CN113742077B - Computing migration method based on 5G Internet of vehicles - Google Patents
Computing migration method based on 5G Internet of vehicles Download PDFInfo
- Publication number
- CN113742077B CN113742077B CN202111048558.3A CN202111048558A CN113742077B CN 113742077 B CN113742077 B CN 113742077B CN 202111048558 A CN202111048558 A CN 202111048558A CN 113742077 B CN113742077 B CN 113742077B
- Authority
- CN
- China
- Prior art keywords
- vehicle
- calculation
- server
- population
- individuals
- 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
- 230000005012 migration Effects 0.000 title claims abstract description 73
- 238000013508 migration Methods 0.000 title claims abstract description 73
- 238000000034 method Methods 0.000 title claims abstract description 54
- 238000004364 calculation method Methods 0.000 claims abstract description 108
- 238000005265 energy consumption Methods 0.000 claims abstract description 36
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 18
- 238000013178 mathematical model Methods 0.000 claims abstract description 18
- 238000005516 engineering process Methods 0.000 claims abstract description 8
- 230000005540 biological transmission Effects 0.000 claims description 24
- 230000000739 chaotic effect Effects 0.000 claims description 15
- 210000000349 chromosome Anatomy 0.000 claims description 6
- 230000006870 function Effects 0.000 claims description 6
- 235000006679 Mentha X verticillata Nutrition 0.000 claims description 3
- 235000002899 Mentha suaveolens Nutrition 0.000 claims description 3
- 235000001636 Mentha x rotundifolia Nutrition 0.000 claims description 3
- 230000015572 biosynthetic process Effects 0.000 claims description 3
- 230000035772 mutation Effects 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 3
- 108090000623 proteins and genes Proteins 0.000 claims description 3
- 238000006467 substitution reaction Methods 0.000 claims description 3
- 238000003786 synthesis reaction Methods 0.000 claims description 3
- 230000006978 adaptation Effects 0.000 claims 1
- 230000009286 beneficial effect Effects 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 8
- 238000004088 simulation Methods 0.000 description 6
- 238000004891 communication Methods 0.000 description 3
- 230000006855 networking Effects 0.000 description 3
- 230000001934 delay Effects 0.000 description 1
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
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5061—Partitioning or combining of resources
- G06F9/5072—Grid computing
-
- 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
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5005—Allocation of resources, e.g. of the central processing unit [CPU] to service a request
- G06F9/5027—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resource being a machine, e.g. CPUs, Servers, Terminals
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a calculation migration method based on 5G Internet of vehicles, which comprises the following steps: the first step, building a calculation migration scene: selecting a 5G-based Internet of vehicles scene as a calculation migration scene, wherein the calculation migration scene mainly comprises an EC server, an RSU, a vehicle and a cloud server, and the EC server, the RSU, the vehicle and the cloud server are communicated by using a 5G technology; step two, obtaining a calculation migration mathematical model according to the calculation migration scene; thirdly, obtaining an optimal solution of the mathematical model based on an improved chaos-differential evolution algorithm; and fourthly, taking a calculation migration method corresponding to the optimal solution of the mathematical model as a final calculation migration method. The beneficial effects are that: the method and the device can break the resource limitation of the vehicle, reduce the calculated amount of the vehicle, reduce the loss of the battery capacity of the vehicle, save the storage resources of the vehicle and the like, and can also improve the user experience. Through the improved chaos-differential evolution algorithm, optimal time delay and energy consumption can be obtained, and the requirements of time delay and energy consumption in the environment of the Internet of vehicles can be met.
Description
Technical Field
The invention relates to a calculation migration method, in particular to a calculation migration method based on a 5G Internet of vehicles.
Background
Currently, as the internet of vehicles and autopilot applications are receiving attention, more communication technology (Communication Technology, CT) enterprises sequentially consider the internet of vehicles as an important application direction of Edge Computing, and in addition, information technology (Information Technology, IT) enterprises start to enter the internet of vehicles field based on Edge Computing (EC).
The advent of the 5G age has resulted in massive amounts of data and connections. In the scene of the internet of vehicles, because the computing capacity and storage resources of the vehicles are limited, some computing tasks with large computing capacity and high safety cannot be independently realized in the vehicle-mounted units, and the computing tasks which cannot be processed by the vehicles need to be considered to be migrated to other devices with idle resources.
Disclosure of Invention
The invention aims to solve the problem that a vehicle-mounted terminal resource is limited under an edge computing environment and a task with large data volume cannot be processed in real time, and provides a computing migration method based on a 5G vehicle networking.
The invention provides a 5G car networking-based calculation migration method, which comprises the following steps:
the first step, building a calculation migration scene: selecting a 5G-based Internet of vehicles scene as a calculation migration scene, wherein the calculation migration scene mainly comprises an EC server, an RSU, a vehicle and a cloud server, and the EC server, the RSU, the vehicle and the cloud server are communicated by using a 5G technology;
step two, obtaining a calculation migration mathematical model according to the calculation migration scene;
thirdly, obtaining an optimal solution of the mathematical model based on an improved chaos-differential evolution algorithm;
and fourthly, taking a calculation migration method corresponding to the optimal solution of the mathematical model as a final calculation migration method.
In the second step, the specific steps for obtaining the computational migration mathematical model according to the computational migration scene are as follows:
according to the calculation migration scene, the problem of reasonable distribution of calculation tasks generated by N EC servers and a single vehicle to the vehicle is researched, and the assumption is made that the single vehicle has one calculation task K= { K 1 ,k 2 ,...k i ,...,k M Required processing, subtask k i Is k i =(a i ,b i ,c i ) The vehicle computing tasks K are processed by the vehicle terminal or the EC server, and P computing tasks K are selected i Processing M-P computing tasks k at vehicle-mounted terminal i Migration to EC Server processing, each subtask k i One of the EC servers, either local or N, must be selected for processing as follows:
wherein a is i Indicating completion of task k i The required cycle number, b i Representing the size of the uploaded data, c i Represents the size of the returned data, M, P represents the number of subtasks, and M > P, x i Representing a calculation migration decision, determining a calculation task k i Whether processing is performed locally or on the EC server, y i,n Indicating whether task k is to be calculated i Assigned to EC server n;
x i ={0,1},x i when=1, the calculation task k is represented i Processing locally, x i When=0, it indicates that the computing task is in EC serviceOn-board processing, y i,n ={0,1},y i,n When=1, the calculation task k is represented i Assigned to EC servers n, y i,n When=0, the calculation task k is represented i Not assigned to EC server n;
step two, time delay and energy consumption generated by processing a vehicle computing task in a vehicle-mounted terminal;
if some or all of the subtasks in the vehicle computing task are processed locally, the computing latency is expressed as follows:
wherein T is V Representing time delay generated by processing vehicle computing task in vehicle-mounted terminal, f V Representing the computing power of the vehicle V, i.e. the number of CPU cycles per unit time;
if some or all of the subtasks in the vehicle calculation task are processed in the vehicle-mounted terminal, the calculation energy consumption is expressed as follows:
wherein E is V Calculation energy consumption P generated by processing vehicle calculation task in vehicle-mounted terminal V The energy consumption power is realized in a unit CPU period of the vehicle;
step three, processing generated time delay and energy consumption of the vehicle computing task on the EC server:
if some or all of the subtasks in the vehicle computing task need to be migrated to the EC server for processing, the generated delay includes an uploading delay, a processing delay and a backhaul delay, which are expressed as follows:
wherein T is C Representing the time delay generated by uploading the vehicle computing task to the EC server, the time delay of processing the computing task by the EC server and the time delay and returnThe sum of the return delays of the result to the vehicle, f C Representing the computing power, tr, of an EC server VR Representing the transmission rate between the vehicle and the RSU, tr RE Representing the transmission rate between an RSU and an EC server, tr EV Representing a transmission rate between the EC server and the vehicle;
if some or all of the subtasks in the vehicle computing task are migrated to the EC server for processing, the wireless transmission energy consumption of the vehicle is expressed as follows:
wherein E is C Refers to wireless transmission energy consumption of a vehicle, P u Refers to the transmission power of an automobile on a wireless transmission channel, P refers to the transmission power between an RSU and an EC, and P r Refers to the received power of the automobile;
in the calculation migration scene, the time delay and the energy consumption determine the quality of the selected calculation migration method, and the vehicle and the EC server process the vehicle calculation tasks in parallel, so that T is determined V And T is C The maximum value of (2) is taken as the final time delay, so the final time delay T and the energy consumption E are expressed as follows:
E=E V +E C (7)
step five, the total target of the optimal calculation migration method is expressed as follows:
f={minT,minE} (8)。
the third step is based on the improved chaos-differential evolution algorithm to obtain the optimal solution of the mathematical model, which comprises the following steps:
step one, determining that the population scale NP is 50, the scaling factor F is 0.6, and the maximum iteration number T max The crossover probability factor CR is 0.5;
step two, enabling the iteration number to be 0, and generating individuals by chaotic initialization populationNotably therein->Is a 1 XM dimensional list, < >>Corresponding to each element in (a) is x which occurs in a specific step in the second step i Is a value of->Is an M x N dimensional list,/->Corresponding to each element of the second step is the y which occurs in the specific step i,n Is a value of (2);
wherein,the ith chromosome representing the 0 th generation in the population,/->The j-th gene representing the i-th chromosome of the 0 th generation, M is the total number of vehicle calculation subtasks, N is the total number of EC servers, r k Is a random number between 0,1, mu 1 Is a control parameter, and the value is 3.6;
for the generated population individualsAnd (3) performing upward rounding to adapt to the solution of the formula (8) in the specific step of the second step, and obtaining the method:
step three, performing mutation operation to generate variant individualsWherein->Is a 1 XM dimensional list, < >>Corresponding to each element in the second specific step is x i Is a value of->Is an M x N dimensional list,/->Corresponding to each element of the second specific step is y i,n Is a value of (2);
the individual variation is realized by a differential strategy, firstly, the individual variation is a differential vector taking an individual variation as a parent, and each vector is obtained from the parent, namely the T-th generation populationVector synthesis is carried out on two different individuals after vector scaling, namely:
wherein,is a variant individual, is->All are parent individuals, wherein i is not equal to r1 not equal to r2 not equal to r3, F represents the influence degree of the differential vector on the next generation of individuals, and the value is 0.6;
for the generated variant individualsRounding and rounding are carried out to adapt to the solution of the formula (8), and then the method can be obtained:
step four, performing cross operation to generate experimental individualsHere population individuals->The constraint requirement of the formula (1) in the specific step of the second step is to be met;
for the ith individual in the T generation populationAnd variant thereof>Performing cross operation among populations, wherein the cross operation is expressed as the following formula:
wherein rand () is a random decimal number between 0-1;
step five, selecting operation is carried out to generate the next generation individualGenerating population individuals->Is between NP and 2 NP;
the selection operation is to make the experimental subjectIs->Competing, if one party is Pareto the other party, the dominant party enters a new population, if the two individuals do not have a dominant relationship, the dominant party will +.>And->Simultaneously adding a new population, and selecting an operation expression as the following formula:
is the fitness function in the algorithm and is also the formula (8) in the second step;
step six, selecting the top NP individuals by adopting the rapid non-dominant sorting and the crowding degree calculation in the intensity pareto algorithmForming a new population, and recording the number np1 of individuals with the quality grade of 1 in the new population;
step seven, when np1 is not equal to NP, and judging the population diversity lambda T1 ,λ T2 Whether or not to be less than lambda min ,λ min If the value is the threshold value, chaotic substitution operation is carried out, and a chaotic standby new population is generated
Population diversity lambda T1 ,λ T2 The following formula is given:
the specific chaos operation process is as follows:
in the evolution process, when the diversity of the population is low, replacing individuals in the current population with individuals in the generated chaotic standby population according to the set probability so as to lead an algorithm to pick out the current local optimum, wherein the replacement probability P of the ith individual is determined by the individual in sequence in the population, namely the following formula:
P=i/NP (26)
where i is the ordering of individuals in the population, NP is the population size, atOn the basis of (1) generating a new chaotic standby individualIs>Wherein->From the following formulas (27), (28), (29), wherein>Is a 1 XM dimensional list, < >>Corresponding to each element in the second specific step is x i Is a value of->Is an M x N dimensional list,/->The second step corresponds to each element of the listY occurring in a specific step i,n The specific formula is as follows:
wherein mu 2 Is a control parameter, and the value is 3.8;
step eight, if T is less than T max T=t+1 and returns to step three; if t=t max All individuals with the quality grade of 1 in the population form a non-inferior optimal solution of a single-objective optimization problem, and all fitness functions corresponding to all optimal individuals, namely the value of a formula (8), are taken as a final set G to be output, wherein T refers to the iteration times;
and step nine, outputting the optimal individual corresponding to the minimum value in the set G obtained in the step eight and ending operation, wherein the calculation migration method corresponding to the selected optimal individual is used as a final calculation migration method.
The invention has the beneficial effects that:
the calculation migration method based on the 5G Internet of vehicles can break the resource limitation of vehicles, reduce the calculation amount of the vehicles, reduce the consumption of the battery capacity of the vehicles, save the storage resources of the vehicles and the like, and can also improve the user experience. The EC server is close to the vehicle side, and compared with the vehicle, the EC server has more sufficient calculation and storage resources, and can assist the vehicle to better process the vehicle calculation task. When the calculation task amount of the vehicle is too large, the vehicle-mounted terminal and the EC server can independently and parallelly process the calculation task of the vehicle, so that the problem that the vehicle cannot be timely processed due to the large calculation task amount is avoided to a certain extent. According to the invention, through an improved chaos-differential evolution algorithm, optimal time delay and energy consumption can be obtained, and the requirements of time delay and energy consumption in the environment of the Internet of vehicles can be met.
Drawings
Fig. 1 is a schematic diagram of a computing migration scenario according to the present invention.
Fig. 2 is a schematic flow chart of a computing migration method based on the 5G internet of vehicles according to the present invention.
Fig. 3 is a schematic diagram of a time delay simulation according to the present invention.
Fig. 4 is a schematic diagram of the energy consumption simulation according to the present invention.
Detailed Description
Please refer to fig. 1 to 4:
fig. 1 is a schematic diagram of a computing migration scenario, in fig. 1, there are vehicles, RSUs, EC servers, which communicate through 5G, the present invention only contemplates communication between a single vehicle and multiple EC servers.
The invention provides a 5G car networking-based calculation migration method, which comprises the following steps:
the first step, building a calculation migration scene: selecting a 5G-based Internet of vehicles scene as a calculation migration scene, wherein the calculation migration scene mainly comprises an EC server, an RSU, a vehicle and a cloud server, and the EC server, the RSU, the vehicle and the cloud server are communicated by using a 5G technology;
step two, obtaining a calculation migration mathematical model according to the calculation migration scene;
thirdly, obtaining an optimal solution of the mathematical model based on an improved chaos-differential evolution algorithm;
and fourthly, taking a calculation migration method corresponding to the optimal solution of the mathematical model as a final calculation migration method.
In the second step, the specific steps for obtaining the computational migration mathematical model according to the computational migration scene are as follows:
according to the calculation migration scene, the problem of reasonable distribution of calculation tasks generated by N EC servers and a single vehicle to the vehicle is researched, and the assumption is made that the single vehicle has one calculation task K= { K 1 ,k 2 ,...k i ,...,k M Required processing, subtask k i Is k i =(a i ,b i ,c i ) The vehicle computing tasks K are processed by the vehicle terminal or the EC server, and P computing tasks K are selected i Processing M-P computing tasks k at vehicle-mounted terminal i Migration to EC Server processing, each subtask k i One of the EC servers, either local or N, must be selected for processing as follows:
wherein a is i Indicating completion of task k i The required cycle number, b i Representing the size of the uploaded data, c i Represents the size of the returned data, M, P represents the number of subtasks, and M > P, x i Representing a calculation migration decision, determining a calculation task k i Whether processing is performed locally or on the EC server, y i,n Indicating whether task k is to be calculated i Assigned to EC server n;
x i ={0,1},x i when=1, the calculation task k is represented i Processing locally, x i When=0, it means that the calculation task is processed on the EC server, y i,n ={0,1},y i,n When=1, the calculation task k is represented i Assigned to EC servers n, y i,n When=0, the calculation task k is represented i Not assigned to EC server n;
step two, time delay and energy consumption generated by processing a vehicle computing task in a vehicle-mounted terminal;
if some or all of the subtasks in the vehicle computing task are processed locally, the computing latency is expressed as follows:
wherein T is V Representing time delay generated by processing vehicle computing task in vehicle-mounted terminal, f V Representing the computing power of the vehicle V, i.e. the number of CPU cycles per unit time;
if some or all of the subtasks in the vehicle calculation task are processed in the vehicle-mounted terminal, the calculation energy consumption is expressed as follows:
wherein E is V Calculation energy consumption P generated by processing vehicle calculation task in vehicle-mounted terminal V The energy consumption power is realized in a unit CPU period of the vehicle;
step three, processing generated time delay and energy consumption of the vehicle computing task on the EC server:
if some or all of the subtasks in the vehicle computing task need to be migrated to the EC server for processing, the generated delay includes an uploading delay, a processing delay and a backhaul delay, which are expressed as follows:
wherein T is C Representing the sum of time delay generated by uploading the calculation task to the EC server by the vehicle, time delay of processing the calculation task by the EC server and time delay of returning the result to the vehicle, f C Representing the computing power, tr, of an EC server VR Representing the transmission rate between the vehicle and the RSU, tr RE Representing the transmission rate between an RSU and an EC server, tr EV Representing a transmission rate between the EC server and the vehicle;
if some or all of the subtasks in the vehicle computing task are migrated to the EC server for processing, the wireless transmission energy consumption of the vehicle is expressed as follows:
wherein E is C Refers to wireless transmission energy consumption of a vehicle, P u Refers to the transmission power of an automobile on a wireless transmission channel, P refers to the transmission power between an RSU and an EC, and P r Refers to the received power of the automobile;
in the calculation migration scene, the time delay and the energy consumption determine the quality of the selected calculation migration method, and the vehicle and the EC server process the vehicle calculation tasks in parallel, so that T is determined V And T is C The maximum value of (2) is taken as the final time delay, so the final time delay T and the energy consumption E are expressed as follows:
E=E V +E C (7)
step five, the total target of the optimal calculation migration method is expressed as follows:
f={minT,minE} (8)。
the third step is based on the improved chaos-differential evolution algorithm to obtain the optimal solution of the mathematical model, which comprises the following steps:
step one, determining that the population scale NP is 50, the scaling factor F is 0.6, and the maximum iteration number T max The crossover probability factor CR is 0.5;
step two, enabling the iteration number to be 0, and generating individuals by chaotic initialization populationNotably therein->Is a 1 XM dimensional list, < >>Corresponding to each element in (a) is x which occurs in a specific step in the second step i Is a value of->Is an M x N dimensional list,/->Corresponds to the first elementY occurring in a particular step of the two steps i,n Is a value of (2);
wherein,the ith chromosome representing the 0 th generation in the population,/->The j-th gene representing the i-th chromosome of the 0 th generation, M is the total number of vehicle calculation subtasks, N is the total number of EC servers, r k Is a random number between 0,1, mu 1 Is a control parameter, and the value is 3.6;
for the generated population individualsAnd (3) performing upward rounding to adapt to the solution of the formula (8) in the specific step of the second step, and obtaining the method:
step three, performing mutation operation to generate variant individualsWherein->Is a 1 XM dimensional list, < >>Corresponding to each element in the second specific step is x i Is a value of->Is an M x N dimensional list,/->Corresponding to each element of the second specific step is y i,n Is a value of (2);
the individual variation is realized by a differential strategy, firstly, the individual variation is a differential vector taking an individual variation as a parent, and each vector is obtained from the parent, namely the T-th generation populationVector synthesis is carried out on two different individuals after vector scaling, namely:
wherein,is a variant individual, is->All are parent individuals, wherein i is not equal to r1 not equal to r2 not equal to r3, F represents the influence degree of the differential vector on the next generation of individuals, and the value is 0.6;
for the generated variant individualsRounding and rounding are carried out to adapt to the solution of the formula (8), and then the method can be obtained:
step four, performing cross operation to generate experimental individualsHere population individuals->The constraint requirement of the formula (1) in the specific step of the second step is to be met;
for the ith individual in the T generation populationAnd variant thereof>Performing cross operation among populations, wherein the cross operation is expressed as the following formula:
wherein rand () is a random decimal number between 0-1;
step five, selecting operation is carried out to generate the next generation individualGenerating population individuals->Is between NP and 2 NP;
the selection operation is to make the experimental subjectIs->Competing, if one party is Pareto the other party, the dominant party enters a new population, if the two individuals do not have a dominant relationship, the dominant party will +.>And->Simultaneously adding a new population, and selecting an operation expression as the following formula:
is the fitness function in the algorithm and is also the formula (8) in the second step;
step six, selecting the top NP individuals by adopting the rapid non-dominant sorting and the crowding degree calculation in the intensity pareto algorithmComposition newAnd recording the number np1 of individuals with a merit grade of 1 in the newly generated population;
step seven, when np1 is not equal to NP, and judging the population diversity lambda T1 ,λ T2 Whether or not to be less than lambda min ,λ min If the value is the threshold value, chaotic substitution operation is carried out, and a chaotic standby new population is generated
Population diversity lambda T1 ,λ T2 The following formula is given:
the specific chaos operation process is as follows:
in the evolution process, when the diversity of the population is low, replacing individuals in the current population with individuals in the generated chaotic standby population according to the set probability so as to lead an algorithm to pick out the current local optimum, wherein the replacement probability P of the ith individual is determined by the individual in sequence in the population, namely the following formula:
P=i/NP (26)
where i is the ordering of individuals in the population, NP is the population size, atOn the basis of (1) generating a new chaotic standby individualIs>Wherein->From the following formulas (27), (28), (29), wherein>Is a 1 XM dimensional list, < >>Corresponding to each element in the second specific step is x i Is a value of->Is an M x N dimensional list,/->Corresponding to each element of the second specific step is y i,n The specific formula is as follows:
wherein mu 2 Is a control parameter, and the value is 3.8;
step eight, if T is less than T max T=t+1 and returns to step three; if t=t max All individuals with the quality grade of 1 in the population form a non-inferior optimal solution of a single-objective optimization problem, and all fitness functions corresponding to all optimal individuals, namely the value of a formula (8), are taken as a final set G to be output, wherein T refers to the iteration times;
and step nine, outputting the optimal individual corresponding to the minimum value in the set G obtained in the step eight and ending operation, wherein the calculation migration method corresponding to the selected optimal individual is used as a final calculation migration method.
In the above embodiment, assuming that the vehicles are random and independently present and the speed of the vehicles is fixed, the vehicles on the straight road are subject to the poisson distribution. The vehicles and the EC server, and the EC server are communicated by using 5G. The vehicle each computing task i selects one of the N EC servers to handle, and each EC server computing storage resources is plentiful.
Python language programming is performed on a window7 system by using a spyder platform to obtain simulation diagrams, and the simulation diagrams are shown in fig. 3 and 4. Fig. 3 is a resulting time delay simulation diagram, and fig. 4 is an energy consumption simulation diagram.
In FIG. 3, the abscissa is the calculated amount required for the task, i.e., a occurs in the second step i The ordinate represents the value of the time delay T that occurs in the formula (8). From the figure it can be seen that a i When the value is from 400Megacycles to 1800Megacycles, the time delay T is changed between 0 and 0.24 s. It can be seen that the time delay of the calculation migration method obtained by using the improved chaos-differential evolution method is controlled within 0.24s, which meets the time delay requirement in the Internet of vehicles.
In FIG. 4, the abscissa is the calculated amount required for the task, i.e., a occurring in the second specific step i The ordinate represents the value of the energy consumption E appearing in the formula (8). From the figure it can be seen that a i When the value is from 400Megacycles to 1800Megacycles, the energy consumption E changes between 0J and 0.37J. It can be seen that the energy consumption is not great under the condition of meeting the requirement of the Internet of vehicles on time delay, and the calculation migration method obtained by using the improved chaos-differential evolution method is excellent.
Claims (1)
1. A calculation migration method based on 5G Internet of vehicles is characterized by comprising the following steps: the method is as follows:
the first step, building a calculation migration scene: selecting a 5G-based Internet of vehicles scene as a calculation migration scene, wherein the calculation migration scene mainly comprises an EC server, an RSU, a vehicle and a cloud server, and the EC server, the RSU, the vehicle and the cloud server are communicated by using a 5G technology;
step two, obtaining a calculation migration mathematical model according to the calculation migration scene; the specific steps for obtaining the computational migration mathematical model according to the computational migration scene are as follows:
according to the calculation migration scene, the problem of reasonable distribution of calculation tasks generated by N EC servers and a single vehicle to the vehicle is researched, and the assumption is made that the single vehicle has one calculation task K= { K 1 ,k 2 ,...k i ,...,k M Required processing, subtask k i Is k i =(a i ,b i ,c i ) The vehicle computing tasks K are processed by the vehicle terminal or the EC server, and P computing tasks K are selected i Processing M-P computing tasks k at vehicle-mounted terminal i Migration to EC Server processing, each subtask k i One of the EC servers, either local or N, must be selected for processing as follows:
wherein a is i Indicating completion of task k i The required cycle number, b i Representing the size of the uploaded data, c i Represents the size of the returned data, M, P represents the number of subtasks, and M > P, x i Representing a calculation migration decision, determining a calculation task k i Whether processing is performed locally or on the EC server, y i,n Indicating whether task k is to be calculated i Assigned to EC server n;
x i ={0,1},x i when=1Representing a computing task k i Processing locally, x i When=0, it means that the calculation task is processed on the EC server, y i,n ={0,1},y i,n When=1, the calculation task k is represented i Assigned to EC servers n, y i,n When=0, the calculation task k is represented i Not assigned to EC server n;
step two, time delay and energy consumption generated by processing a vehicle computing task in a vehicle-mounted terminal;
if some or all of the subtasks in the vehicle computing task are processed locally, the computing latency is expressed as follows:
wherein T is V Representing time delay generated by processing vehicle computing task in vehicle-mounted terminal, f V Representing the computing power of the vehicle V, i.e. the number of CPU cycles per unit time;
if some or all of the subtasks in the vehicle calculation task are processed in the vehicle-mounted terminal, the calculation energy consumption is expressed as follows:
wherein E is V Calculation energy consumption P generated by processing vehicle calculation task in vehicle-mounted terminal V The energy consumption power is realized in a unit CPU period of the vehicle;
step three, processing generated time delay and energy consumption of the vehicle computing task on the EC server:
if some or all of the subtasks in the vehicle computing task need to be migrated to the EC server for processing, the generated delay includes an uploading delay, a processing delay and a backhaul delay, which are expressed as follows:
wherein T is C Representing the sum of time delay generated by uploading the calculation task to the EC server by the vehicle, time delay of processing the calculation task by the EC server and time delay of returning the result to the vehicle, f C Representing the computing power, tr, of an EC server VR Representing the transmission rate between the vehicle and the RSU, tr RE Representing the transmission rate between an RSU and an EC server, tr EV Representing a transmission rate between the EC server and the vehicle;
if some or all of the subtasks in the vehicle computing task are migrated to the EC server for processing, the wireless transmission energy consumption of the vehicle is expressed as follows:
wherein E is C Refers to wireless transmission energy consumption of a vehicle, P u Refers to the transmission power of an automobile on a wireless transmission channel, P refers to the transmission power between an RSU and an EC, and P r Refers to the received power of the automobile;
in the calculation migration scene, the time delay and the energy consumption determine the quality of the selected calculation migration method, and the vehicle and the EC server process the vehicle calculation tasks in parallel, so that T is determined V And T is C The maximum value of (2) is taken as the final time delay, so the final time delay T and the energy consumption E are expressed as follows:
E=E V +E C (7)
step five, the total target of the optimal calculation migration method is expressed as follows:
f={minT,minE} (8);
thirdly, obtaining an optimal solution of the mathematical model based on an improved chaos-differential evolution algorithm; the method for obtaining the optimal solution of the mathematical model based on the improved chaos-differential evolution algorithm comprises the following steps:
step one, determining that the population scale NP is 50, the scaling factor F is 0.6, and the maximum iteration number T max The crossover probability factor CR is 0.5;
step two, enabling the iteration number to be 0, and generating individuals by chaotic initialization populationNotably therein->Is a 1 XM dimensional list, < >>Corresponding to each element in (a) is x which occurs in a specific step in the second step i Is a value of->Is an M x N dimensional list,/->Corresponding to each element of the second step is the y which occurs in the specific step i,n Is a value of (2);
wherein,the ith chromosome representing the 0 th generation in the population,/->The j-th gene representing the i-th chromosome of the 0 th generation, M is the total number of vehicle calculation subtasks, N is the total number of EC servers, r k Is a random number between 0,1, mu 1 Is a control parameter, and the value is 3.6;
for the generated population individualsAnd (3) performing upward rounding to adapt to the solution of the formula (8) in the specific step of the second step, and obtaining the method:
step three, performing mutation operation to generate variant individualsWherein->Is a 1 XM dimensional list, < >>Corresponding to each element in the second specific step is x i Is a value of->Is an M x N dimensional list,/->Corresponding to each element of the second specific step is y i,n Is a value of (2);
the individual variation is realized by a differential strategy, firstly, the individual variation is a differential vector taking an individual variation as a parent, and each vector is obtained from the parent, namely the T-th generation populationVector synthesis is carried out on two different individuals after vector scaling, namely:
wherein,is a variant individual, is->All are parent individuals, wherein i is not equal to r1 not equal to r2 not equal to r3, F represents the influence degree of the differential vector on the next generation of individuals, and the value is 0.6;
for the generated variant individualsRounding and roundingSolving an adaptation formula (8) to obtain:
step four, performing cross operation to generate experimental individualsHere population individuals->The constraint requirement of the formula (1) in the specific step of the second step is to be met;
for the ith individual in the T generation populationAnd variant thereof>Performing cross operation among populations, wherein the cross operation is expressed as the following formula:
wherein rand () is a random decimal number between 0-1;
step five, selecting operation is carried out to generate the next generation individualGenerating population individuals->Is between NP and 2 NP;
the selection operation is to make the experimental subjectIs->Competing, if one party is Pareto the other party, the dominant party enters a new population, if the two individuals do not have a dominant relationship, the dominant party will +.>And->Simultaneously adding a new population, and selecting an operation expression as the following formula:
is the fitness function in the algorithm and is also the formula (8) in the second step;
step six, selecting the top NP individuals by adopting the rapid non-dominant sorting and the crowding degree calculation in the intensity pareto algorithmForming a new population, and recording the number np1 of individuals with the quality grade of 1 in the new population;
step seven, when np1 is not equal to NP, and judging the population diversity lambda T1 ,λ T2 Whether or not to be less than lambda min ,λ min If the value is the threshold value, chaotic substitution operation is carried out, and a chaotic standby new population is generated
Population diversity lambda T1 ,λ T2 The following formula is given:
the specific chaos operation process is as follows:
in the evolution process, when the diversity of the population is low, replacing individuals in the current population with individuals in the generated chaotic standby population according to the set probability so as to lead an algorithm to pick out the current local optimum, wherein the replacement probability P of the ith individual is determined by the individual in sequence in the population, namely the following formula:
P=i/NP (26)
where i is the ordering of individuals in the population, NP is the population size, atOn the basis of (1) generating a new chaotic standby individualIs>Wherein-> From the following formulas (27), (28), (29), wherein>Is a 1 XM dimensional list, < >>Corresponding to each element in the second specific step is x i Is a value of->Is an M x N dimensional list,/->Corresponding to each element of the second specific step is y i,n The specific formula is as follows:
wherein mu 2 Is a control parameter, and the value is 3.8;
step eight, if T is less than T max T=t+1 and returns to step three; if t=t max All individuals with the quality grade of 1 in the population form a non-inferior optimal solution of a single-objective optimization problem, and all fitness functions corresponding to all optimal individuals, namely the value of a formula (8), are taken as a final set G to be output, wherein T refers to the iteration times;
outputting the optimal individual corresponding to the minimum value in the set G obtained in the step eight and ending operation, wherein the calculation migration method corresponding to the selected optimal individual is used as a final calculation migration method;
and fourthly, taking a calculation migration method corresponding to the optimal solution of the mathematical model as a final calculation migration method.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111048558.3A CN113742077B (en) | 2021-09-08 | 2021-09-08 | Computing migration method based on 5G Internet of vehicles |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111048558.3A CN113742077B (en) | 2021-09-08 | 2021-09-08 | Computing migration method based on 5G Internet of vehicles |
Publications (2)
Publication Number | Publication Date |
---|---|
CN113742077A CN113742077A (en) | 2021-12-03 |
CN113742077B true CN113742077B (en) | 2024-03-01 |
Family
ID=78736925
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202111048558.3A Active CN113742077B (en) | 2021-09-08 | 2021-09-08 | Computing migration method based on 5G Internet of vehicles |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113742077B (en) |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109710374A (en) * | 2018-12-05 | 2019-05-03 | 重庆邮电大学 | The VM migration strategy of task unloading expense is minimized under mobile edge calculations environment |
CN109840154A (en) * | 2019-01-08 | 2019-06-04 | 南京邮电大学 | A kind of computation migration method that task based access control relies under mobile cloud environment |
CN111984419A (en) * | 2020-08-28 | 2020-11-24 | 华侨大学 | Complex task computing and transferring method for marginal environment reliability constraint |
CN112995289A (en) * | 2021-02-04 | 2021-06-18 | 天津理工大学 | Internet of vehicles multi-target computing task unloading scheduling method based on non-dominated sorting genetic strategy |
CN113139639A (en) * | 2021-05-17 | 2021-07-20 | 华侨大学 | MOMBI-based smart city application-oriented multi-target calculation migration method and device |
-
2021
- 2021-09-08 CN CN202111048558.3A patent/CN113742077B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109710374A (en) * | 2018-12-05 | 2019-05-03 | 重庆邮电大学 | The VM migration strategy of task unloading expense is minimized under mobile edge calculations environment |
CN109840154A (en) * | 2019-01-08 | 2019-06-04 | 南京邮电大学 | A kind of computation migration method that task based access control relies under mobile cloud environment |
CN111984419A (en) * | 2020-08-28 | 2020-11-24 | 华侨大学 | Complex task computing and transferring method for marginal environment reliability constraint |
CN112995289A (en) * | 2021-02-04 | 2021-06-18 | 天津理工大学 | Internet of vehicles multi-target computing task unloading scheduling method based on non-dominated sorting genetic strategy |
CN113139639A (en) * | 2021-05-17 | 2021-07-20 | 华侨大学 | MOMBI-based smart city application-oriented multi-target calculation migration method and device |
Also Published As
Publication number | Publication date |
---|---|
CN113742077A (en) | 2021-12-03 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111507601B (en) | Resource optimization allocation decision method based on deep reinforcement learning and block chain consensus | |
Cao et al. | Resource allocation in 5G IoV architecture based on SDN and fog-cloud computing | |
CN107172166B (en) | Cloud and mist computing system for industrial intelligent service | |
CN108182115B (en) | Virtual machine load balancing method in cloud environment | |
Zhang et al. | New method of traffic flow forecasting based on quantum particle swarm optimization strategy for intelligent transportation system | |
Jiang et al. | Distributed resource scheduling for large-scale MEC systems: A multiagent ensemble deep reinforcement learning with imitation acceleration | |
Wang et al. | A deep learning based energy-efficient computational offloading method in Internet of vehicles | |
Jian et al. | A cloud edge-based two-level hybrid scheduling learning model in cloud manufacturing | |
CN111611062B (en) | Cloud-edge collaborative hierarchical computing method and cloud-edge collaborative hierarchical computing system | |
CN112822234A (en) | Task unloading method based on deep reinforcement learning in Internet of vehicles | |
CN113781002B (en) | Low-cost workflow application migration method based on agent model and multiple group optimization in cloud edge cooperative network | |
CN115374853A (en) | Asynchronous federal learning method and system based on T-Step polymerization algorithm | |
CN113778691B (en) | Task migration decision method, device and system | |
CN116541106B (en) | Computing task unloading method, computing device and storage medium | |
Zhao et al. | Adaptive Swarm Intelligent Offloading Based on Digital Twin-assisted Prediction in VEC | |
CN113742077B (en) | Computing migration method based on 5G Internet of vehicles | |
CN116599860B (en) | Network traffic gray prediction method based on reinforcement learning | |
CN114650321A (en) | Task scheduling method for edge computing and edge computing terminal | |
Zhao et al. | Communication-efficient federated learning for digital twin systems of industrial Internet of Things | |
CN112380006A (en) | Data center resource allocation method and device | |
CN116645130A (en) | Automobile order demand prediction method based on combination of federal learning and GRU | |
CN112764932B (en) | Deep reinforcement learning-based calculation-intensive workload high-energy-efficiency distribution method | |
CN115150335A (en) | Optimal flow segmentation method and system based on deep reinforcement learning | |
Ren et al. | Balanced allocation method of physical education distance education resources based on linear prediction | |
Wang et al. | Resource allocation algorithm for MEC based on Deep Reinforcement Learning |
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 |