Method and system for selecting optimal edge computing node in internet of vehicle environment
The present disclosure provides a method and system for selecting an optimal edge computing node in an Internet of vehicle (IoV) environment. The method includes: acquiring and analyzing properties of computing tasks of a vehicle in the IoV environment; acquiring and analyzing properties of different edge computing nodes; computing matching degrees between the properties of the computing tasks and the properties of the nodes; analyzing computing demands of different tasks, and assigning weights to different types of matching degrees; and selecting a node having an optimal sum for products of the matching degrees and the weights as an optimal edge computing node to compute each of the computing tasks of the vehicle.
1. A method for selecting an optimal edge computing node in an Internet of vehicle (IoV) environment, comprising the following steps:
step 1: acquiring and analyzing properties of computing tasks of a vehicle in the IoV environment as well as properties of different edge computing nodes;
step 2: computing matching degrees between the properties of the computing tasks of the vehicle and the properties of the edge computing nodes, wherein a data volume of each of the computing tasks of the vehicle is matched with data transmission speeds of the edge computing nodes, and a number of central processing unit (CPU) cycles required by each of the computing tasks of the vehicle is matched with computing resources allocated by the edge computing nodes;
step 3: analyzing computing demands of different computing tasks, assigning weights to different types of matching degrees, and computing comprehensive matching degrees; and
step 4: comparing the comprehensive matching degrees in step 3, and selecting an optimal edge computing node according to a comparison result to compute each of the computing tasks of the vehicle,
wherein, in step 1, the properties of the computing tasks of the vehicle comprise: data volumes D j of the computing tasks, numbers C j of CPU cycles required by the computing tasks, maximum time T j max required to complete the tasks, and distances r ij between the vehicle and the nodes;
the properties of the edge computing nodes comprise bandwidths B j of the edge computing nodes, percentages b ij of time slots allocated by the edge computing nodes to the vehicle in unit time, computing resources f ij allocated by the edge computing nodes to the vehicle, and average signal-to-noise ratios (SNRs)
S
N
after the edge computing nodes are connected to the vehicle;
wherein matching the properties of the tasks with the properties of the nodes in step 2 specifically comprises:
matching the data volume of each of the computing tasks of the vehicle with the data transmission speeds of the edge computing nodes, to obtain matching degrees
J
ij
=
t
ij
Smin
t
ij
S
,
J
ij
∈
(
0
,
1
)
between the data volume of each of the computing tasks of the vehicle and the data transmission speeds of the edge computing nodes,
wherein, the data transmission speeds v ij of the edge computing nodes are computed according to an equation
v
ij
=
b
ij
v
j
=
b
ij
B
j
log
2
(
1
+
S
N
)
and transmission time of all base stations for a data volume of each of computing tasks of a vehicle j is computed according to an equation
t
ij
S
=
D
j
v
ij
to obtain minimum transmission time t ij S min for each of the computing task of the vehicle j, with an equation of t ij S min =min {t 1j S , t 2j S , t 3j S . . . t nj S };
matching the number of CPU cycles required by each of the computing tasks of the vehicle with the computing resources allocated by the edge computing nodes, to obtain matching degrees
K
ij
=
t
ij
Emin
t
ij
E
,
K
ij
∈
(
0
,
1
)
between the number of CPU cycles required by each of the computing tasks of the vehicle and the computing resources allocated by the edge computing nodes,
wherein execution time of all base stations for the data volume of each of the computing tasks of the vehicle j is computed according to an equation
t
ij
E
=
C
j
f
ij
to obtain minimum transmission time t ij E min =min {t 1j E , t 2j E , t 3j E . . . t nj E } for each of the computing task of the vehicle j;
matching maximum completion time for each of the computing tasks of the vehicle with time required by the nodes to compute the task, an equation for computing matching degrees L ij being
L
ij
=
T
j
m
a
x
-
t
ij
T
j
m
a
x
,
wherein time t ij required to complete each of the tasks is computed according to an equation
t
ij
=
D
j
v
ij
+
C
j
f
ij
;
and
computing matching degrees with the distances r ij between the vehicle and the nodes and a maximum distance r ij max =max{r 1j , r 2j , r 3j . . . r nj } according to an equation
R
ij
=
r
ij
m
a
x
-
r
ij
r
ij
m
a
x
.
2. The method for selecting an optimal edge computing node in an IoV environment according to claim 1 , wherein the assigning weights to different types of matching degrees in step 3 comprises:
respectively computing a mean and a variance for each of D j , C j , T j max according to the properties T j ={D j , C j , T j max }j∈M of the computing tasks,
wherein for the data volumes D j of the computing tasks and the numbers C j of CPU cycles required by the computing tasks, the D j and C j each have a demand degree O D , O C of 3 if being greater than or equal to a sum of a mean and a variance thereof; the D j and C j each have a demand degree O D , O C of 2 if being between the sum of the mean and the variance thereof and a difference between the mean and the variance thereof; and the D j and C j each have a demand degree O D , O C of 1 if being less than or equal to the difference between the mean and the variance thereof
wherien for the maximum time T j max required to complete the tasks, the T j max has a demand degree O T of 3 if being less than or equal to a difference between a mean and a variance thereof; the T j max has a demand degree O T of 2 if being between a sum of the mean and the variance thereof and the difference between the mean and the variance thereof; and the T j max has a demand degree O T of 1 if being greater than or equal to the sum of the mean and the variance thereof;
wherein distances toward base stations of the edge computing nodes for each of the computing tasks have a demand degree O R of 2;
D j , C j , T j max respectively correspond to a transmission speed demand, an execution speed demand and a task completion time demand, with corresponding matching degrees of J ij , K ij , L ij ;
and after the demand degrees O D , O C , O T , O R are obtained, weights w R , w J , w K , w L are computed according to equations:
w
J
=
O
D
O
D
+
O
C
+
O
T
+
O
R
w
K
=
O
C
O
D
+
O
C
+
O
T
+
O
R
w
L
=
O
T
O
D
+
O
C
+
O
T
+
O
R
w
R
=
O
R
O
D
+
O
C
+
O
T
+
O
R
,
wherein O D , O C , O T , O R respectively represent the transmission speed demand, the execution speed demand, the task completion time demand and the distance demand, and proportions of the O D , O C , O T , O R in a total demand are the weights of the matching degrees J ij , K ij , L ij , R ij ; and
an equation for computing comprehensive matching degrees is:
Z ij =R ij w R +J ij w J +K ij w K +L ij w L .
3. The method for selecting an optimal edge computing node in an IoV environment according to claim 2 , wherein the comparing the comprehensive matching degrees in step 4 comprises:
matching each of the computing tasks of the vehicle with multiple edge computing nodes, assigning values to weights of different matching degrees, and computing multiple comprehensive matching degrees,
wherein when the comprehensive matching degrees are compared, if only one maximum value is present among the comprehensive matching degrees for the computing task of the vehicle j, a node where the value is located is selected as an optimal edge computing node for the computing task of the vehicle j.
4. The method for selecting an optimal edge computing node in an IoV environment according to claim 3 , wherein in the step 4, when the comprehensive matching degrees are compared, if two or more same maximum values are present among the computed comprehensive matching degrees, a comparison sequence for the matching degrees J ij , K ij , L ij , R ij is determined according to the weights w R , w J , w K , w L , a matching degree having a larger weight is compared first, and a node having a largest matching degree is selected to execute the computing task.
5. A system for selecting an optimal edge computing node in an Internet of vehicle (IoV) environment, comprising:
a property acquisition module, configured to acquire properties of computing tasks of a vehicle in the IoV environment as well as properties of different edge computing nodes, wherein the properties of the computing tasks of the vehicle comprise: data volumes D j of the computing tasks, numbers C j of central processing unit (CPU) cycles required by the computing tasks, maximum time T j max required to complete the tasks, and distances r ij between the vehicle and the nodes; and the properties of the edge computing nodes comprise: bandwidths B j of the edge computing nodes, percentages b ij of time slots allocated by the edge computing nodes to the vehicle in unit time, computing resources f ij allocated by the edge computing nodes to the vehicle, and average signal-to-noise ratios (SNRs)
S
N
after the edge computing nodes are connected to the vehicle;
a matching degree acquisition module, configured to compute matching degrees between the properties of the computing tasks of the vehicle and the properties of the edge computing nodes, wherein a data volume of each of the computing tasks of the vehicle is matched with the data transmission speeds of the edge computing nodes, and a number of CPU cycles required by each of the computing tasks of the vehicle is matched with the computing resources allocated by the edge computing nodes;
a comprehensive matching degree acquisition module, configured to analyze computing demands of different computing tasks, assign weights to different types of matching degrees, and compute comprehensive matching degrees; and
a comparison and selection module, configured to compare the comprehensive matching degrees acquired by the comprehensive matching degree acquisition module, and select an optimal edge computing node according to a comparison result to compute each of the computing tasks of the vehicle; wherein
the matching degree acquisition module comprises:
a first acquisition module, configured to match the data volume of each of the computing tasks of the vehicle with the data transmission speeds of the edge computing nodes, to obtain matching degrees
J
ij
=
t
ij
Smin
t
ij
S
,
J
ij
∈
(
0
,
1
)
between the data volume of each of the computing tasks of the vehicle and the data transmission speeds of the edge computing nodes,
wherein the data transmission speeds v ij of the edge computing nodes are computed according to an equation
v
ij
=
b
ij
v
j
=
b
ij
B
j
log
2
(
1
+
S
N
)
;
and transmission time of all base stations for a data volume of each of computing tasks of a vehicle j is computed according to an equation
t
ij
S
=
D
j
v
ij
to obtain minimum transmission time t ij S min for each of the computing tasks of the vehicle j, with an equation of t ij S min =min {t 1j S , t 2j S , t 3j S . . . t nj S };
a second acquisition module, configured to match the number of CPU cycles required by each of the computing tasks of the vehicle with the computing resources allocated by the edge computing nodes, to obtain matching degrees
K
ij
=
t
ij
Emin
t
ij
E
,
K
ij
∈
(
0
,
1
)
between the number of CPU cycles required by the each of the computing tasks of the vehicle and the computing resources allocated by the edge computing nodes,
wherein execution time of all base stations for the data volume of each of the computing tasks of the vehicle j is computed according to an equation
t
ij
E
=
C
j
f
ij
to obtain minimum transmission time t ij E min =min {t 1j E , t 2j E , t 3j E . . . t nj E } for each of the computing tasks of the vehicle j;
a third acquisition module, configured to match maximum completion time for each of the computing tasks of the vehicle with time required by the nodes to compute the task, an equation for computing matching degrees L ij being
L
ij
=
T
j
m
a
x
-
t
ij
T
j
m
a
x
,
wherein time t ij required to complete each of the tasks is computed according to an equation
t
ij
=
D
j
v
ij
+
C
j
f
ij
;
and
a fourth acquisition module, configured to compute matching degrees with the distances r ij between the vehicle and the nodes and a maximum distance r ij max =max{r 1j , r 2j , r 3j . . . r nj } according to an equation
R
ij
=
r
ij
m
a
x
-
r
ij
r
ij
m
a
x
.
6. The system for selecting an optimal edge computing node in an IoV environment according to claim 5 , wherein in the matching degree acquisition module, the weights are assigned to the different types of matching degrees as follows:
respectively computing a mean and a variance for each of D j , C j , T j max according to the properties T j ={D j , C j , T j max }j∈M of the computing tasks,
wherein for the data volumes D j of the computing tasks and the numbers C j of CPU cycles required by the computing tasks, the D j and C j each have a demand degree O D , O C of 3 if being greater than or equal to a sum of a mean and a variance thereof; the D j and C j each have a demand degree O D , O C of 2 if being between the sum of the mean and the variance thereof and a difference between the mean and the variance thereof; and the D j and C j each have a demand degree O D , O C of 1 if being less than or equal to the difference between the mean and the variance thereof;
wherein for the maximum time T j max required to complete the tasks, the T j max has a demand degree O T of 3 if being less than or equal to a difference between a mean and a variance thereof; the T j max has a demand degree O T of 2 if being between a sum of the mean and the variance thereof and the difference between the mean and the variance thereof; and the T j max has a demand degree O T of 1 if being greater than or equal to the sum of the mean and the variance thereof
wherein distances toward base stations of the edge computing nodes for the computing tasks have a demand degree O R of 2;
D j , C j , T j max respectively correspond to a transmission speed demand, an execution speed demand and a task completion time demand, with corresponding matching degrees of J ij , K ij , L ij ;
and after the demand degrees O D , O C , O T , O R are obtained, weights w R , w J , w K , w L are computed according to equations:
w
J
=
O
D
O
D
+
O
C
+
O
T
+
O
R
w
K
=
O
C
O
D
+
O
C
+
O
T
+
O
R
w
L
=
O
T
O
D
+
O
C
+
O
T
+
O
R
w
R
=
O
R
O
D
+
O
C
+
O
T
+
O
R
,
wherein O D , O C , O T , O R respectively represent the transmission speed demand, the execution speed demand, the task completion time demand and the distance demand, and proportions of the O D , O C , O T , O R in a total demand are the weights of the matching degrees J ij , K ij , L ij , R ij ; and
an equation for computing comprehensive matching degrees is:
Z ij =R ij w R +J ij w J +K ij w K +L ij w L .
7. The system for selecting an optimal edge computing node in an IoV environment according to claim 6 , wherein in the comprehensive matching degree acquisition module, the comprehensive matching degrees are compared as follows:
matching each of the computing tasks of the vehicle with multiple edge computing nodes, assigning values to weights of different matching degrees, and computing multiple comprehensive matching degrees,
wherein when the comprehensive matching degrees are compared, if only one maximum value is present among the comprehensive matching degrees for the computing task of the vehicle j, a node where the value is located is selected as an optimal edge computing node for the computing task of the vehicle j; and if two or more same maximum values are present among the computed comprehensive matching degrees, a comparison sequence for the matching degrees J ij , K ij , L ij , R ij is determined according to the weights w R , w J , w K , w L , a matching degree having a larger weight is compared first, and a node having a largest matching degree is selected to execute the computing task.
8. The system for selecting an optimal edge computing node in an IoV environment according to claim 5 , wherein each of the edge computing nodes is provided with one base station; base stations of different edge computing nodes are connected in a wired manner; and the vehicle communicates with the base stations through time division multiple address (TDMA).