunt, and the value is between between.

[0085] Step S2.3.2: The red-billed blue magpie population forms different teams and uses strategies such as fast pecking and jumping to catch prey. When the teams are formed into small groups, the red-billed blue magpie individuals When capturing prey, the position parameters are updated as follows:

[0086] ;

[0087] in The initialization optimal individual position information corresponding to the transmission power selected by the user's mobile device, the amount of task data unloaded by the user to the small base station, the base station index selected by the user's mobile device, the encryption algorithm index selected by the user's mobile device, and the channel index selected by the user's mobile device are respectively:

[0088] ;

[0089] Dynamic Factor and angle variables It is used to dynamically adjust the position of the red-billed blue magpie to prevent it from falling into the local optimum. The position parameters are updated as follows:

[0090] ;

[0091] Step S2.3.3: In addition to finding and catching prey, the red-billed blue magpie also stores excess food. This operation is to compare the individual fitness values ​​before and after the updated position. The position with a large fitness is regarded as the latest position parameters of the red-billed blue magpie individual.

[0092] ;

[0093] in The update formula is as shown above. The same update method is also performed.

[0094] Step S2.3.4: Calculate and compare the fitness of the red-billed blue magpie after the position change, regard the individual with the largest fitness as the current best red-billed blue magpie individual, and compare its fitness with the best individual initialized in the population. If the current individual fitness value is better than the fitness of the historical best individual, replace the initialized best red-billed blue magpie individual with the current individual as the historical best individual.

[0095] The detailed process of step S3 is as follows: a team consisting of an eel and a grouper is regarded as a predator agent, the position information of all red-billed blue magpies in the target population is used to encode and initialize the position information of the predator agent in the eel-grouper population, and the optimal individual is found according to the fitness; then the algorithm is iterated, and the process is as follows: in the process of grouper tracking prey, the adaptive factor Find the right position and the best fitness to avoid falling into the local optimum; use the coefficients in the signal transmission stage Different actions are distinguished when groupers send signals to eels to share the best position information, so that eels also continuously update their positions; in the cooperative attack phase, groupers and eels capture prey in different ways, and use different weight coefficients to update the positions of predator agents according to the positions of groupers and eels, and find the current optimal individual. After the iteration, the current optimal individual in the population is selected and regarded as the historical best individual with the better individual among the initialized optimal individuals, and the relevant information parameters are output.

[0096] The specific process of step S3 is as follows:

[0097] Step S3.1: Inherit the relevant position information of all red-billed blue magpie individuals in the original population to the predator proxy individuals in the eel-grouper population, and calculate the fitness value of the proxy individuals at the same time.

[0098] Step S3.2: The predator agent with the best fitness value in the current position is regarded as the best individual in history.

[0099] Step S3.3: Determine the current number of iterations Whether the maximum number of iterations has been exceeded , if the current iteration number Less than or equal to the maximum number of iterations , then the eel-grouper population will implement internal signal transmission and cooperative attack operations; if the current iteration number Greater than the maximum number of iterations The information encoding and fitness value of the best eel-grouper individual in history are directly output.

[0100] Step S3.4: The specific operation of signal transmission is as follows:

[0101] ;

[0102] in , , Express Take the absolute value.

[0103] Step S3.5: The specific operation of the cooperative attack is as follows:

[0104] ;

[0105] Where b is the iterative random coefficient, Represents a natural constant of Second power, is the ratio of pi.

[0106] Step S3.6: Set the number of iterations .

[0107] Step S3.7: Perform energy-efficient task offloading related configuration according to the desired optimal individual related parameters.

[0108] The effects of the implementation cases of the present invention can be further illustrated by simulation.

[0109] The established network model is: base stations and user mobile devices are randomly distributed in a cellular cell with a diameter of one kilometer; on this basis, a single macro base station and 15 small base stations are considered, and each user has three subtasks; the system bandwidth is 20MHz, the maximum user transmission power is 25dBm, the maximum computing power of the base station is 2GHz, and the maximum computing power of the user's mobile device is 2GHz.

[0110] Figure 1 The figure is a schematic diagram of the influence of user mobile device density on total energy consumption disclosed by the present invention. When the network user mobile device density in a single macro cell changes from 20 to 30, the total energy consumption of the three algorithms, IRBMO, EGO and RBMEGO, increases with the increase in the number of user mobile devices, but the total energy consumption of the RBMEG0 algorithm proposed by the present invention is always lower than that of the other two algorithms.

[0111] Figure 2 The present invention discloses the effect of user mobile device density on fitness value. When the network user mobile device density in a single macro cell changes from 20 to 30, the fitness of the three algorithms IRBMO, EGO and RBMEGO decreases as the number of network user density increases, but the fitness value of the RBMEG0 algorithm proposed in the present invention is always better than the other two algorithms.

[0112] 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. An energy-efficient collaborative offloading method for ultra-dense NOMA-MEC networks, characterized in that: Here are the steps: Step S1: Obtain network basic information of the ultra-dense NOMA-MEC network to build a network architecture, wherein the network architecture includes a communication model, a computing model, and a security model; configure the optimization problem under multiple constraints in the network architecture; Step S2: Initialize the population according to the optimization problem. The initialization process includes generating a set of random candidate solutions within the constraints of the optimization problem and treating them as the initial solution set, defining them as the original population after chaotic mapping, and finding the initial best individual of the population. After the initialization, the improved red-billed blue magpie algorithm iteration process is performed: in the stage of searching for prey, the red-billed blue magpie forms different types of teams with different numbers, and hunts in the form of teams, and sets the balance coefficient; When catching small prey, they form groups to carry out activities. The number of red-billed blue magpies in the group is is to randomly select 2 from all individuals 5; they act in groups when catching large prey. The number of red-billed blue magpies in the group It is to randomly select more than 10 individuals from all individuals; after forming a team, use the dynamic change coefficient in the exploration of food and cluster search operation Adjust position; use random numbers when catching prey Assign different attack strategies to different teams, using standard normal distribution data Dynamically adjust the fitness value; in the food storage stage, the individual selects the optimal value based on the comparison of the fitness values ​​before and after, and determines whether to perform the Levy flight operation based on the random number; then, in the iterative process, the prey search, prey capture, and food storage operations are repeated continuously to update the information of the optimal individual in the current population, and output the location information code of all individuals in the red-billed blue magpie population; Step S3: Encode the position of the initial eel-grouper predator agent using the position information of all red-billed blue magpies in the target population, select the optimal initialization individual, and use the eel-grouper algorithm to perform further search, update the position of the eel-grouper predator agent, select the best individual between the current population best individual and the initialized best individual as the historical best individual, and output the historical best position information of the optimal eel-grouper predator agent in the population; Step S4: performing energy-efficient task offloading resource allocation according to the historical optimal position information of the optimal eel-grouper predator agent individual in the population; In step S3, a team consisting of an eel and a grouper is regarded as a predator agent. The position information of all red-billed blue magpies in the target population is used to encode and initialize the position information of the predator agent in the eel-grouper population. The optimal individual is found based on the fitness. Then the algorithm is iterated. The process is as follows: In the process of grouper tracking prey, the adaptive factor Find the right position and the best fitness to avoid falling into the local optimum; use the coefficients in the signal transmission stage Different actions are distinguished when groupers send signals to eels to share the best position information, so that eels also continuously update their positions; in the cooperative attack phase, groupers and eels capture prey in different ways, and use different weight coefficients to update the positions of predator agents according to the positions of groupers and eels, and find the current optimal individual. After the iteration, the current optimal individual in the population is selected and regarded as the historical best individual with the better individual among the initialized optimal individuals, and the relevant information parameters are output.

2. The energy-efficient collaborative offloading method for ultra-dense NOMA-MEC networks according to claim 1 is characterized in that: In step S1, obtaining basic network information includes: In the ultra-dense NOMA-MEC network, each small base station is configured Server, all small base stations are connected to nearby macro base stations. The number of small base stations is not less than the number of user mobile devices. In the ultra-dense NOMA-MEC network, there is a macro base station and small base stations, the index of the macro base station is defined as 0, and the small base station index set is , is the number of small base stations, then the index set of all base stations is ; Each user mobile device The task index set is , is the number of tasks, from arrive The security protection level index set is , Indicates the encryption algorithm security level, L is the number of encryption algorithms; each user mobile device has an independent task that needs to be executed within the security vulnerability cost and the specified deadline; according to the different location distribution of small base stations, K-means clustering technology is used to divide them into clusters, each of which has Orthogonal sub-channels are provided to the small base stations in the cluster and to the user mobile devices associated with the small base stations. By using the same sub-channel in NOMA mode, Subchannels in the set From 1 to to index.

3. The energy-efficient collaborative offloading method for ultra-dense NOMA-MEC networks according to claim 2 is characterized in that: The expression of the optimization problem generated in step S1 is as follows: ； Among them, the index set of task offloading association decision is Is the user's mobile device and small base stations The associated index between them, the index set of task safety decision is Move devices to your devices Mission Encryption Algorithm The usage decision of , the index set of subchannel decision is , Mobile device for users In subchannel The associated decision on is a subchannel set; the index of the uplink transmission power of the user's mobile device is , Mobile device for users The transmission power of the mobile device is the index set of the task data volume unloaded from the user to the base station. , the index of the task data volume offloaded from the small base station to the macro base station is ; Represents the total running time of the task, represents the maximum constraint delay of the task, Encryption protection cost for tasks, total user mobile devices, with a maximum of The user's mobile device index set is , each user has There are at most tasks, the task index set is ,have There are encryption algorithms with different robustness, arranged from small to large, from 1 to Indicates the level, the encryption algorithm index is ; Minimum uplink transmit power Arbitrarily choose a value small enough to prevent division by zero, Select a base station to associate with the user's mobile device. is a collection of base stations; Indicates that the task execution time of the user's mobile device cannot exceed its maximum constraint delay ; Indicates that the total encryption protection cost of the user's mobile device cannot exceed its maximum constraint cost ; Represents the association decision when a user's mobile device associates with a base station , otherwise 0, This indicates that a user mobile device can only be associated with one base station at a time; Indicates the selection of encryption algorithm Time Decision , otherwise 0, Represents the tasks of the user's mobile device Only one encryption algorithm can be selected; Indicates that a user's mobile device can only select one sub-channel. Indicates that when the subchannel When selected , otherwise 0; The lower limit of the transmission power of the user's mobile device is given and upper boundary value ; Indicates the amount of data uninstalled and All are greater than or equal to 0, but less than or equal to the tasks performed by the user's mobile device Data size ,in addition Must be less than or equal to , For small base stations The amount of tasks offloaded to nearby macro base stations.

4. The energy-efficient collaborative offloading method for ultra-dense NOMA-MEC networks according to claim 3 is characterized in that: Step S2 operation steps are as follows: Step S2.1: Initialize the maximum number of iterations of the improved red-billed blue magpie algorithm IRBMO , and the initial number of iterations Set to 1; Step S2.2: Assume that the total number of red-billed blue magpies in the population is , where the original population of the improved red-billed blue magpie algorithm RBMO is Any individual of the red-billed blue magpie , the objective function solved by each red-billed blue magpie individual after encoding is the fitness value, and the objective function is used to calculate the fitness values ​​of all red-billed blue magpie individuals in the population and find the best individual; the optimization parameters Replace with , where the base station index set , Represents the red-billed blue magpie individual and the user's mobile device Base station index when associated, encryption algorithm level set , Represents the red-billed blue magpie individual and the user's mobile device The associated encryption algorithm level, signal index collection , Represents the user's mobile device in the Red-billed Blue Magpie individual Selected signal index, uplink transmit power set , Represents the user's mobile device in the Red-billed Blue Magpie individual Uplink transmission power, data volume set , Represents the user's mobile device in the Red-billed Blue Magpie individual The amount of data and task volume offloaded to the associated small base station , Represents the user's mobile device in the Red-billed Blue Magpie individual Or the amount of tasks offloaded from the small base station associated with it to the macro base station; The fitness function is used to calculate the fitness values ​​of all red-billed blue magpies in the original population, and the individual with the largest fitness value is taken as the best red-billed blue magpie individual in history; Step S2.3: Determine the current number of iterations Whether the maximum number of iterations has been exceeded , if the current iteration number Less than or equal to the maximum number of iterations , then search for prey, capture prey, and store food for the red-billed blue magpie population; if the current number of iterations is Greater than the maximum number of iterations Then directly output the information coding of all red-billed blue magpies individuals.

5. The energy-efficient collaborative offloading method for ultra-dense NOMA-MEC networks according to claim 4 is characterized in that: Red-billed blue magpie Searching for prey in small groups The process of position update is expressed as: ； in Respectively represent The transmission power selected by the user's mobile device during the round iteration, the amount of task data unloaded by the user to the small base station, the amount of task data unloaded by the user to the small base station, the base station index selected by the user's mobile device, the encryption algorithm index selected by the user's mobile device, and the channel index selected by the user's mobile device; similarly Respectively Iterate the transmission power selected by the user's mobile device, the amount of task data unloaded by the user to the small base station, the amount of task data unloaded by the user to the small base station, the base station index selected by the user's mobile device, the encryption algorithm index selected by the user's mobile device, and the channel index selected by the user's mobile device; They are respectively the transmission power selected by the user's mobile device, the amount of task data offloaded by the user to the small base station, the amount of task data offloaded by the user to the small base station, the base station index selected by the user's mobile device, the encryption algorithm index selected by the user's mobile device, and the channel index selected by the user's mobile device; is the iteration coefficient, through which the individual position information of the red-billed blue magpie is continuously optimized according to the number of iterations; They respectively represent the transmission power selected by the user mobile device randomly selected in the population, the amount of task data unloaded by the user to the small base station, the amount of task data unloaded by the user to the small base station, the base station index selected by the user mobile device, the encryption algorithm index selected by the user mobile device, and the location information of the channel index selected by the user mobile device, Express Round down, Representation output A random number between Indicates the number of individuals in the red-billed blue magpie group.

6. The energy-efficient collaborative offloading method for ultra-dense NOMA-MEC networks according to claim 5 is characterized in that: When the red-billed blue magpie The process of updating the position parameters when forming a cluster to search for prey is expressed as: ； It indicates the number of individuals of red-billed blue magpies that form small groups to hunt, and the value is between between.

7. The energy-efficient collaborative offloading method for ultra-dense NOMA-MEC networks according to claim 6 is characterized in that: The red-billed blue magpie population forms different teams to catch prey. When the teams are formed into groups, the red-billed blue magpie individuals When capturing prey, the position parameters are updated as follows: ； in, The initialization optimal individual position information corresponding to the transmission power selected by the user's mobile device, the amount of task data unloaded by the user to the small base station, the base station index selected by the user's mobile device, the encryption algorithm index selected by the user's mobile device, and the channel index selected by the user's mobile device are respectively: ； Dynamic Factor and angle variables It is used to dynamically adjust the position of the red-billed blue magpie to prevent it from falling into the local optimum; when the team is composed of a cluster, the individual red-billed blue magpie The position parameters are updated as follows: 。 8. The energy-efficient collaborative offloading method for ultra-dense NOMA-MEC networks according to claim 7 is characterized in that: The specific process of step S3 is as follows: Step S3.1: Inherit the relevant position information of all red-billed blue magpies in the original population to the predator proxy individuals in the eel-grouper population, and calculate the fitness value of the proxy individuals; Step S3.2: The predator agent with the best fitness value in the current position is regarded as the best individual in history; Step S3.3: Determine the current number of iterations Whether the maximum number of iterations has been exceeded , if the current iteration number Less than or equal to the maximum number of iterations , then the eel-grouper population will implement internal signal transmission and cooperative attack operations; if the current iteration number Greater than the maximum number of iterations Then directly output the information coding and fitness value of the best eel-grouper individual in history; Step S3.4: The specific operation of signal transmission is as follows: ； in , , Express Take the absolute value; Step S3.5: The specific operation of the cooperative attack is as follows: ； Where b is the iterative random coefficient, Represents a natural constant of Second power, is the ratio of pi; Step S3.6: Set the number of iterations ; Step S3.7: Perform energy-efficient task offloading related configuration according to the desired optimal individual related parameters.

Images