Resolution method intent-based network resource conflicts and apparatus thereof
Disclosed are a resolution method for intent-based wireless network resource conflicts and an apparatus thereof. The method includes the following steps: periodically acquiring configured resource information about a current wireless network; pre-allocating resources required for a newly-added intent; and determining whether a pre-allocated resource configuration policy is incompatible with a resource configuration policy being executed in a current network. The apparatus includes: an interface module; a solution module; a feedback module; and a storage module. Through the relative priority of intent requirements, a bias of an implementation satisfaction degree of a network configuration policy is determined, and network resource conflicts of intent requirements with different priorities are solved.
1 . A resolution method for intent-based wireless network resource conflicts, comprising the following steps:
step 1: periodically acquiring configured resource information about a current wireless network and an intent performance state in the current network;
step 2: pre-allocating resources required for a newly-added intent when the newly-added intent occurs in the wireless network, and generating a pre-allocation resource configuration policy,
determining whether the pre-allocation resource configuration policy is incompatible with a resource configuration policy being executed in the current network according to the configured resource information about the wireless network and the intent performance state in the current network,
resource conflicts existing in the wireless network resource configuration policy if the pre-allocation resource configuration policy is incompatible with the resource configuration policy being executed in the current network, and network resources being redeployed according to stored historical intent requirement information and the distribution of resources in the wireless network; and
step 3: issuing a resource pre-allocation policy to be executed in a radio access network (RAN) if a current allocation resource configuration policy is compatible with the resource configuration policy being executed in the current network, and collecting and storing an execution state of a current wireless network resource configuration policy and latest intent performance state data, and
a number of original intents in the wireless network is , and an expression of a preset intent satisfaction degree function is as follows:
S
R
(
H
n
(
t
)
)
=
H
n
(
t
)
-
H
n
,
worst
H
n
,
best
-
H
n
,
worst
,
(
1
)
H n (t) representing a multidimensional performance of a certain intent n at a time t, and H n,best and H n,worst representing best and worst multidimensional performance function values of the certain intent n;
an expression of the multidimensional performance is as follows:
H
n
(
t
)
=
F
n
[
r
V
s
,
n
(
t
)
,
R
s
(
·
)
,
a
s
n
]
,
(
2
)
F n [⋅] representing a multidimensional performance function of the certain intent n, R V s,n (t) representing a corresponding network resource vector allocated to the intent n for a performance metric s, R s (⋅) representing a specific performance numerical function corresponding to the performance metric s, and
a
s
n
representing a preset weight of the intent n to the performance metric s, the greater the weight, the higher the importance of the performance metrics to the intent n.
2 . The resolution method for intent-based wireless network resource conflicts according to claim 1 , wherein in step 2, an acted network area range and an executed lifecycle by a pre-allocation policy of the newly-added intent are compared with corresponding data information of a network configuration policy being executed in a network, and if an overlapped acted network area range and/or an overlapped executed time period do not exist, the pre-allocation policy of the newly-added intent is issued to be executed in a low-level network infrastructure through network management and service technologies, and
otherwise, a satisfaction degree of a user in the network after executing a pre-allocation configuration policy is predicted using a satisfaction degree function by combining the pre-allocation configuration policy of the newly-added intent with parameter values and wireless network state information required for the network configuration policy being executed in the network,
the intent referring to a desired state or effect of the RAN, being in a natural language form or being a combination of relevant keywords, one intent being capable of serving one or more users, and a current intent referring to an intent to implement execution in the RAN.
3 . The resolution method for intent-based wireless network resource conflicts according to claim 1 , wherein in step 1,
an expression of the multidimensional performance function F n [⋅] of the intent n is as follows:
F
n
[
r
V
s
,
n
(
t
)
,
R
s
(
·
)
,
a
s
n
]
=
∑
u
⊂
U
n
[
a
1
n
R
1
(
r
V
1
,
u
(
t
)
)
R
1
max
+
a
2
n
R
2
(
r
V
2
,
u
(
t
)
)
R
2
max
+
,
…
,
+
a
s
n
R
s
(
r
V
s
,
u
(
t
)
)
R
S
max
]
s
·
U
n
,
(
3
)
U n representing a set of user equipment u belonging to the intent n in the wireless network,
a
1
n
,
a
2
n
,
…
,
a
s
n
representing a preset weight of the intent n to a performance metric 1, 2, . . . , s, r V s,n (t) representing a corresponding network resource vector allocated to the intent n for the performance metric s, and
R
1
max
,
R
2
max
,
…
,
R
s
max
representing a maximum theoretical performance value of the performance metric 1, 2, . . . , s in the wireless network, the higher the value of R s (⋅), the lower the performance metric s, corresponding R s (⋅) and
R
s
max
in the formula taking reciprocal forms.
4 . The resolution method for intent-based wireless network resource conflicts according to claim 1 , wherein step 2 specifically comprises:
S 201 : comparing the acted network area range and the executed lifecycle by the pre-allocation policy required for the newly-added intent with the corresponding data information of the network configuration policy being executed in the network, and determining whether the overlapped acted network area range and/or the overlapped executed time period occur, step S 202 being executed if so and step 3 being executed if not;
S 202 : determining whether the pre-allocation policy of the newly-added intent will result in an intent satisfaction degree in the network being equal to/lower than a satisfaction degree threshold value, step 203 being executed if so and step 3 being executed if not;
S 203 : canceling a pre-allocation policy estimated to cause multidimensional resource conflicts, and redeploying the network resources according to the stored historical intent requirement information and the distribution of resources in the wireless network;
S 2041 : establishing a multi-intent Stackelberg game model to perform intent-level coarse-grained network resource deployment according to an intent priority difference and a behavior of network resource competition in the wireless network;
S 2042 : dynamically adjusting network resources allocated to each user under the intent according to a bias difference of the user to different performance metrics under the intent, and performing user-level fine-grained network resources deployment; and
S 2043 : determining whether a redeployment policy of the network resources results in the intent satisfaction degree in the network being equal to/lower than the satisfaction degree threshold value, step S 2041 being executed if not and step 3 being executed if so.
5 . The resolution method for intent-based wireless network resource conflicts according to claim 4 , wherein in step S 2041 , the intent-level coarse-grained network resource deployment is modeled using the Stackelberg game model, three basic elements are defined: a participant set, a policy set of a participant and a utility set of the participant, and an expression of a game problem G is as follows:
G
=
[
N
,
(
X
i
)
i
∈
N
,
(
U
i
)
i
∈
N
]
,
(
4
)
N={1, 2, . . . , n} representing a participant set in the game problem, namely, a set of intents to participate in resource competition, X representing a policy set of a participant, namely, a policy set for acquiring network resources for the intents to participate in resource competition, and U representing a utility function of the participant, a metric being configured to measure benefits gained from the intents to participate in the game;
an expression of a utility function for a follower low priority intent participating in the game is as follows:
U
n
(
r
V
s
,
n
(
t
)
)
=
F
n
[
r
V
s
,
n
(
t
)
,
R
s
(
·
)
,
a
s
n
]
-
∑
s
∈
S
∑
u
∈
U
n
p
V
s
N
l
r
V
s
,
u
(
t
)
,
n
∈
N
l
,
(
5
)
an expression of a utility function for a leader high priority intent participating in the game is as follows:
U
n
(
r
V
s
,
n
(
t
)
)
=
F
n
[
r
V
s
,
n
(
t
)
,
R
s
(
·
)
,
a
s
n
]
-
∑
s
∈
S
∑
u
∈
U
n
p
V
s
N
h
r
V
s
,
u
(
t
)
,
n
∈
N
h
,
(
6
)
U n representing a set of user equipment u belonging to an intent n in the wireless network, F n [⋅] representing a multidimensional performance function of a certain intent n, r V s,n (t) representing a corresponding network resource vector allocated to the intent n for a performance metric s, R s (⋅) representing a specific performance numerical function corresponding to the performance metric s,
a
s
n
representing a preset weight of the intent n to the performance metric s, S being an intent performance metric set, p V s representing unit pricing factors for resources Vs belonging to different dimensions,
p
V
s
N
h
and
p
V
s
N
l
representing unit resource prices of a high priority intent and a low priority intent, N l representing a set of high priority network intents, and N h representing a set of low priority network intents;
expressions of coarse-grained resource allocation problems for the leader and follower are as follows:
max
U
n
(
r
V
s
,
u
(
t
)
)
,
n
∈
N
h
or
N
l
C
1
:
∑
s
∈
S
∑
u
∈
U
n
r
V
s
,
u
(
t
)
≤
V
S
C
2
:
∑
u
∈
U
n
r
V
s
,
u
(
t
)
≤
r
V
s
,
u
(
t
)
subject
to
:
C
3
:
R
s
(
r
V
s
,
u
(
t
)
)
≥
R
s
,
u
thr
or
R
s
(
r
V
s
,
u
(
t
)
)
≤
R
s
,
u
thr
,
s
∈
T
C
4
:
a
s
n
∈
[
0
,
1
)
C
5
:
Σ
s
∈
S
a
s
n
=
1
,
(
7
)
a constraint condition C1 representing a constraint on the number of network resources provided by the network, V s representing corresponding multidimensional network resources available for deployment for the performance metric s, a constraint condition C2 representing a constraint on the number of network resources allocated to each user, a constraint condition C3 representing that a threshold value performance metric must be greater than a preset threshold value,
R
s
,
u
thr
representing a corresponding minimum threshold value of the user u for the performance metric s, T being a threshold value performance metric set, a constraint condition C4 representing a weight of the performance metric, and a constraint condition C5 representing that a sum of weights of the performance metrics is limited; and
an expression of an optimal solution of coarse-grained resource allocation based on Stackelberg game equilibrium is as follows:
U
n
r
V
s
,
u
*
(
t
)
≥
U
n
r
V
s
,
u
(
t
)
∀
r
V
s
,
u
(
t
)
,
n
∈
N
h
or
N
l
,
(
8
)
r* V s,u (t) representing an optimal coarse-grained resource allocation policy of the wireless network to the intent n.
6 . The resolution method for intent-based wireless network resource conflicts according to claim 4 , wherein in step S 2042 , a user-level fine-grained network resource deployment problem is described as a multi-objective optimization problem with O+T optimization objectives, and an expression is as follows:
max
u
∈
U
n
,
r
V
s
,
n
(
t
)
[
R
1
(
r
V
1
,
u
(
t
)
)
,
R
2
(
r
V
2
,
u
(
t
)
)
,
…
,
R
T
(
r
V
T
,
u
(
t
)
)
,
R
1
+
T
(
r
V
1
+
T
,
u
(
t
)
)
,
R
2
+
T
(
r
V
2
+
T
,
u
(
t
)
)
,
…
,
R
O
+
T
(
r
V
O
+
T
,
u
(
t
)
)
]
(
9
)
subject
to
:
C
1
,
C
2
:
∑
u
∈
U
n
r
V
s
,
u
(
t
)
≤
r
V
s
,
u
(
t
)
C
3
:
R
s
(
r
V
s
,
u
(
t
)
)
≥
R
s
thr
or
R
s
(
r
V
s
,
u
(
t
)
)
≤
R
s
thr
,
s
∈
T
,
U n representing a set of user equipment u belonging to an intent n in the wireless network, r V s,n (t) representing a corresponding network resource vector allocated to the intent n for a performance metric s, R 1 (⋅), R 2 (⋅), . . . , R O+T (⋅) representing a specific performance numerical function corresponding to a performance metric 1, 2, . . . , O+T, R s (⋅) representing a specific performance numerical function corresponding to the performance metric s,
R
s
thr
representing a corresponding minimum threshold value of the performance metric s, O being an optimized performance metric set, T being a threshold value performance metric set, S being an intent performance metric set, O+T=S, a constraint condition C2 representing a constraint on the number of network resources allocated to each user, and a constraint condition C3 representing that a threshold value performance metric must be greater than a preset threshold value.
7 . The resolution method for intent-based wireless network resource conflicts according to claim 6 , wherein the multi-objective optimization problem is solved by a low complexity solution based on a weighted sum, and the multi-objective optimization problem is converted into:
max
u
∈
U
n
,
r
V
s
,
n
(
t
)
∑
s
∈
O
⋃
T
a
s
n
R
s
(
r
V
s
,
u
(
t
)
)
subject
to
:
C
1
∼
C
3
;
C
4
:
a
s
n
∈
[
0
,
1
)
C
5
:
∑
s
∈
S
a
s
n
=
1
,
(
10
)
a
s
n
representing a preset weight of an intent n to a performance metric s, a constraint condition C4 representing a weight of the performance metric, and a constraint condition C5 representing that a sum of weights of the performance metrics is limited.