CN115915278B - Task unloading method for Internet of vehicles - Google Patents

Task unloading method for Internet of vehicles Download PDF

Info

Publication number
CN115915278B
CN115915278B CN202310194757.8A CN202310194757A CN115915278B CN 115915278 B CN115915278 B CN 115915278B CN 202310194757 A CN202310194757 A CN 202310194757A CN 115915278 B CN115915278 B CN 115915278B
Authority
CN
China
Prior art keywords
user
system loss
solution
task
representing
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
Application number
CN202310194757.8A
Other languages
Chinese (zh)
Other versions
CN115915278A (en
Inventor
沈艳
胡辉
陈姣
史奎锐
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Chengdu University of Information Technology
Original Assignee
Chengdu University of Information Technology
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Chengdu University of Information Technology filed Critical Chengdu University of Information Technology
Priority to CN202310194757.8A priority Critical patent/CN115915278B/en
Publication of CN115915278A publication Critical patent/CN115915278A/en
Application granted granted Critical
Publication of CN115915278B publication Critical patent/CN115915278B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • 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
    • Y02D30/00—Reducing energy consumption in communication networks
    • Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks

Landscapes

  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

The invention discloses a task unloading method oriented to the Internet of vehicles, which relates to the field of mobile edge calculation and comprises the following steps: s1, constructing a system loss model; s2, calculating the system loss and the overall constraint violation degree of an initial solution of a system loss model; s3, ordering initial solutions of the system loss model, and initializing weights of the initial solutions of the system loss model; when the global constraint violation of the non-feasible solution is at the thresholdεWhen inside, as a feasible solution; s4, solving a system loss model to obtainMNew solutions; s5, obtaining the current optimal feasible solution, and repeating the step S3 to the step until the preset times to obtain the optimal feasible solution which minimizes the system loss; and S6, taking the optimal feasible solution as an internet of vehicles task unloading strategy. The invention considers the problem that the residence time of the user in the RSU service range is smaller than the round trip time of the unloading task between the two, reduces the system loss and improves the frequency utilization rate.

Description

Task unloading method for Internet of vehicles
Technical Field
The invention relates to the field of mobile edge calculation, in particular to a task unloading method oriented to the Internet of vehicles.
Background
Thanks to the advent of mobile edge computing (mobile edge computing, MEC), vehicle users can offload part of the computation-intensive tasks to Roadside Service Units (RSUs) for execution to alleviate the trouble of local resource limitation, thereby reducing the computation latency and the energy consumption of the costs of the tasks. Currently, there have been many related studies including: considering the situation of concurrent multiple multi-priority computing tasks and uneven resource load of the MEC server, providing an unloading strategy based on a genetic algorithm so as to improve the unloading success rate of the security type tasks; the method comprises the steps of providing a combined calculation unloading, calculation resource and wireless resource allocation algorithm under a cloud and mist mixed network architecture, and minimizing system energy consumption and resource cost on the premise of meeting time delay requirements; an orthogonal frequency division multiple access (orthogonal frequencydivision multiple access, OFDMA) technology is adopted for communication among nodes, one sub-channel can only be used by one user in a certain time, and the spectrum utilization rate is low; a policy that minimizes reasonable time allocation and computation offloading of MEC system computation time periods for a plurality of wireless sensor devices to minimize computation periods; a deep learning network-based gaming algorithm optimizes user offload latency and energy consumption, however, RSU service coverage is limited and users are typically mobile, so that if the user's residence time in the RSU service coverage is less than the round trip time of the offload task between the two, the task result will fail back.
Disclosure of Invention
Aiming at the defects in the prior art, the task unloading method for the Internet of vehicles solves the problems of low spectrum utilization rate, low inspection efficiency and high system loss in the prior art.
In order to achieve the aim of the invention, the invention adopts the following technical scheme: a task unloading method facing the Internet of vehicles comprises the following steps:
s1, acquiring user data and constructing a system loss model;
s2, initializing parameters of a system loss model, and calculating system loss and overall constraint violation degree of an initial solution of the system loss model;
s3, sorting initial solutions of the system loss model according to the system loss, and initializing weights of the initial solutions of the system loss model; when the global constraint violation of the non-feasible solution is at the thresholdεWhen the solution is within, the solution is regarded as a feasible solution to be subjected to sorting treatment;
s4, solving a system loss model through an ant colony algorithm to obtain a sub-channel comprising a user task unloading proportion, user transmission power and user usageMNew solutions;
s5, sorting the new solutions and the initial solutions of the system loss model to obtain the current optimal feasible solution, and repeating the step S3 to the step until the preset times to obtain the optimal feasible solution which minimizes the system loss;
and S6, taking the optimal feasible solution which minimizes the system loss as an Internet of vehicles task unloading strategy.
Further, the specific implementation manner of step S1 is as follows:
constructing a system loss model:
Figure SMS_1
Figure SMS_2
Figure SMS_3
Figure SMS_4
Figure SMS_5
Figure SMS_6
Figure SMS_7
wherein ,
Figure SMS_36
representing minimum system loss; />
Figure SMS_41
Representing a system loss function; />
Figure SMS_46
A solution representing a system loss function; />
Figure SMS_9
Representing user +.>
Figure SMS_18
Task unloading time; />
Figure SMS_24
Representing user +.>
Figure SMS_30
The length of the resident time RSU service section in the RSU service range is +.>
Figure SMS_28
User->
Figure SMS_33
Is +.>
Figure SMS_38
，/>
Figure SMS_43
The position of (2) is +.>
Figure SMS_35
；/>
Figure SMS_40
Representing sharing ofUA user;CHrepresenting sharing ofCHSubchannel, & gt>
Figure SMS_45
；/>
Figure SMS_49
Representing user +.>
Figure SMS_34
Task offload ratio of (2); />
Figure SMS_39
Representing user +.>
Figure SMS_44
Is used for the transmission power of the (a); />
Figure SMS_48
Representing user +.>
Figure SMS_8
The sub-channels used; />
Figure SMS_16
Representing the user's preference between latency and energy consumption; />
Figure SMS_22
Representing user +.>
Figure SMS_27
The computation delay of all tasks executed locally; />
Figure SMS_12
Representing user +.>
Figure SMS_14
The energy consumed by all tasks performed locally; />
Figure SMS_20
Representing user +.>
Figure SMS_26
Total energy consumption required by a task; />
Figure SMS_13
For user->
Figure SMS_17
Actual computation delay of task,/->
Figure SMS_23
For user->
Figure SMS_29
Local calculation time delay of the task; />
Figure SMS_11
Representing a user
Figure SMS_19
In subchannel->
Figure SMS_25
Channel gain on; />
Figure SMS_32
Is->
Figure SMS_31
At->
Figure SMS_37
Power gain on; />
Figure SMS_42
Is the path loss coefficient; />
Figure SMS_47
Representing a transmission power threshold; />
Figure SMS_10
Representing user +.>
Figure SMS_15
In subchannel->
Figure SMS_21
Channel gain on the upper.
Further, in step S1, the user
Figure SMS_50
Total energy consumption required for a task, user +.>
Figure SMS_51
The specific implementation mode of the task unloading time is as follows: />
According to the formula:
Figure SMS_52
Figure SMS_53
obtaining local calculation energy consumption
Figure SMS_55
And user->
Figure SMS_58
Local computation delay of task->
Figure SMS_61
； wherein ,/>
Figure SMS_56
For user->
Figure SMS_57
CPU average power,/,>
Figure SMS_60
is->
Figure SMS_63
CPU power consumption coefficient of the same; />
Figure SMS_54
For user->
Figure SMS_59
The working frequency of the local CPU; />
Figure SMS_62
Representing the size of the task data; />
Figure SMS_64
Representing the number of CPU cycles required to calculate each bit of data;
according to the formula:
Figure SMS_65
obtaining the user received by the RSU
Figure SMS_67
Signal-to-interference-plus-noise ratio of signal>
Figure SMS_71
The method comprises the steps of carrying out a first treatment on the surface of the The number of users in the set isS；
Figure SMS_74
Indicating all used subchannels>
Figure SMS_69
Is a set of users; />
Figure SMS_72
For users
Figure SMS_75
Distance between RSU and->
Figure SMS_77
Distance between RSU and service section; />
Figure SMS_66
Representing user +.>
Figure SMS_70
Is used for the transmission power of the (a); />
Figure SMS_73
Maximum transmission power for the user; />
Figure SMS_76
Is Gaussian white noise power spectral density; />
Figure SMS_68
The bandwidth allocated for each sub-channel for the RSU;
according to the formula:
Figure SMS_78
obtaining the user
Figure SMS_79
Transmission rate of->
Figure SMS_80
；
According to the formula:
Figure SMS_81
Figure SMS_82
/>
obtaining the user
Figure SMS_83
Transmission delay during task offloading>
Figure SMS_84
And transmission energy consumption->
Figure SMS_85
；
According to the formula:
Figure SMS_86
obtaining the calculation time delay required by the RSU to execute the task
Figure SMS_87
； wherein ,/>
Figure SMS_88
For RSU server for user +.>
Figure SMS_89
CPU working frequency allocated by the unloaded task;
according to the formula:
Figure SMS_90
Figure SMS_91
obtaining the user
Figure SMS_92
Task offloading latency->
Figure SMS_93
And user->
Figure SMS_94
Total energy consumption required for a task->
Figure SMS_95
。
Further, the specific implementation manner of step S2 is as follows:
s2-1, initializing the number of initial solutions in the ant colonyKInitializing weights of an initial solution
Figure SMS_96
Number of antsMUpper limit of iteration numberT，Overall constraint violation mean of all non-viable solutions in ant colonyε；
S2-2, according to the formula:
Figure SMS_97
obtaining an overall constraint violation function
Figure SMS_98
； wherein ,/>
Figure SMS_99
A violation degree function for time constraint;
Figure SMS_100
a violation degree function for the power constraint; when->
Figure SMS_101
When (I)>
Figure SMS_102
For the system loss function->
Figure SMS_103
Or else, a non-viable solution;
s2-3, calculating the system loss of all solutions according to the system loss function, and calculating the overall constraint violation of all solutions according to the overall constraint violation function.
Further, the specific implementation manner of step S3 is as follows:
s3-1, according to the formula:
Figure SMS_104
/>
based on the possible and non-possible solutions together in an ant colonyεIs a comparison ordering of (2);
Figure SMS_106
indicate->
Figure SMS_109
System loss of individual solutions,/->
Figure SMS_111
Indicate->
Figure SMS_107
System loss of the solution; />
Figure SMS_108
Indicate->
Figure SMS_110
A solution power constraint violation degree; />
Figure SMS_112
Indicate->
Figure SMS_105
Power constraint violation degrees of the individual solutions;
s3-2, according to the formula:
Figure SMS_113
Figure SMS_114
obtaining a subchannel
Figure SMS_129
Weight of +.>
Figure SMS_117
And initial solution->
Figure SMS_124
Weight of +.>
Figure SMS_118
； wherein ,/>
Figure SMS_126
Is a super parameter; />
Figure SMS_119
The more preferred is that,
Figure SMS_123
the greater the +.>
Figure SMS_128
The thicker the pheromone is left, the ant is +.>
Figure SMS_131
Is->
Figure SMS_115
and />
Figure SMS_122
The greater the probability of nearby movement;
Figure SMS_120
for users in ant colony->
Figure SMS_125
Sub-channel +.>
Figure SMS_130
Weights of the optimal solution of ∈1->
Figure SMS_132
For users in ant colony->
Figure SMS_116
Sub-channel +.>
Figure SMS_121
The number of solutions of>
Figure SMS_127
Indicating the number of sub-channels that are not used by the users in all solutions in the ant colony.
Further, the specific implementation manner of step S4 is as follows:
s4-1, according to the formula:
Figure SMS_133
obtaining a random solution
Figure SMS_134
；/>
Figure SMS_135
Is->
Figure SMS_136
Weights of (2);crepresent the firstcPerforming solution;
s4-2, according to the formula:
Figure SMS_137
get ants to move to
Figure SMS_138
Continuous zoneProbability of transition at any point in the space +.>
Figure SMS_139
；/>
Figure SMS_140
Solution selected for ants->
Figure SMS_141
Middle user->
Figure SMS_142
Task offload ratio of (2); />
Figure SMS_143
，/>
Figure SMS_144
Is a super parameter; />
S4-3, according to the transition probability
Figure SMS_145
Randomly move to interval +.>
Figure SMS_146
Any point in the above, get the user in the new solution
Figure SMS_147
Task offloading ratio of->
Figure SMS_148
；
S4-4, according to the formula:
Figure SMS_149
obtaining user power transition probability
Figure SMS_150
； wherein ,/>
Figure SMS_151
，/>
Figure SMS_152
；
S4-5, according to the transition probability
Figure SMS_153
In the continuous interval +.>
Figure SMS_154
Randomly moving to any point to obtain user +.>
Figure SMS_155
Is +.>
Figure SMS_156
；
S4-6, according to the formula:
Figure SMS_157
obtaining sub-channel transition probability of user
Figure SMS_158
； wherein ,/>
Figure SMS_159
Is sub-channel->
Figure SMS_160
Weights of (2);
s4-7, according to the sub-channel transition probability of the user
Figure SMS_161
Move to discrete interval +.>
Figure SMS_162
Inner one point->
Figure SMS_163
Obtaining the user's->
Figure SMS_164
Subchannel number->
Figure SMS_165
；
S4-8, according to the users in the new solution
Figure SMS_168
Subchannel number->
Figure SMS_170
User->
Figure SMS_172
Is +.>
Figure SMS_167
And user ∈>
Figure SMS_169
Task offloading ratio of->
Figure SMS_171
Obtaining ant->
Figure SMS_173
A new solution of 3 XU times in the value space of each variable>
Figure SMS_166
；
S4-9, repeating the steps S4-1 to S4-8 untilMNext, obtainMNew solutions are provided.
Further, the specific implementation manner of step S5 is as follows:
s5-1, to be obtainedMNew solutions and current existing solutionsKThe solutions are subjected to epsilon-based comparison and sorting, and used beforeKThe original solution in the ant colony is replaced by the optimal solution to obtain a replaced solution;
s5-2, judging whether the current iteration reaches the set times, if so, outputting an optimal feasible solution which minimizes the system loss; otherwise, enter step S5-3;
s5-3, calculating the pheromone of the replaced solution, namely the weight of the replaced solution;
s5-4, according to the formula:
Figure SMS_174
/>
obtaining the current iteration times
Figure SMS_177
Adding 1 to the iteration times and returning to the step S5-1; wherein (1)>
Figure SMS_179
For the current iteration number>
Figure SMS_182
For the upper limit of the iteration number, +.>
Figure SMS_176
For the initialized global constraint violation mean value of all non-feasible solutions in the ant colony->
Figure SMS_180
For the proportion of feasible solutions in the current ant colony, < >>
Figure SMS_183
and />
Figure SMS_184
Are super parameters; when->
Figure SMS_175
In the time-course of which the first and second contact surfaces,εthe value of (2) decreases exponentially with increasing iteration number, when +.>
Figure SMS_178
In the time-course of which the first and second contact surfaces,εthe value of (2) is adjusted to approximately the initial value +.>
Figure SMS_181
The proportion of the non-feasible solution is too low, so that the acceptance degree of the non-feasible solution is properly improved; the initial value of the iteration number is 0.
The beneficial effects of the invention are as follows: the invention comprehensively considers the RSU service range pairsEstablishing a system loss model under the constraint of unloading decision, and adding a model based on a mixed variable ant colony algorithmεThe constraint processing technology of the system has smaller system loss and high spectrum utilization rate, and effectively improves the task unloading efficiency of the user.
Drawings
FIG. 1 is a flow chart of the present invention;
FIG. 2 is a block diagram of a system according to the present invention;
FIG. 3 is a graph comparing the effects of user speed on system loss for different approaches;
FIG. 4 is a graph comparing the effect of the number of users on the system loss for different methods.
Detailed Description
The following description of the embodiments of the present invention is provided to facilitate understanding of the present invention by those skilled in the art, but it should be understood that the present invention is not limited to the scope of the embodiments, and all the inventions which make use of the inventive concept are protected by the spirit and scope of the present invention as defined and defined in the appended claims to those skilled in the art.
As shown in fig. 1, a task unloading method for internet of vehicles includes the following steps:
s1, acquiring user data and constructing a system loss model;
s2, initializing parameters of a system loss model, and calculating system loss and overall constraint violation degree of an initial solution of the system loss model;
s3, sorting initial solutions of the system loss model according to the system loss, and initializing weights of the initial solutions of the system loss model; when the global constraint violation of the non-feasible solution is at the thresholdεWhen the solution is within, the solution is regarded as a feasible solution to be subjected to sorting treatment;
s4, solving a system loss model through an ant colony algorithm to obtain a sub-channel comprising a user task unloading proportion, user transmission power and user usageMNew solutions;
s5, sorting the new solutions and the initial solutions of the system loss model to obtain the current optimal feasible solution, and repeating the step S3 to the step until the preset times to obtain the optimal feasible solution which minimizes the system loss;
and S6, taking the optimal feasible solution which minimizes the system loss as an Internet of vehicles task unloading strategy.
Further, the specific implementation manner of step S1 is as follows:
constructing a system loss model:
Figure SMS_185
/>
Figure SMS_186
Figure SMS_187
Figure SMS_188
Figure SMS_189
Figure SMS_190
Figure SMS_191
wherein ,
Figure SMS_218
representing minimum system loss; />
Figure SMS_223
Representing a system loss function; />
Figure SMS_228
A solution representing a system loss function; />
Figure SMS_196
Representing user +.>
Figure SMS_202
Task unloading time; />
Figure SMS_208
Representing user +.>
Figure SMS_214
The residence time within the RSU service range; the length of the RSU service section is +.>
Figure SMS_217
User->
Figure SMS_222
Is +.>
Figure SMS_227
，/>
Figure SMS_231
The position of (2) is +.>
Figure SMS_219
；/>
Figure SMS_224
Representing sharing ofUA user;CHrepresenting sharing ofCHSubchannel, & gt>
Figure SMS_229
；/>
Figure SMS_232
Representing user +.>
Figure SMS_210
Task offload ratio of (2); />
Figure SMS_216
Representing user +.>
Figure SMS_221
Is used for the transmission power of the (a); />
Figure SMS_226
Representing user +.>
Figure SMS_192
The sub-channels used; />
Figure SMS_198
Representing the user's preference between latency and energy consumption; />
Figure SMS_204
Representing user +.>
Figure SMS_211
The computation delay of all tasks executed locally; />
Figure SMS_197
Representing user +.>
Figure SMS_203
The energy consumed by all tasks performed locally; />
Figure SMS_209
Representing user +.>
Figure SMS_215
Total energy consumption required by a task; />
Figure SMS_220
For user->
Figure SMS_225
Actual computation delay of task,/->
Figure SMS_230
For user->
Figure SMS_233
Local calculation time delay of the task; />
Figure SMS_194
Representing user +.>
Figure SMS_199
In subchannel->
Figure SMS_205
Channel gain on; />
Figure SMS_212
Is->
Figure SMS_195
At->
Figure SMS_201
Power gain on; />
Figure SMS_207
Is the path loss coefficient; />
Figure SMS_213
Representing a transmission power threshold; />
Figure SMS_193
Representing user +.>
Figure SMS_200
In subchannel->
Figure SMS_206
Channel gain on the upper.
In step S1, the user
Figure SMS_234
Total energy consumption required for a task, user +.>
Figure SMS_235
The specific implementation mode of the task unloading time is as follows:
according to the formula:
Figure SMS_236
Figure SMS_237
obtaining local calculation energy consumption
Figure SMS_240
And user->
Figure SMS_243
Local computation delay of task->
Figure SMS_246
； wherein ,/>
Figure SMS_239
For user->
Figure SMS_242
CPU average power,/,>
Figure SMS_245
is->
Figure SMS_248
CPU power consumption coefficient of the same; />
Figure SMS_238
For user->
Figure SMS_241
The working frequency of the local CPU; />
Figure SMS_244
Representing the size of the task data; />
Figure SMS_247
Representing the number of CPU cycles required to calculate each bit of data;
according to the formula:
Figure SMS_250
get user received by RSU->
Figure SMS_254
Signal-to-interference-plus-noise ratio of signal>
Figure SMS_258
The method comprises the steps of carrying out a first treatment on the surface of the The number of users in the set isS；/>
Figure SMS_251
Indicating all used subchannels>
Figure SMS_255
Is a set of users; />
Figure SMS_259
For user->
Figure SMS_261
Distance between RSU and->
Figure SMS_249
Distance between RSU and service section; />
Figure SMS_253
Representing user +.>
Figure SMS_257
Is used for the transmission power of the (a); />
Figure SMS_260
Maximum transmission power for the user; />
Figure SMS_252
Is Gaussian white noise power spectral density; />
Figure SMS_256
The bandwidth allocated for each sub-channel for the RSU;
according to the formula:
Figure SMS_262
obtaining the user
Figure SMS_263
Transmission rate of->
Figure SMS_264
；
According to the formula:
Figure SMS_265
Figure SMS_266
obtaining the user
Figure SMS_267
Transmission delay during task offloading>
Figure SMS_268
And transmission energy consumption->
Figure SMS_269
；
According to the formula:
Figure SMS_270
obtaining the calculation time delay required by the RSU to execute the task
Figure SMS_271
； wherein ,/>
Figure SMS_272
For RSU server for user +.>
Figure SMS_273
CPU working frequency allocated by the unloaded task;
according to the formula:
Figure SMS_274
Figure SMS_275
obtaining the user
Figure SMS_276
Task offloading latency->
Figure SMS_277
And user->
Figure SMS_278
Total energy consumption required for a task->
Figure SMS_279
。
The specific implementation manner of the step S2 is as follows:
s2-1, initializing the number of initial solutions in the ant colonyKInitializing weights of an initial solution
Figure SMS_280
Number of antsMUpper limit of iteration numberT，Overall constraint violation mean of all non-viable solutions in ant colonyε；
S2-2, according to the formula:
Figure SMS_281
obtaining an overall constraint violation function
Figure SMS_282
； wherein ,/>
Figure SMS_283
A violation degree function for time constraint;
Figure SMS_284
a violation degree function for the power constraint; when->
Figure SMS_285
When (I)>
Figure SMS_286
For the system loss function->
Figure SMS_287
Or else, a non-viable solution;
s2-3, calculating the system loss of all solutions according to the system loss function, and calculating the overall constraint violation of all solutions according to the overall constraint violation function.
The specific implementation manner of the step S3 is as follows:
s3-1, according to the formula:
Figure SMS_288
based on the possible and non-possible solutions together in an ant colonyεIs a comparison ordering of (2);
Figure SMS_291
indicate->
Figure SMS_293
System loss of individual solutions,/->
Figure SMS_295
Indicate->
Figure SMS_290
System loss of the solution; />
Figure SMS_292
Indicate->
Figure SMS_294
A solution power constraint violation degree; />
Figure SMS_296
Indicate->
Figure SMS_289
Power constraint violation for individual solutionsA degree;
s3-2, according to the formula:
Figure SMS_297
Figure SMS_298
obtaining a subchannel
Figure SMS_314
Weight of +.>
Figure SMS_302
And initial solution->
Figure SMS_309
Weight of +.>
Figure SMS_311
； wherein ,/>
Figure SMS_315
Is a super parameter; />
Figure SMS_312
The more preferred is that,
Figure SMS_316
the greater the +.>
Figure SMS_304
The thicker the pheromone is left, the ant is +.>
Figure SMS_308
Is->
Figure SMS_299
and />
Figure SMS_305
The greater the probability of nearby movement;
Figure SMS_303
for users in ant colony->
Figure SMS_310
Sub-channel +.>
Figure SMS_307
Weights of the optimal solution of ∈1->
Figure SMS_313
For users in ant colony->
Figure SMS_300
Sub-channel +.>
Figure SMS_306
The number of solutions of>
Figure SMS_301
Indicating the number of sub-channels that are not used by the users in all solutions in the ant colony.
The specific implementation manner of the step S4 is as follows:
s4-1, according to the formula:
Figure SMS_317
obtaining a random solution
Figure SMS_318
；/>
Figure SMS_319
Is->
Figure SMS_320
Weights of (2);crepresent the firstcPerforming solution;
s4-2, according to the formula:
Figure SMS_321
get ants to move to
Figure SMS_322
Transition probability of any point in continuous interval +.>
Figure SMS_323
；/>
Figure SMS_324
Solution selected for ants->
Figure SMS_325
Middle user->
Figure SMS_326
Task offload ratio of (2); />
Figure SMS_327
，/>
Figure SMS_328
Is a super parameter;
s4-3, according to the transition probability
Figure SMS_329
Randomly move to interval +.>
Figure SMS_330
Any point in the above, get the user in the new solution
Figure SMS_331
Task offloading ratio of->
Figure SMS_332
；
S4-4, according to the formula:
Figure SMS_333
obtaining user power transition probability
Figure SMS_334
； wherein ,/>
Figure SMS_335
，/>
Figure SMS_336
；
S4-5, according to the transition probability
Figure SMS_337
In the continuous interval +.>
Figure SMS_338
Randomly moving to any point to obtain user +.>
Figure SMS_339
Is +.>
Figure SMS_340
；
S4-6, according to the formula:
Figure SMS_341
obtaining sub-channel transition probability of user
Figure SMS_342
； wherein ,/>
Figure SMS_343
Is sub-channel->
Figure SMS_344
Weights of (2);
s4-7, according to the sub-channel transition probability of the user
Figure SMS_345
Move to discrete interval +.>
Figure SMS_346
Inner one point->
Figure SMS_347
Obtaining the user's->
Figure SMS_348
Subchannel number->
Figure SMS_349
；
S4-8, according to the users in the new solution
Figure SMS_351
Subchannel number->
Figure SMS_353
User->
Figure SMS_355
Is +.>
Figure SMS_352
And user ∈>
Figure SMS_354
Task offloading ratio of->
Figure SMS_356
Obtaining ant->
Figure SMS_357
A new solution of 3 XU times in the value space of each variable>
Figure SMS_350
；
S4-9, repeating the steps S4-1 to S4-8 untilMNext, obtainMNew solutions are provided.
The specific implementation manner of the step S5 is as follows:
s5-1, to be obtainedMNew solutions and current existing solutionsKThe solutions are subjected to epsilon-based comparison and sorting, and used beforeKThe original solution in the ant colony is replaced by the optimal solution to obtain a replaced solution;
s5-2, judging whether the current iteration reaches the set times, if so, outputting an optimal feasible solution which minimizes the system loss; otherwise, enter step S5-3;
s5-3, calculating the pheromone of the replaced solution, namely the weight of the replaced solution;
s5-4, according to the formula:
Figure SMS_358
obtaining the current iteration times
Figure SMS_360
Adding 1 to the iteration times and returning to the step S5-1; wherein (1)>
Figure SMS_363
For the current iteration number>
Figure SMS_366
For the upper limit of the iteration number, +.>
Figure SMS_361
For the initialized global constraint violation mean value of all non-feasible solutions in the ant colony->
Figure SMS_364
For the proportion of feasible solutions in the current ant colony, < >>
Figure SMS_367
and />
Figure SMS_368
Are super parameters; when->
Figure SMS_359
In the time-course of which the first and second contact surfaces,εthe value of (2) decreases exponentially with increasing iteration number, when +.>
Figure SMS_362
In the time-course of which the first and second contact surfaces,εthe value of (2) is adjusted to approximately the initial value +.>
Figure SMS_365
The proportion of the non-feasible solution is too low, so that the acceptance degree of the non-feasible solution is properly improved; the initial value of the iteration number is 0.
As shown in fig. 2, NOMA-basedA single RSU in a car networking environment serves multiple users, with the RSU being equipped with edge servers to assist in computing the tasks offloaded by the users. NOMA allows each sub-channel to be used by a plurality of users at the same time, so that the task unloading efficiency of the users can be greatly improved; each user generates a plurality of tasks which are not decomposed and are not associated with each other, and the network topology structure and the channel gain between the RSU and the user are kept unchanged from the beginning of task unloading of the user to the end of returning the calculation result to the user by the RSU. RSU hasCHAnd the available mutually orthogonal sub-channels. Within the service range of RSUUThe number of users is far greater than the number of sub-channels, i.e
Figure SMS_369
。
As shown in fig. 3 and fig. 4, as the number of users increases, the number of computing resources allocated to each task by the RSU server decreases, increasing the computation delay for offloading tasks, and at the same time, as the inter-channel interference increases, the transmission energy consumption and the delay increase, which results in a corresponding increase in system loss of all algorithms. It can also be seen that the performance of the present invention is similar to NOMA-B and OFDMA when the number of users is small, whereas the system loss of the present invention is lower than other algorithms when the number of users increases. When (when)ＵWhen=29, the system loss of the present invention is 21.14%, 17.33% and 6.79% lower than AL, OFDMA and NOMA-B, respectively. When the upper limit of the transmission power of the user is changed between 17dBm and 31dBm, the system loss of the invention is lower than that of other algorithms. When the upper limit of the user transmission power is 31dBm, the system loss of the present invention is 38.00%, 15.74% and 8.54% lower than AL, OFDMA and NOMA-B, respectively.
In one embodiment of the present invention, the present invention is compared with the system losses of AL, NOMA-B and OFDMA under the same parameters, and the simulation environment parameters and the policy-related parameter results are shown in tables 1 and 2.
TABLE 1
Figure SMS_370
TABLE 2
Figure SMS_371
As can be seen from tables 1 and 2, the system loss of the invention is lower than other strategies at different speeds, when the user speed is gradually increased from 10m/s to 40m/s, the residence time of the user in the service range of the RSU is reduced, so that the user can reduce the task proportion unloaded to the RSU in order to ensure that the RSU can return the calculation result to the requesting vehicle within the residence time of the RSU, and the local calculation energy consumption is increased. The system losses of the present invention are reduced by 34.62%, 17.55% and 8.23% compared to AL, OFDMA and NOMA-B, respectively, when the user speed reaches 40m/s, and by 38.05%, 14.28% and 7.08% when the user speed is 10m/s, respectively.
The invention comprehensively considers the constraint of the RSU service range on the unloading decision, establishes a system loss model, and adds a load-based algorithm based on the mixed variable ant colony algorithmεThe constraint processing technology of the system has smaller system loss and high spectrum utilization rate, and effectively improves the task unloading efficiency of the user.

Claims (1)

1. The task unloading method for the Internet of vehicles is characterized by comprising the following steps of:
s1, acquiring user data and constructing a system loss model;
s2, initializing parameters of a system loss model, and calculating system loss and overall constraint violation degree of an initial solution of the system loss model;
s3, sorting initial solutions of the system loss model according to the system loss, and initializing weights of the initial solutions of the system loss model; when the global constraint violation of the non-feasible solution is at the thresholdεWhen the solution is within, the solution is regarded as a feasible solution to be subjected to sorting treatment;
s4, solving a system loss model through an ant colony algorithm to obtain a sub-channel comprising a user task unloading proportion, user transmission power and user usageMNew solutions;
s5, sorting the new solutions and the initial solutions of the system loss model to obtain the current optimal feasible solution, and repeating the step S3 to the step until the preset times to obtain the optimal feasible solution which minimizes the system loss;
s6, taking the optimal feasible solution which minimizes the system loss as an Internet of vehicles task unloading strategy;
the specific implementation manner of the step S1 is as follows:
constructing a system loss model:
Figure QLYQS_1
Figure QLYQS_2
Figure QLYQS_3
Figure QLYQS_4
Figure QLYQS_5
Figure QLYQS_6
Figure QLYQS_7
wherein ,
Figure QLYQS_24
representing minimum system loss; />
Figure QLYQS_28
Representing a system loss function; />
Figure QLYQS_32
A solution representing a system loss function; />
Figure QLYQS_11
Representing user +.>
Figure QLYQS_12
Task unloading time; />
Figure QLYQS_16
Representing user +.>
Figure QLYQS_20
The residence time in the RSU service area, the length of the RSU service section is +.>
Figure QLYQS_36
User->
Figure QLYQS_40
Is +.>
Figure QLYQS_43
，/>
Figure QLYQS_46
The position of (2) is +.>
Figure QLYQS_41
；/>
Figure QLYQS_44
Representing sharing ofUA user;CHrepresenting sharing ofCHSubchannel, & gt>
Figure QLYQS_47
；/>
Figure QLYQS_49
Representing user +.>
Figure QLYQS_27
Task offload ratio of (2); />
Figure QLYQS_31
Representing a user
Figure QLYQS_35
Is used for the transmission power of the (a); />
Figure QLYQS_39
Representing user +.>
Figure QLYQS_8
The sub-channels used; />
Figure QLYQS_15
Representing the user's preference between latency and energy consumption; />
Figure QLYQS_19
Representing user +.>
Figure QLYQS_23
The computation delay of all tasks executed locally; />
Figure QLYQS_9
Representing user +.>
Figure QLYQS_13
The energy consumed by all tasks performed locally; />
Figure QLYQS_17
Representing user +.>
Figure QLYQS_21
Total energy consumption required by a task; />
Figure QLYQS_25
For user->
Figure QLYQS_29
Actual computation delay of task,/->
Figure QLYQS_33
For user->
Figure QLYQS_37
Local calculation time delay of the task; />
Figure QLYQS_22
Representing user +.>
Figure QLYQS_26
In subchannel->
Figure QLYQS_30
Channel gain on; />
Figure QLYQS_34
Is->
Figure QLYQS_38
At->
Figure QLYQS_42
Power gain on; />
Figure QLYQS_45
Is the path loss coefficient; />
Figure QLYQS_48
Representing user +.>
Figure QLYQS_10
In subchannel->
Figure QLYQS_14
Channel gain on; />
Figure QLYQS_18
Representing a transmission power threshold;
in step S1, the user
Figure QLYQS_50
Total energy consumption required for a task, user +.>
Figure QLYQS_51
The specific implementation mode of the task unloading time is as follows:
according to the formula:
Figure QLYQS_52
Figure QLYQS_53
obtaining local calculation energy consumption
Figure QLYQS_55
And user->
Figure QLYQS_58
Local computation delay of task->
Figure QLYQS_61
； wherein ,/>
Figure QLYQS_56
For users
Figure QLYQS_59
CPU average power,/,>
Figure QLYQS_62
is->
Figure QLYQS_64
CPU power consumption coefficient of the same; />
Figure QLYQS_54
For user->
Figure QLYQS_57
The working frequency of the local CPU; />
Figure QLYQS_60
Representing the size of the task data; />
Figure QLYQS_63
Representing the number of CPU cycles required to calculate each bit of data;
according to the formula:
Figure QLYQS_65
obtaining the user received by the RSU
Figure QLYQS_67
Signal-to-interference-plus-noise ratio of signal>
Figure QLYQS_70
The method comprises the steps of carrying out a first treatment on the surface of the The number of users in the set is S; />
Figure QLYQS_73
Indicating all used subchannels>
Figure QLYQS_69
Is a set of users; />
Figure QLYQS_72
For user->
Figure QLYQS_75
And the distance between the RSUs,
Figure QLYQS_77
distance between RSU and service section; />
Figure QLYQS_66
Representing user +.>
Figure QLYQS_71
Is used for the transmission power of the (a); />
Figure QLYQS_74
Maximum transmission power for the user; />
Figure QLYQS_76
Is Gaussian white noise power spectral density; />
Figure QLYQS_68
The bandwidth allocated for each sub-channel for the RSU;
according to the formula:
Figure QLYQS_78
obtaining the user
Figure QLYQS_79
Transmission rate of->
Figure QLYQS_80
；
According to the formula:
Figure QLYQS_81
Figure QLYQS_82
obtaining the user
Figure QLYQS_83
Transmission delay during task offloading>
Figure QLYQS_84
And transmission energy consumption->
Figure QLYQS_85
；
According to the formula:
Figure QLYQS_86
/>
obtaining the calculation time delay required by the RSU to execute the task
Figure QLYQS_87
； wherein ,/>
Figure QLYQS_88
For RSU server for user +.>
Figure QLYQS_89
CPU working frequency allocated by the unloaded task;
according to the formula:
Figure QLYQS_90
Figure QLYQS_91
obtaining the user
Figure QLYQS_92
Task offloading latency->
Figure QLYQS_93
And user->
Figure QLYQS_94
Total energy consumption required for a task->
Figure QLYQS_95
；
The specific implementation manner of the step S2 is as follows:
s2-1, initializing the number of initial solutions in the ant colonyKInitializing weights of an initial solution
Figure QLYQS_96
Number of antsMUpper limit of iteration numberT，Overall constraint violation mean of all non-viable solutions in ant colonyε；
S2-2, according to the formula:
Figure QLYQS_97
obtaining an overall constraint violation function
Figure QLYQS_98
； wherein ,/>
Figure QLYQS_99
A violation degree function for time constraint;
Figure QLYQS_100
a violation degree function for the power constraint; when->
Figure QLYQS_101
When (I)>
Figure QLYQS_102
For the system loss function->
Figure QLYQS_103
Or else, a non-viable solution;
s2-3, calculating the system loss of all solutions according to the system loss function, and calculating the overall constraint violation of all solutions according to the overall constraint violation function;
the specific implementation manner of the step S3 is as follows:
s3-1, according to the formula:
Figure QLYQS_104
based on the possible and non-possible solutions together in an ant colonyεIs a comparison ordering of (2);
Figure QLYQS_107
indicate->
Figure QLYQS_108
System loss of individual solutions,/->
Figure QLYQS_110
Indicate->
Figure QLYQS_106
System loss of the solution; />
Figure QLYQS_109
Indicate->
Figure QLYQS_111
A solution power constraint violation degree;
Figure QLYQS_112
indicate->
Figure QLYQS_105
Power constraint violation degrees of the individual solutions;
s3-2, according to the formula:
Figure QLYQS_113
/>
Figure QLYQS_114
obtaining a subchannel
Figure QLYQS_126
Weight of +.>
Figure QLYQS_116
And initial solution->
Figure QLYQS_122
Weight of +.>
Figure QLYQS_123
； wherein ,/>
Figure QLYQS_127
Is a super parameter; />
Figure QLYQS_130
The more optimal (i.e. ->
Figure QLYQS_132
The greater the +.>
Figure QLYQS_124
The thicker the pheromone is left, the ant is +.>
Figure QLYQS_128
Is->
Figure QLYQS_115
and />
Figure QLYQS_120
The greater the probability of nearby movement; />
Figure QLYQS_121
For users in ant colony->
Figure QLYQS_125
Sub-channel +.>
Figure QLYQS_129
Weights of the optimal solution of ∈1->
Figure QLYQS_131
For users in ant colony->
Figure QLYQS_117
Sub-channels are used
Figure QLYQS_119
The number of solutions of>
Figure QLYQS_118
Representing the number of sub-channels that are not used by the users in all solutions in the ant colony;
the specific implementation manner of the step S4 is as follows:
s4-1, according to the formula:
Figure QLYQS_133
obtaining a random solution
Figure QLYQS_134
；/>
Figure QLYQS_135
Is->
Figure QLYQS_136
Weights of (2);crepresent the firstcPerforming solution;
s4-2, according to the formula:
Figure QLYQS_137
get ants to move to
Figure QLYQS_138
Continuous zoneProbability of transition at any point in the space +.>
Figure QLYQS_139
；/>
Figure QLYQS_140
Solution selected for ants->
Figure QLYQS_141
Middle user->
Figure QLYQS_142
Task offload ratio of (2); />
Figure QLYQS_143
，/>
Figure QLYQS_144
Is a super parameter;
s4-3, according to the transition probability
Figure QLYQS_145
Randomly move to interval +.>
Figure QLYQS_146
At any point above, get the user +.>
Figure QLYQS_147
Task offloading ratio of->
Figure QLYQS_148
；
S4-4, according to the formula:
Figure QLYQS_149
obtaining user power transition probability
Figure QLYQS_150
； wherein ,/>
Figure QLYQS_151
，/>
Figure QLYQS_152
；
S4-5, according to the transition probability
Figure QLYQS_153
In the continuous interval +.>
Figure QLYQS_154
Randomly moving to any point to obtain user +.>
Figure QLYQS_155
Is +.>
Figure QLYQS_156
；
S4-6, according to the formula:
Figure QLYQS_158
/>
obtaining sub-channel transition probability of user
Figure QLYQS_159
； wherein ,/>
Figure QLYQS_160
Is sub-channel->
Figure QLYQS_161
Weights of (2);
s4-7, according to the sub-channel transition probability of the user
Figure QLYQS_162
Move to discrete interval +.>
Figure QLYQS_163
Inner one point->
Figure QLYQS_164
Obtaining the user's->
Figure QLYQS_165
Subchannel number->
Figure QLYQS_166
；
S4-8, according to the users in the new solution
Figure QLYQS_169
Subchannel number->
Figure QLYQS_170
User->
Figure QLYQS_172
Is +.>
Figure QLYQS_168
And user ∈>
Figure QLYQS_171
Task offloading ratio of->
Figure QLYQS_173
Obtaining ant->
Figure QLYQS_174
A new solution of 3 XU times in the value space of each variable>
Figure QLYQS_167
；
S4-9, repeating the steps S4-1 to S4-8 untilMNext, obtainMNew solutions are obtained;
the specific implementation manner of the step S5 is as follows:
s5-1, to be obtainedMNew solutions and current existing solutionsKThe solutions are subjected to epsilon-based comparison and sorting, and used beforeKThe original solution in the ant colony is replaced by the optimal solution to obtain a replaced solution;
s5-2, judging whether the current iteration reaches the set times, if so, outputting an optimal feasible solution which minimizes the system loss; otherwise, enter step S5-3;
s5-3, calculating the pheromone of the replaced solution, namely the weight of the replaced solution;
s5-4, according to the formula:
Figure QLYQS_175
obtaining the current iteration times
Figure QLYQS_176
Adding 1 to the iteration times and returning to the step S5-1; wherein (1)>
Figure QLYQS_177
For the current number of iterations,
Figure QLYQS_178
for the upper limit of the iteration number, +.>
Figure QLYQS_179
For the initialized global constraint violation mean value of all non-feasible solutions in the ant colony->
Figure QLYQS_180
For the proportion of feasible solutions in the current ant colony, < >>
Figure QLYQS_181
and />
Figure QLYQS_182
Are super parameters; the initial value of the iteration number is 0./>
CN202310194757.8A 2023-03-03 2023-03-03 Task unloading method for Internet of vehicles Active CN115915278B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310194757.8A CN115915278B (en) 2023-03-03 2023-03-03 Task unloading method for Internet of vehicles

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310194757.8A CN115915278B (en) 2023-03-03 2023-03-03 Task unloading method for Internet of vehicles

Publications (2)

Publication Number Publication Date
CN115915278A CN115915278A (en) 2023-04-04
CN115915278B true CN115915278B (en) 2023-05-05

Family

ID=85751883

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310194757.8A Active CN115915278B (en) 2023-03-03 2023-03-03 Task unloading method for Internet of vehicles

Country Status (1)

Country Link
CN (1) CN115915278B (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111741050A (en) * 2020-04-07 2020-10-02 北京邮电大学 Data distribution method and system based on roadside unit
CN112367640A (en) * 2020-11-09 2021-02-12 中科怡海高新技术发展江苏股份公司 V2V mode multitask unloading method and system based on mobile edge calculation
CN114827191A (en) * 2022-03-15 2022-07-29 华南理工大学 Dynamic task unloading method for fusing NOMA in vehicle-road cooperative system
EP4037352A1 (en) * 2019-09-27 2022-08-03 Tencent Technology (Shenzhen) Company Limited Access processing method for user equipment, device, electronic apparatus, and computer storage medium

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP4037352A1 (en) * 2019-09-27 2022-08-03 Tencent Technology (Shenzhen) Company Limited Access processing method for user equipment, device, electronic apparatus, and computer storage medium
CN111741050A (en) * 2020-04-07 2020-10-02 北京邮电大学 Data distribution method and system based on roadside unit
CN112367640A (en) * 2020-11-09 2021-02-12 中科怡海高新技术发展江苏股份公司 V2V mode multitask unloading method and system based on mobile edge calculation
CN114827191A (en) * 2022-03-15 2022-07-29 华南理工大学 Dynamic task unloading method for fusing NOMA in vehicle-road cooperative system

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
5G网络下移动云计算节能措施研究;李继蕊;李小勇;高云全;高雅丽;;计算机学报(07);全文 *
自供电路侧单元能量-时延均衡分组调度策略;代亮;翟一鸣;汪贵平;;交通运输工程学报(02);全文 *

Also Published As

Publication number Publication date
CN115915278A (en) 2023-04-04

Similar Documents

Publication Publication Date Title
Liu et al. A computation offloading algorithm based on game theory for vehicular edge networks
CN108809695B (en) Distributed uplink unloading strategy facing mobile edge calculation
Qin et al. Power-constrained edge computing with maximum processing capacity for IoT networks
CN109767117B (en) Power distribution method for joint task scheduling in mobile edge computing
CN110233755B (en) Computing resource and frequency spectrum resource allocation method for fog computing in Internet of things
CN107995660A (en) Support Joint Task scheduling and the resource allocation methods of D2D- Edge Servers unloading
Hou et al. Joint allocation of wireless resource and computing capability in MEC-enabled vehicular network
CN111124531B (en) Method for dynamically unloading calculation tasks based on energy consumption and delay balance in vehicle fog calculation
Saleem et al. Performance guaranteed partial offloading for mobile edge computing
CN110149401B (en) Method and system for optimizing edge calculation task
Wei et al. Optimal offloading in fog computing systems with non-orthogonal multiple access
Liu et al. Fog computing vehicular network resource management based on chemical reaction optimization
Wang et al. Multi-agent reinforcement learning-based user pairing in multi-carrier NOMA systems
CN113286329A (en) Communication and computing resource joint optimization method based on mobile edge computing
Zu et al. Smeto: Stable matching for energy-minimized task offloading in cloud-fog networks
Trrad et al. Application of fuzzy logic to cognitive wireless communications
Nur et al. Priority-based offloading and caching in mobile edge cloud
Zhang et al. Energy minimization task offloading mechanism with edge-cloud collaboration in IoT networks
CN115915278B (en) Task unloading method for Internet of vehicles
Fan et al. Task caching and computation offloading for muti-user mobile edge computing network
CN114928611B (en) IEEE802.11p protocol-based energy-saving calculation unloading optimization method for Internet of vehicles
Jothi et al. Nelder mead-based spider monkey optimization for optimal power and channel allocation in MANET
Jin et al. Computation offloading and resource allocation for MEC in C-RAN: A deep reinforcement learning approach
Wang et al. Task and resource allocation in mobile edge computing: An improved reinforcement learning approach
Malik et al. A user-centric qos-aware multi-path service provisioning in mobile edge computing

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