Variable configuration method and variable configuration device thereof
A variable configuration method and a variable configuration device thereof are disclosed. The variable configuration method for the variable configuration device includes receiving an incident prediction notification from a network at a first time instant, determining at least one variable at a second time instant in response to the incident prediction notification, and outputting the at least one variable at a third time instant. The incident prediction notification is used to instruct the variable configuration device to determine the at least one variable to respond to a quality of service (QoS) violation prediction at a fourth time instant later than the third time instant.
1. A variable configuration method, for a variable configuration device, comprising:
receiving an incident prediction notification from a network at a first time instant;
determining at least one variable at a second time instant in response to the incident prediction notification, wherein the step of determining the at least one variable at the second time instant comprises maximizing a plurality of posterior probabilities of assigning a plurality of subdata of a grounding data to a plurality of observation functions and a causal structure to generate a causal graph, wherein one of the plurality of posterior probabilities is proportional to Π t=0 T P(w i,t |s t-1 ,C, ƒ i ), where S t-1 denotes a plurality of states of a plurality of causal variables at a time instant t- 1 , T denotes a fifth time instant earlier than a fourth time instant, C denotes the causal structure, ƒ i denotes one of the plurality of observation functions, which is corresponding to an i-th causal variable of the plurality of causal variables, W i,t denotes one of the plurality of subdata, which is corresponding to the i-th causal variable, at a time instant t; and
outputting the at least one variable at a third time instant, wherein the incident prediction notification is used to instruct the variable configuration device to determine the at least one variable to respond to a quality of service (QoS) violation prediction at the fourth time instant, and the third time instant is earlier than the fourth time instant.
2. The variable configuration method of claim 1 , wherein the network is configured to predict inability to meet at least one QoS requirement at the fourth time instant according to a continuous time structural equation modeling to output the incident prediction notification.
3. The variable configuration method of claim 1 , wherein the network is configured to compare at least one temporal causal variable at the fourth time instant with the at least one variable at the fourth time instant to perform the QoS violation prediction, and the at least one temporal causal variable constitutes a vector η h (t) which satisfies
η
h
(
t
)
=
e
A
(
t
-
t
0
)
η
h
(
t
0
)
+
A
-
1
[
e
A
(
t
-
t
0
)
-
I
]
ξ
h
+
A
-
1
[
e
A
(
t
-
t
0
)
-
I
]
Bz
h
+
M
∑
u
x
h
,
u
δ
(
t
-
u
)
+
∫
t
0
t
e
A
(
t
-
s
)
GdW
h
(
s
)
,
where A denotes a qualitative matrix, t 0 denotes an initial time instant, I denotes an identity matrix, ξ h denotes a random vector, B denotes a transformation matrix, Z h denotes a time-independent predictor vector, M denotes a coefficient matrix, X h,u , denotes a time-dependent predictor vector, u denotes a pulse time instant, W h (s) denotes a random walk vector, and G denotes a lower triangular matrix.
4. The variable configuration method of claim 1 , wherein
at least one temporal causal variable comprises a lowest flow bit rate, a highest flow bit rate, a packet loss rate, a session-aggregate bit rate, a user equipment-aggregate bit rate, a packet delay time, a packet error rate, or a data burst volume, and
the at least one variable comprises a guaranteed flow bit rate, a maximum flow bit rate, a maximum packet loss rate, a session-aggregate maximum bit rate, a user equipment-aggregate maximum bit rate, a packet delay budget, a maximum packet error rate, or a maximum data burst volume.
5. The variable configuration method of claim 1 , wherein the at least one variable is decided at the second time instant according to Bayesian optimization, causal Bayesian optimization, or dynamic causal Bayesian optimization.
6. The variable configuration method of claim 1 , wherein the network is configured to modify at least one QoS parameter or at least one QoS characteristic according to the at least one variable at the fifth time instant earlier than the fourth time instant, and the at least one variable is mapped to a minimum value of a loss function at the fifth time instant.
7. The variable configuration method of claim 1 , wherein the causal structure of the causal graph between a loss function and the plurality of causal variables and the plurality of causal variables of the causal graph are determined together, and the plurality of causal variables comprising the at least one variable are selected from a plurality of independent variables.
8. The variable configuration method of claim 1 , wherein the plurality of observation functions are obtained based on a causal semantic generative model.
9. The variable configuration method of claim 1 , wherein a loss function at the fifth time instant is a function of the plurality of causal variables at the fifth time instant, the plurality of causal variables at least one time instant earlier than the fifth time instant, or the loss function at the at least one time instant.
10. A variable configuration device, comprising:
a storage circuit, configured to store a program code, wherein the program code comprises:
receiving an incident prediction notification from a network at a first time instant;
determining at least one variable at a second time instant in response to the incident prediction notification, wherein the step of determining the at least one variable at the second time instant comprises maximizing a plurality of posterior probabilities of assigning a plurality of subdata of a grounding data to a plurality of observation functions and a causal structure to generate a causal graph, wherein one of the plurality of posterior probabilities is proportional to Π t=0 T P(w i,t |s t-1 ,C, ƒ i ), s t-1 denotes a plurality of states of a plurality of causal variables at a time instant t-1, T denotes a fifth time instant earlier than a fourth time instant, C denotes the causal structure, ƒ i denotes one of the plurality of observation functions, which is corresponding to an i-th causal variable of the plurality of causal variables, w i,t denotes one of the plurality of subdata, which is corresponding to the i-th causal variable, at a time instant t; and
outputting the at least one variable at a third time instant, wherein the incident prediction notification is used to instruct the variable configuration device to determine the at least one variable to respond to a quality of service (QoS) violation prediction at the fourth time instant, and the third time instant is earlier than the fourth time instant; and
a processing circuit, coupled to the storage circuit and configured to execute the program code stored in the storage circuit.
11. The variable configuration device of claim 10 , wherein the network is configured to predict inability to meet at least one QoS requirement at the fourth time instant according to a continuous time structural equation modeling to output the incident prediction notification.
12. The variable configuration device of claim 10 , wherein the network is configured to compare at least one temporal causal variable at the fourth time instant with the at least one variable at the fourth time instant to perform the QoS violation prediction, and the at least one temporal causal variable constitutes a vector η h (t) which satisfies
η
h
(
t
)
=
e
A
(
t
-
t
0
)
η
h
(
t
0
)
+
A
-
1
[
e
A
(
t
-
t
0
)
-
I
]
ξ
h
+
A
-
1
[
e
A
(
t
-
t
0
)
-
I
]
Bz
h
+
M
∑
u
x
h
,
u
δ
(
t
-
u
)
+
∫
t
0
t
e
A
(
t
-
s
)
GdW
h
(
s
)
,
where A denotes a qualitative matrix, t 0 denotes an initial time instant, I denotes an identity matrix, ξ h denotes a random vector, B denotes a transformation matrix, Z h denotes a time-independent predictor vector, M denotes a coefficient matrix, x h,u denotes a time-dependent predictor vector, u denotes a pulse time instant, W h (s) denotes a random walk vector, and G denotes a lower triangular matrix.
13. The variable configuration device of claim 10 , wherein
at least one temporal causal variable comprises a lowest flow bit rate, a highest flow bit rate, a packet loss rate, a session-aggregate bit rate, a user equipment-aggregate bit rate, a packet delay time, a packet error rate, or a data burst volume, and
the at least one variable comprises a guaranteed flow bit rate, a maximum flow bit rate, a maximum packet loss rate, a session-aggregate maximum bit rate, a user equipment-aggregate maximum bit rate, a packet delay budget, a maximum packet error rate, or a maximum data burst volume.
14. The variable configuration device of claim 10 , wherein the at least one variable is decided at the second time instant according to Bayesian optimization, causal Bayesian optimization, or dynamic causal Bayesian optimization.
15. The variable configuration device of claim 10 , wherein the network is configured to modify at least one QoS parameter or at least one QoS characteristic according to the at least one variable at the fifth time instant earlier than the fourth time instant, and the at least one variable is mapped to a minimum value of a loss function at the fifth time instant.
16. The variable configuration device of claim 10 , wherein the causal structure of the causal graph between a loss function and the plurality of causal variables and the plurality of causal variables of the causal graph are determined together, and the plurality of causal variables comprising the at least one variable are searched out from a plurality of independent variables.
17. The variable configuration device of claim 10 , wherein the plurality of observation functions are obtained based on a causal semantic generative model.
18. The variable configuration device of claim 10 , wherein a loss function at the fifth time instant is a function of the plurality of causal variables at the fifth time instant, the plurality of causal variables a sixth time instant earlier than the fifth time instant, or the loss function at the sixth time instant.