CN114936075B - Method for unloading computing tasks of mobile audit equipment in edge computing environment - Google Patents
Method for unloading computing tasks of mobile audit equipment in edge computing environment Download PDFInfo
- Publication number
- CN114936075B CN114936075B CN202210339714.XA CN202210339714A CN114936075B CN 114936075 B CN114936075 B CN 114936075B CN 202210339714 A CN202210339714 A CN 202210339714A CN 114936075 B CN114936075 B CN 114936075B
- Authority
- CN
- China
- Prior art keywords
- task
- wolf
- sequence
- wolfs
- unloading
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/48—Program initiating; Program switching, e.g. by interrupt
- G06F9/4806—Task transfer initiation or dispatching
- G06F9/4843—Task transfer initiation or dispatching by program, e.g. task dispatcher, supervisor, operating system
- G06F9/4881—Scheduling strategies for dispatcher, e.g. round robin, multi-level priority queues
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/44—Arrangements for executing specific programs
- G06F9/445—Program loading or initiating
- G06F9/44594—Unloading
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Abstract
The invention discloses a method for unloading a computing task of mobile audit equipment in an edge computing environment. Firstly, reading user equipment information and related information of a task on equipment; secondly, inputting the read equipment information and the task information into a gray wolf algorithm to obtain an unloading scheme of the task, wherein the gray wolf algorithm at least comprises a mapping method for mapping gray wolfs into task sequences based on position information of gray wolf individuals (hereinafter referred to as gray wolfs), a strategy for distributing transmitting antenna power for each task, and a task sequence disturbance method based on a double-point hybridization strategy; and finally, the device unloads the task to the edge server to be executed according to the unloading scheme. The invention can effectively shorten the time for processing the calculation task by the mobile audit equipment in the marginal calculation environment and improve the audit work efficiency.
Description
Technical Field
The invention relates to the field of edge computing, in particular to a method for unloading computing tasks of mobile auditing equipment in an edge computing environment.
Background
With the gradual increase of the popularity of the mobile intelligent equipment and the arrival of the world of everything interconnection, the mobile intelligent equipment is also widely applied to the field of auditing. The use of mobile auditing equipment (including smart phones, notebooks, software and hardware integrated customized terminals and the like, hereinafter referred to as equipment for short) greatly improves the efficiency of auditing work. However, the battery energy storage and the memory space of the device are limited, which becomes an important factor that restricts the device from timely processing a large number of calculation tasks (including audit data association comparison, audit model operation, audit doubtful point discovery and the like, hereinafter referred to as tasks) generated during audit. The edge computing is to deploy an edge server (such as a mobile audit car) at a wireless network (such as a wifi network, a 4G network or a 5G network) access place near a user, wherein the edge computing server has strong computing capability and can provide computing services for equipment. The user unloads tasks with complex calculation amount on the equipment to the edge server for execution, so that the energy consumption on the equipment is saved, and the time delay caused by huge calculation amount of the tasks is reduced. Obviously, the calculation task generated during the audit of the mobile audit equipment is unloaded to the edge server to be executed, so that the time length of the audit work can be greatly shortened, and the audit efficiency is effectively improved.
Currently, scholars have achieved certain results in the research of task scheduling problems of edge computing, but still have some defects. For example, the document "MAUl: mapping smart class load with code offload" discloses a method for real-time application offload decision-making, which has the feature of formalizing the offload decision as a 0-1 linear programming problem, which makes this method have the disadvantage of too high computational cost for solving integer programming, making it unsuitable for the edge task offload problem. The document "clone cloud" Elastic execution between mobile device and cloud "discloses a method of cloning a cloud model, which has a feature that an off-line algorithm can be used to calculate an unloading decision based on a given network transmission rate and equipment computing power, but the feature makes the method have the disadvantage of excessively high optimization complexity. The document 'enclosing interactive performance applications on mobile devices' discloses a task unloading method based on a greedy mechanism, which has the characteristics of high algorithm solving speed and dependence on network bandwidth, and the characteristics ensure that the method has the defects of long difference between the optimal solution of unloading decision and the actual solution and poor execution efficiency when the network bandwidth is limited a lot.
Disclosure of Invention
The invention aims to provide a method for unloading a computing task of mobile audit equipment in an edge environment, which comprises the steps of reading user equipment information and related information of the task on the equipment, inputting the read equipment information and the read task information into a computing task unloading method for obtaining an unloading scheme of the task based on a wolf algorithm, and unloading the task to an edge server by the equipment according to the unloading scheme for execution. The invention can effectively shorten the time for processing the calculation task by the mobile audit equipment in the marginal calculation environment and improve the audit work efficiency.
A method for unloading computing tasks of mobile auditing equipment in an edge environment is characterized in that an unloading scheme is obtained by using a wolf algorithm, and the tasks are unloaded to an edge server according to the obtained unloading scheme for execution, and the method mainly comprises the following steps:
step S1, reading equipment information and relevant information of a task on equipment;
s2, inputting the read task information into a calculation task unloading method for obtaining an unloading scheme of a task based on a gray wolf algorithm, wherein the gray wolf algorithm at least comprises a mapping method for mapping gray wolfs into a task sequence based on position information of a gray wolf individual (hereinafter, referred to as gray wolf), a strategy for distributing transmitting antenna power for each task, and a task sequence disturbance method based on a double-dot hybridization strategy;
and S3, the equipment unloads the tasks to the edge server to execute according to the unloading scheme.
The device information and the related information of the task on the device read in step S1 include, but are not limited to, the number of tasks, the data size of the task, the number of CPU cycles required for each data size of the task, and the like; obtaining n independent computing tasks to be processed by a user mobile device, each task being offloaded and scheduled, as described in detail as pi = { pi = { [ pi ]) 1 ，π 2, …,π n }。
The gray wolf algorithm in step S2 includes the following main steps:
step S21, initializing a wolf population;
step S22, applying a mapping method for mapping the wolf into a task sequence based on the wolf position information;
step S23, applying a strategy of distributing transmitting antenna power for each task;
step S24, calculating the fitness value of each wolf according to the corresponding task sequence;
step S25, sorting the fitness values of all the wolfs in a descending order, and sequentially recording the optimal 3 wolf individuals as alpha wolfs, beta wolfs and delta wolfs;
step S26, comparing the fitness value of the alpha wolf and the best wolf, if the fitness value of the alpha wolf is better than that of the best wolf, updating the best wolf into the alpha wolf, otherwise, the best wolf is not changed;
step S27, judging whether the iteration times are reached, if so, outputting an optimal solution (unloading scheme), and turning to the step S3, otherwise, turning to the step S28;
step S28, updating the gray wolf position information by the gray wolf algorithm position updating mode;
step S29, updating the wolf population by a probability method based on fitness value according to the thought of survival of the fittest;
and step S210, hybridizing the task sequences corresponding to all the wolfsbane by using a task sequence perturbation method based on a double-dot hybridization strategy, and repeating the steps S22 to S27.
Initializing the gray wolf population in step S21, i.e. randomly generating x gray wolfs, gray wolf omega m Is represented by an n-dimensional array
Wherein
Is a real number, expressed as the value at position l of the grey wolf m, l =1,2, \ 8230;, n, n is the number of tasks;
in the mapping method for mapping the grayish wolf into the task sequence based on the grayish wolf position information in step S22, firstly, Ω of each wolf is randomly generated m As a historical optimal solution for this wolf
The task offload sequence of (1). Then, we pass the formula
To calculate the historical optimal solution of inheriting the wolf individual at the wolf n-dimensional position
While obtaining a random number λ between 0 and 1 when ρ is the probability ρ of the task corresponding to the position i>At λ timeWe will
Copying tasks corresponding to position i on the sequence to a new solution ζ m On the corresponding i position, otherwise add the task to
The above. Finally, for
Task of (1) randomly inserts into ζ m Form an unloading sequence ζ of a complete new task m ；
The application of step S23 allocates the transmit antenna power to each task, i.e. obtains the task unloading sequence ζ m Then, the selection of the unloading power when the task is unloaded to the MEC server depends on the strategy of power distribution; in order to ensure that the overall energy consumption can meet the maximum energy consumption requirement on the equipment, assuming that the unloading power of all tasks is the lowest power, and calculating the transmission energy consumption E for transmitting all tasks from the equipment end to the MEC server; if the overall transmission energy consumption of the task is less than the upper limit E of the energy consumption of the equipment max Sequentially increasing the power of the tasks by using a breadth-first method until the upper limit of energy consumption is reached;
step S24, calculating the fitness value of each wolf according to the corresponding task sequence, wherein the fitness value of each wolf is the reciprocal of the maximum completion time of all tasks, and the larger the fitness value is, the smaller the maximum completion time of each wolf is;
the output optimal solution (unloading scheme) of step S27 includes the unloading sequence of the task, the power allocation condition of the task sequence and the maximum completion time of the task sequence;
step S28, updating the gray wolf position, namely guiding other wolfs to complete the search to the optimal area by the best 3-wolf alpha wolf, beta wolf and delta wolf in each generation of population, continuously iterating and updating the population, and finally realizing the process of global optimization;
and step S29, updating the grey wolf population by a probability method, namely updating the grey wolf population by adopting a wheel disc betting mode, adding the fitness values of all grey wolf individuals, calculating the ratio of the fitness value of each grey wolf to the total fitness value to obtain the probability of transmitting each grey wolf to the next generation population, and dividing the area percentage of sectors occupied on the wheel disc according to the probability. Then we obtain a random number between 0 and 1, and the corresponding wolf individual of the area on the wheel represented by the random number is selected to the next generation population. Repeating this step until the population number of wolfs is met;
in the step S210, the task sequence perturbation method based on the double-dot hybridization strategy performs hybridization on the task sequences corresponding to all the sirius, that is, performs double-dot hybridization on the task sequences corresponding to any two sirius. Specifically, two gray wolves m are randomly selected 1 And m 2 Randomly generating two unequal integers a and b, wherein a, b belongs to (0, n), and exchanging zeta in task sequences corresponding to the two gray wolves m1 And ζ m2 Generates a new sequence of ζ 'from the sequence of tasks between the a-th task and the b-th task of (c)' m1 And ζ' m2 And carrying out power distribution on the new sequence, calculating the fitness value, updating the sequence of the grey wolf into the new sequence and updating the fitness value of the grey wolf if the new sequence is superior to the original sequence. Repeating the steps until all the wolfs complete double-dot hybridization, and each wolf is only carried out once in each iteration;
compared with the prior art, the invention has the beneficial effects that:
1) The method is designed based on the meta-starting algorithm, and can better provide an effective solution for the task unloading problem in the edge computing environment;
2) The method can effectively avoid the situation of falling into local optimum, and the obtained result is close to the actual optimum solution of task unloading.
Drawings
Fig. 1 is a drawing of the abstract of the present invention, illustrating the process of the method as a whole.
FIG. 2 is a detailed flow chart of the method of the present invention.
Detailed Description
The technical scheme of the invention is explained in detail with reference to fig. 2, and the specific steps comprise:
step S1, reading equipment information and relevant information of tasks on equipment, wherein the relevant information includes but is not limited to the number of the tasks, the data volume of the tasks, the number of CPU cycles required by each bit of data volume of the tasks and the like;
step S2, inputting the read task information into a calculation task unloading method for obtaining an unloading scheme of a task based on a gray wolf algorithm, wherein the gray wolf algorithm at least comprises a mapping method for mapping gray wolfs into task sequences based on position information of gray wolf individuals (hereinafter referred to as gray wolfs), a strategy for distributing transmitting antenna power for each task, and a task sequence disturbance method based on a double-point hybridization strategy;
and S3, the equipment unloads the task to the edge server to be executed according to the unloading scheme.
Wherein, step S2 includes the following main steps:
step S21, initializing a wolf population, including the number of the wolf populations and wolf position information;
step S22, a mapping method for mapping the wolf into the task sequence based on the wolf position information is applied, and the specific process is that omega of each wolf is randomly generated firstly m And as a historical optimal solution for this wolf
Then, we pass the formula
To calculate a historical optimal solution for inheriting an individual graying wolf at a graying wolf n-dimensional position
While obtaining a random number λ between 0 and 1, when ρ>At λ, we will
Copying tasks corresponding to position i on the sequence to a new solution ζ m Corresponding i bitPut on, otherwise add the task to
The above. Finally, for
Task of (1) randomly inserts into ζ m Form a complete new solution ζ m ；
Step S23, applying a strategy of distributing transmitting antenna power for each task; in order to ensure that the overall energy consumption can meet the maximum energy consumption requirement on the equipment, assuming that the unloading power of all tasks is the lowest power, and calculating the transmission energy consumption E for transmitting all tasks from the equipment end to the MEC server. If the overall transmission energy consumption E is less than the upper limit E of the equipment energy consumption max Sequentially increasing the power of the tasks by using a breadth-first method until the upper limit of energy consumption is reached;
step S24, calculating the fitness value of each wolf according to the corresponding task sequence;
s25, sorting the fitness values of all the gray wolves in a descending order, and sequentially recording the optimal 3 gray wolves as alpha wolves, beta wolves and delta wolves;
step S26, comparing the fitness value of the alpha wolf and the best wolf, if the fitness value of the alpha wolf is better than that of the best wolf, updating the best wolf into the alpha wolf, otherwise, the best wolf is not changed;
step S27, judging whether the iteration times are reached, if so, outputting an optimal solution (unloading scheme), and turning to step S3, otherwise, turning to step S28;
step S28, updating the gray wolf position information by the gray wolf algorithm position updating mode;
and S29, updating the population of the wolfs by a probability method, namely adding the fitness values of all wolfs, calculating the ratio of the fitness value of each wolf to the total fitness value, obtaining the probability of each wolf being transmitted to the next generation of population, and dividing the area percentage of the sector on the wheel disc according to the probability. Then we obtain a random number between 0 and 1, and the corresponding wolf individual of the area on the wheel represented by the random number is selected to the next generation population. Repeating this step until the population number of wolfs is met;
step S210, a task sequence perturbation method based on a double-dot hybridization strategy is applied to hybridize the task sequences corresponding to all the sirius, that is, to perform double-dot hybridization on the task sequences corresponding to any two sirius, and each sirius is performed only once in each iteration, and steps S22 to S27 are repeated.
Examples
The following describes an embodiment in detail with reference to fig. 2, in this embodiment, the user's mobile device needs to process 4 tasks, and the following tasks are required to be offloaded to the MEC server to execute an optimal solution with a minimum completion time. The data volume of the task is as follows:
| task numbering | Data volume (bit) | CPU cycle number (cycle/bit) |
| 0 | 619 | 1193 |
| 1 | 1626 | 658 |
| 2 | 339 | 1443 |
| 3 | 1805 | 1457 |
Step S1, acquiring a calculation task to be processed by user mobile equipment, wherein the calculation task comprises the data volume of the task and the number of CPU cycles required by each data volume, and defining the task set as pi = { pi = (pi) = 0 ,π 1 ,π 2 ,π 3 In which pi 0 ,π 1 ,π 2 ,π 3 The data volume of (2) is 619, 1626, 339 and 1805 (bit), the CPU period number of each data volume is 1193, 658, 1443 and 1457 (cycle/bit), and a relevant unloading model is established;
step S21, initializing a wolf population, defining the wolf population number to be 10, and setting the position information of the wolf as the following table;
step S22, applying a mapping method for mapping the wolf into the task sequence based on the wolf position information: firstly randomly generating omega of each wolf m ζ, for gray wolf Ω 1 Randomly generating an initial solution ζ = (π) 0 ,π 3 ,π 1 ,π 2 ) Calculating the probability rho of each task being relayed on the solution, and pi for the 0 th task 0 The probability of being relayed is
Then, a random number λ of 0.6162 between 0 and 1 is obtained because ρ>λ, so the task of the 0 th position on the ζ sequence is π 0 Copying to a New solution ζ m At the 0 th position, i.e.
If ρ<λ, will task π 0 Is added to
Performing the following steps; by parity of reasoning, obtain
To pair
After the tasks in (1) are randomly disturbed, insert ζ m In the experiment, ζ was obtained m ＝(π 0 ,π 3 ,π 1 ,π 2 )。
Step S23, applying a strategy of distributing transmitting antenna power for each task: this experiment set the variable power of the antenna to p k E {70,80,90,100}, with the unit mW. For task sequence ζ = (π) 0 ,π 3 ,π 1 ,π 2 ) Firstly, the power of all tasks is set to 70mW, and the overall transmission energy consumption of the tasks is calculated
Wherein p is k Is the antenna transmission power of the task,
for a task pi i R is the transmission rate of the task,
the parameter settings related to R are as follows:
when the transmission power of all tasks was set to 70mW, E =0.0734,e was calculated max Is 0.08 due to E<E max Thus, it adjusts π for the 0 th task in task sequence ζ 0 Of the transmission power of the re-calculation taskOverall transmission energy consumption, if the overall transmission energy consumption E is less than the upper limit E of the equipment energy consumption max Adjusting the transmitting power of the next task until the upper limit of energy consumption is reached; the transmission power of the final task is 80Mw, and the overall transmission energy consumption is E =0.0798;
step S24, calculating the fitness value of each wolf according to the corresponding task sequence; the fitness value of the gray wolf is related to the completion time of a task sequence corresponding to the gray wolf, and the completion time of the task comprises the time when the task is transmitted to the server and the time when the task is executed on the server.
S25, sorting the fitness values of all the gray wolves in a descending order, and sequentially recording the optimal 3 gray wolves as alpha wolves, beta wolves and delta wolves;
step S26, comparing the fitness value of the alpha wolf and the best wolf, if the fitness value of the alpha wolf is better than that of the best wolf, updating the best wolf into the alpha wolf, otherwise, the best wolf is not changed;
step S27, judging whether the iteration times are reached, if so, outputting an optimal solution (unloading scheme), and turning to step S3, otherwise, turning to step S28;
step S28, updating the gray wolf position information by the gray wolf algorithm position updating mode;
and S29, updating the population of the wolfs by a probability method, namely adding the fitness values of all wolfs, calculating the ratio of the fitness value of each wolf to the total fitness value, obtaining the probability of each wolf being transmitted to the next generation of population, and dividing the area percentage of the sector on the wheel disc according to the probability. The following figures:
then, a random number rnd between 0 and 1 is obtained, and if the rnd is 0.7724, the wolf with the probability interval number of 7 is selected into a new population; repeating the steps until the number of the wolf populations is met; the numbers of the 10 grey wolves obtained at this time were 0,2,9,1,7,5,4,3,6,0, respectively;
step S210, a task sequence disturbance method based on a double-dot hybridization strategy is applied to hybridize the task sequences corresponding to all the wolfs, and the task sequences corresponding to all the wolfs are hybridized, firstly, any two wolfs in a population are randomly divided into a group, and then the grouping condition is represented by the following numbers of the wolfs: {0,9}, {7,5}, {2,6}, {0,1} and {4,3}; the gray wolfs with the number 0 and the number 9 are hybridized, and the corresponding task sequences are respectively (pi) 2 ,π 0 ,π 3 ,π 1 ) And (pi) 2 ,π 0 ,π 1 ,π 3 ) Randomly taking two integers m 1 And m 2 1 and 3 respectively, the task sequences between the sequence position 0 and the sequence position 2 are exchanged to obtain the sequence with the Grey wolf number of 0 as (pi) 2 ,π 0 ,π 3 ,π 1 ) And the sequence of Grey wolf number 9 is (pi) 2 ,π 0 ,π 1 ,π 3 ) (ii) a Distributing power for the task sequence and calculating a fitness value; comparing the fitness value of the grayish wolf sequence before and after hybridization, and updating relevant information of the grayish wolf into the information (including task sequence, power distribution and fitness value) after hybridization if the fitness value after hybridization is greater than the fitness value before hybridization; completing the hybridization of the grouped wolfsbane; step S22 to step S27 are repeated.
And S3, the equipment unloads the task to the edge server to be executed according to the unloading scheme.
The invention provides a method for uninstalling computing tasks of mobile audit equipment based on an edge computing environment, and a plurality of methods and ways for implementing the technical scheme, where the above description is only a preferred embodiment of the invention, it should be noted that, for those skilled in the art, a plurality of improvements and modifications may be made without departing from the principle of the invention, and these improvements and modifications should also be regarded as the scope of the invention. All the components not specified in the present embodiment can be realized by the prior art.
Claims (9)
1. A method for unloading computing tasks of mobile auditing equipment in an edge environment is characterized in that a wolf algorithm is used for obtaining an unloading scheme and unloading the tasks to an edge server for execution according to the obtained unloading scheme, and the method mainly comprises the following steps:
step S1, reading equipment information and relevant information of a task on equipment;
s2, inputting the read task information into a gray wolf algorithm to obtain an unloading scheme of the task, wherein the gray wolf algorithm at least comprises a mapping method for mapping gray wolfs into task sequences based on gray wolf position information, a strategy for distributing transmitting antenna power for each task, and a task sequence disturbance method based on a double-dot hybridization strategy; the grey wolf algorithm in the step S2 comprises the following main steps:
step S21, initializing a wolf population; randomly generating x wolfs, wolf omega m Is represented by an n-dimensional array
Wherein
Is a real number, expressed as the value at the position l of the grey wolf m, l =1,2, \8230, n, n is the number of tasks;
step S22, applying a mapping method for mapping the wolf into a task sequence based on the wolf position information;
step S23, applying a strategy of distributing transmitting antenna power for each task;
step S24, calculating the fitness value of each wolf according to the corresponding task sequence;
s25, sorting the fitness values of all the gray wolves in a descending order, and sequentially recording the optimal 3 gray wolves as alpha wolves, beta wolves and delta wolves;
step S26, comparing the fitness value of the alpha wolf and the best wolf, if the fitness value of the alpha wolf is better than that of the best wolf, updating the best wolf into the alpha wolf, otherwise, the best wolf is not changed;
step S27, judging whether the iteration times are reached, if so, outputting an optimal solution, namely, outputting an unloading scheme, turning to step S3, otherwise, turning to step S28;
step S28, updating the gray wolf position information by the gray wolf algorithm position updating mode;
step S29, updating the wolf population by using a probability method based on the fitness value;
step S210, a task sequence disturbance method based on a double-dot hybridization strategy is applied to hybridize the task sequences corresponding to all wolfsbanes, and the steps S22 to S27 are repeated;
and S3, the equipment unloads the task to the edge server to be executed according to the unloading scheme.
2. The method of claim 1, wherein the device information read and related information of the task on the device, including but not limited to the number of tasks, the data size of the task, the number of CPU cycles required for each data size of the task; acquiring n independent computing tasks pi to be processed by a user mobile device, wherein each task can be unloaded and scheduled, and the specific description is pi = { pi = [ pi ]) 1 ，π 2， …，π n }。
3. The method according to claim 1, wherein the mapping method for mapping the grayish wolf into the task sequence based on the grayish wolf location information in step S22: first, each wolf Ω is randomly generated m As a historical optimal solution for this wolf
The task offload sequence of (1); then, by the formula
To calculate a historical optimal solution for inheriting an individual graying wolf at a graying wolf n-dimensional position
While obtaining one, the probability p of the task corresponding to the position iA random number λ between 0 and 1 when ρ>At λ, we will
Replication of tasks corresponding to position i on the sequence to a new solution ζ m On the corresponding i position, otherwise add the task to
The above step (1); finally, for
Task of (1) is randomly inserted into ζ m Form an unloading sequence ζ of a complete new task m 。
4. The method of claim 1, wherein the strategy of allocating transmit antenna power for each task in step S23 is as follows: get task unload sequence ζ m Then, the selection of the unloading power when the task is unloaded to the MEC server depends on the strategy of power distribution; in order to ensure that the overall energy consumption can meet the maximum energy consumption requirement on the equipment, assuming the unloading power of all tasks as the lowest power, calculating the transmission energy consumption E for transmitting all tasks from the equipment end to the MEC server; if the overall transmission energy consumption of the task is less than the upper limit E of the energy consumption of the equipment max And sequentially increasing the power of the tasks by using a breadth-first method until the upper limit of energy consumption is reached.
5. The method according to claim 1, wherein the step S24 of calculating the fitness value of each wolf from the corresponding task sequence: the fitness value of the gray wolf is the reciprocal of the maximum completion time of all tasks, and the larger the fitness value is, the smaller the maximum completion time of the gray wolf is.
6. The method of claim 1, wherein the outputting of the optimal solution of step S27: the unloading scheme comprises an unloading sequence of tasks, a power distribution condition of the task sequence and a maximum completion time of the task sequence.
7. The method according to claim 1, characterized in that the updating of the grey wolf location of step S28: the method is a process of guiding other wolfs to search and finish the optimal area, continuously iterating and updating the population and finally realizing global optimization by the best-expressing 3 wolfs alpha wolfs, beta wolfs and delta wolfs in the population of each generation.
8. The method according to claim 1, wherein the step S29 of updating the population of sirius with a probabilistic method based on fitness values: updating the population of the grey wolfs by adopting a roulette wheel mode, adding the fitness values of all the grey wolfs, calculating the ratio of the fitness value of each grey wolf to the total fitness value, obtaining the probability of each grey wolf being transmitted to the next generation of population, and dividing the area percentage of the sector occupied by the wheel according to the probability; then, a random number between 0 and 1 is obtained, and the corresponding wolf individuals in the area on the wheel disc represented by the random number are selected to the next generation of population; this step is repeated until the population of wolfs is met.
9. The method according to claim 1, wherein the task sequence perturbation method based on the two-point hybridization strategy in step S210: performing double-point hybridization on task sequences corresponding to any two Huidou wolfs, specifically randomly selecting two Huidou wolfs m 1 And m 2 Randomly generating two unequal integers a and b, wherein a, b belongs to (0, n), and exchanging zeta in task sequences corresponding to the two gray wolves m1 And ζ m2 Generates a new sequence of ζ 'from the sequence of tasks between the a-th task and the b-th task of (c)' m1 And ζ' m2 Carrying out power distribution on the new sequence, calculating a fitness value, if the new sequence is superior to the original sequence, updating the sequence of the grey wolf into the new sequence, and updating the fitness value of the grey wolf; this procedure is repeated until all of the wolfs complete the double-dot hybridization, and each wolf is performed only once in each iteration.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210339714.XA CN114936075B (en) | 2022-04-01 | 2022-04-01 | Method for unloading computing tasks of mobile audit equipment in edge computing environment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210339714.XA CN114936075B (en) | 2022-04-01 | 2022-04-01 | Method for unloading computing tasks of mobile audit equipment in edge computing environment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114936075A CN114936075A (en) | 2022-08-23 |
| CN114936075B true CN114936075B (en) | 2023-02-14 |
Family
ID=82863146
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210339714.XA Active CN114936075B (en) | 2022-04-01 | 2022-04-01 | Method for unloading computing tasks of mobile audit equipment in edge computing environment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114936075B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108429633A (en) * | 2018-01-18 | 2018-08-21 | 西南交通大学 | Virtual network function dispositions method based on grey wolf algorithm |
| CN109858739A (en) * | 2018-12-20 | 2019-06-07 | 电子科技大学中山学院 | logistics transportation scheduling method, computer storage medium and electronic equipment |
| CN113364859A (en) * | 2021-06-03 | 2021-09-07 | 吉林大学 | MEC-oriented joint computing resource allocation and unloading decision optimization scheme in Internet of vehicles |
| CN113543176A (en) * | 2021-07-08 | 2021-10-22 | 中国科学院深圳先进技术研究院 | Unloading decision method of mobile edge computing system based on assistance of intelligent reflecting surface |
| CN113743566A (en) * | 2021-08-17 | 2021-12-03 | 北京梧桐车联科技有限责任公司 | Product disassembly sequence optimization method and device, computer equipment and storage medium |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111586607B (en) * | 2020-04-15 | 2022-02-22 | 杭州电力设备制造有限公司 | Intelligent electric meter wireless sensor network layout method based on improved wolf algorithm |
| CN113011612B (en) * | 2021-03-23 | 2022-09-20 | 合肥工业大学 | Production and maintenance scheduling method and system based on improved wolf algorithm |
| CN113342124B (en) * | 2021-06-11 | 2022-08-09 | 中国电建集团华东勘测设计研究院有限公司 | Photovoltaic MPPT method based on improved wolf optimization algorithm |
-
2022
- 2022-04-01 CN CN202210339714.XA patent/CN114936075B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108429633A (en) * | 2018-01-18 | 2018-08-21 | 西南交通大学 | Virtual network function dispositions method based on grey wolf algorithm |
| CN109858739A (en) * | 2018-12-20 | 2019-06-07 | 电子科技大学中山学院 | logistics transportation scheduling method, computer storage medium and electronic equipment |
| CN113364859A (en) * | 2021-06-03 | 2021-09-07 | 吉林大学 | MEC-oriented joint computing resource allocation and unloading decision optimization scheme in Internet of vehicles |
| CN113543176A (en) * | 2021-07-08 | 2021-10-22 | 中国科学院深圳先进技术研究院 | Unloading decision method of mobile edge computing system based on assistance of intelligent reflecting surface |
| CN113743566A (en) * | 2021-08-17 | 2021-12-03 | 北京梧桐车联科技有限责任公司 | Product disassembly sequence optimization method and device, computer equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114936075A (en) | 2022-08-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111507601B (en) | Resource optimization allocation decision method based on deep reinforcement learning and block chain consensus | |
| CN109002358B (en) | Mobile terminal software self-adaptive optimization scheduling method based on deep reinforcement learning | |
| CN110928654B (en) | Distributed online task unloading scheduling method in edge computing system | |
| CN112286677B (en) | Resource-constrained edge cloud-oriented Internet of things application optimization deployment method | |
| CN109144719B (en) | Collaborative unloading method based on Markov decision process in mobile cloud computing system | |
| CN110471762B (en) | Cloud resource allocation method and system based on multi-objective optimization | |
| CN103281374B (en) | A kind of method of data fast dispatch during cloud stores | |
| CN112995289B (en) | Internet of vehicles multi-target computing task unloading scheduling method based on non-dominated sorting genetic strategy | |
| CN105808339B (en) | Big data parallel calculating method and device | |
| CN110162390B (en) | Task allocation method and system for fog computing system | |
| CN112860337B (en) | Method and system for unloading dependent tasks in multi-access edge computing | |
| CN111324422B (en) | Multi-target virtual machine deployment method, device, equipment and storage medium | |
| CN113610227A (en) | Efficient deep convolutional neural network pruning method | |
| CN113590279A (en) | Task scheduling and resource allocation method for multi-core edge computing server | |
| CN114936075B (en) | Method for unloading computing tasks of mobile audit equipment in edge computing environment | |
| Chang et al. | Determining the number of kanbans and lotsizes in a generic kanban system: a simulated annealing approach | |
| Zhang et al. | Quantified edge server placement with quantum encoding in internet of vehicles | |
| CN114980216B (en) | Dependency task unloading system and method based on mobile edge calculation | |
| CN102845028A (en) | Method and device for resource allocation | |
| CN113709817B (en) | Task unloading and resource scheduling method and device under multi-base-station multi-server scene | |
| CN113747500A (en) | High-energy-efficiency low-delay task unloading method based on generation countermeasure network in mobile edge computing environment | |
| CN114912647A (en) | Apparatus, method and machine-readable storage medium for decision making | |
| CN110689320A (en) | Large-scale multi-target project scheduling method based on co-evolution algorithm | |
| CN114399152B (en) | Method and device for optimizing comprehensive energy scheduling of industrial park | |
| CN115065683B (en) | Vehicle edge network task allocation and unloading method based on vehicle clustering |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |