Beruflich Dokumente
Kultur Dokumente
i
x
Input
Layer
(k)
M
Training Process
=Forward Propagation
+Back Propagation
Concealed
Layer
M
(p)
Output
Layer
(1)
i k
x
1 i
x
1 ( ) i k
x
M
History Data: m
(i<=m)
Forward
Propagation
Back
Propagation
i
i
T
Inter-network Resource Allocation
A. Traffic Flow Prediction Model
The neural network model
eural etwork
Original
Data
Forecasting
Data
m
x
2 ( ) m k
x
1
m
x
+
1 ( ) m k
x
Input
Layer
Concealed
Layer
Output
Layer
M
M M
Prediction
Process
(k) (p) (1)
Inter-network Resource Allocation
B. Maximum etwork Resource Utility
In each resource allocation process, we use a utility function of
throughput
Where is the utility of network i for an allocated bandwidth of b
to connection x. Parameters w and are constants indicating the
scale and the shape of the utility function.
( )
,
log( )
con
i x
U w b =
1
B
l
2
m
B
m
2
B
l
1
B
c
2
B
c
3
m
4
m
5
m
2
c
4
c
( )
,
con
i x
U
2 3 1 c
c c B + =
4 5 2 c
c c B + =
Inter-network Resource Allocation
The total utility of the entire network can be obtained from
( ) ( )
( ) ( )
2 2 3 3 1
1
1 2 3
1 2 3
4 4 5 5 2
4 5
4 5
log log log
( , )
log log
l
tol i i
l
m c m c B
m
U m c w
m c m c B
+ + + | |
+ +
|
\
=
+ + +
+ +
(1)
To maximize total utility, an optimization problem is
formulated as follows:
Maximize :
( , )
tol i i
U m c
Subject to :
5
2 3 4 5 2
1
, ,
i m cl c
i
m B c c B c c B
=
= + = + =
1 2
2 2 3 4 5 2 2
1 2
3 2 3 4 5 3 1 2 1
1 4
4 2 3 4 5 4 4
5 1
5 2 3 4 5 5 2 4 2
3 2
2 2 2 3 1 2 1
0
0
0
0
0
total
m
total
m c l
total
m
total
m c l
total
c l
U
m B m m m m m c
U
m B m m m m m B c B
U
m B m m m m m c
U
m B m m m m m B c B
U
c m c m B c B
= =
+
= =
+ +
= =
+
= =
+ +
= =
+ + +
5 4
4 4 4 5 2 4 2
0
total
c l
U
c m c m B c B
= =
+ + +
= =
( )
L t L L = +
Exp CDP=2%,
CDP=3%
2.3 L t t u = +
0.5 L t t u = +
Access Network Selection
Based on the current available resource for handoff
connections or new connections, access network selection will
be done.
When a connection arrivals, it will select a RAN which has the
maximum residual resource according to current network
resource allocation and reservation. If no RANs residual
resources are satisfied its requirement, this connection is
rejected by the system.
Evaluation of proposed solution
1 2 3 4 5
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
N
e
w
A
r
r
i
v
a
l
C
a
l
l
D
r
o
p
p
o
i
n
g
P
r
o
b
a
b
i
l
i
t
y
Areas in System
= 14, CRRM
= 14, RRM
= 28, CRRM
= 28, RRM
= 42, CRRM
= 42, RRM
1 2 3 4 5
0.000
0.002
0.004
0.006
0.008
0.010
0.012
0.014
0.016
0.018
0.020
0.022
0.024
0.026
H
a
n
d
o
f
f
C
a
l
l
D
r
o
p
p
i
n
g
P
r
o
b
a
b
i
l
i
t
y
Area in System
= 14, CRRM
= 14, RRM
= 28, CRRM
= 28, RRM
= 42, CRRM
= 42, RRM
Evaluation of proposed solution
5
6
7
8
9
10
11
12
13
14
15
5
6
7
8
9
10
11
12
13
14
15
0.5
0.6
0.7
0.8
0.9
1
0
5
10
15
20
30
40
50
60
0
0.5
1
1.5
2
2.5
3
3.5
4
x 10
4
Resource utility v.s call arrival rate Total users in system
Conclusions
How to increase the system resource utility is a
challenging problem. Based on the prediction of the
handoff call arrival rate and the variation of traffic in
different hours, we present a new resource management
approach coordinating with IEEE1900.4 architecture ,
which can reserve system resource dynamically and re-
allocat for each RAN, so that the performance of systems
is increased. .
22
T
T
h
h
a
a
n
n
k
k
s
s
!
!