An adaptive bandwidth reservation scheme for 4G cellular networks
using flexible 2-tier cell structure
Chenn-Jung Huang
a,*, Hung-Yen Shen
a, Yi-Ta Chuang
ba
Department of Computer Science and Information Engineering, National Dong Hwa University, Hualien 970, Taiwan b
Department of Computer Science and Information Engineering, National Chiao Tung University, Hsinchu 300, Taiwan
a r t i c l e
i n f o
Keywords: Bandwidth reservation Wireless networks 4G Quality of service Grey prediction theory Particle swarm optimizationa b s t r a c t
Many mechanisms based on bandwidth reservation have been proposed in the literature to decrease con-nection dropping probability for handoffs in cellular communications. The handoff events occur at a much higher rate in packet-switched fourth generation mobile communication networks than in tradi-tional cellular systems. An efficient bandwidth reservation mechanism for the neighboring cells is there-fore critical in the process of handoff during the connection of multimedia calls to avoid the unwillingly forced termination and waste of limited bandwidth in fourth generation mobile communication net-works, particularly when the handoff traffic is heavy. In this paper, an adaptive two-tier scheme, which employs grey prediction theory and swarm intelligence techniques, is proposed to reduce the forced ter-mination probability of multimedia handoffs. The simulation results show that the proposed scheme can achieve superior performance than the representative bandwidth-reserving schemes in the literature when performance metrics are measured in terms of the forced termination probability for the handoffs, the call blocking probability for the new connections and bandwidth utilization.
Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction
With the increasing demand for the provision of multimedia applications, such as Video on Demand (VoD), videoconference, and many WWW-based applications, a great deal of attention is being paid to resource allocation for providing seamless multime-dia access in fourth generation (4G) mobile communication net-works (Huber, 2004; Hui & Yeung, 2003; Jiang & Zhuang, 2004; Schollmeier & Winkler, 2004; Zahariadis, 2003). Since the multi-media applications are very sensitive to the available bandwidth, jitters or delays in the networks, some sorts of service quality guar-antees are desperately needed.
There are two important Quality-of-Service (QoS) parameters considered in wireless networks, namely the handoff call dropping probability (CDP) and new call blocking probability (CBP). Handoff is a mechanism that a mobile host (MH) is transferred from one base station (BS) to another during an ongoing call and the desired bandwidth should be allocated in the new cell in order to provide QoS guarantee for multimedia traffic. The CDP denotes the likeli-hood that an ongoing call is forced to terminate during a handoff process when the allocated resources in the new cell are degener-ated to an unacceptable level, while the CBP represents the possi-bility that a new connection request is denied admission into the
cellular networks. Accordingly, one of the most important QoS is-sues in providing multimedia traffic in wireless networks is to re-duce handoff drops caused by lack of available bandwidth in the new cell while maintaining high bandwidth utilization and low new call blocking rate. In traditional handoffs only signal strength and channel availability are considered, while the following new metrics have been proposed for use in conjunction with signal strength measurements in the envisioned 4G system (McNair & Fang, 2004), such as class of traffic, monetary cost, network condi-tions, include traffic, available bandwidth, network latency, and congestion (packet loss), and mobile node conditions, such as velocity, moving pattern, moving histories, and location informa-tion. The use of the above metrics further increases the complexity of the handoff process and makes the 4G handoff decision more ambiguous.
In recent years, a variety of resource reservation algorithms have been proposed to process handoff in traditional cellular networks (Boumerdassi & Beylot, 1999; Ei-Kadi, Olariu, & Abdel-Wahab, 2002; Kuo, Ko, & Kuo, 2001; Lee, Jung, Yoon, Youm, & Kang, 2000; Lee, Wang, & Tseng, 2001; Levine, Akyildiz, & Naghshineh, 1997; Liu, Bahl, & Chlamtac, 1998; Malla, El-Kadi, & Todorova, 2001; Oliveira, Kim, & Suda, 1998; Wu, Yeung, & Hu, 2000). Among them, Oliviera et al. suggested reserving some bandwidth in the target cells and the neighboring cells at the same time. However, their scheme was unable to adapt to the abrupt oscillation of band-width requirement and bandband-width utilization was deteriorated as well (Oliveira et al., 1998). Levine et al. presented a shadow cluster 0957-4174/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.eswa.2010.02.077
*Corresponding author. Tel.: +886 38633503.
E-mail address:[email protected](C.-J. Huang).
Contents lists available atScienceDirect
Expert Systems with Applications
j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / e s w ascheme to reserve resources with neighboring cells by exchanging information related to the movement pattern and position (Levine et al., 1997). However, the scheme introduces too many communi-cation overheads among the BSs of the cellular system. In Malla et al. (2001), a scheme based on max-min fairness protocol to provide QoS guarantees in wireless multimedia network is proposed. In spite of potentially improving both the CBP and the CDP in this scheme, the users might be subjected to significant bandwidth fluctuations. Lee et al. presented a handoff management scheme using simultaneous multiple bindings that reduces packet loss and generates negligible delays due to handoff in IP-based third-generation cellular systems (Lee et al., 2000). The CDP is probably reduced whereas the bandwidth levels of ongoing multimedia traf-fic are also degraded. Kuo et al. took use of the knowledge of stay-ing time, available time, and the class of the MH to develop a resource semi-reservation scenario and it turns out to be idealistic since the speed of the MH is difficult to detect accurately (Kuo et al., 2001). In Wu et al. (2000), the traffic in a wireless system is first divided into two classes, which are voice calls and video calls, respectively. Then a channel borrowing scheme is proposed to allow voice calls to borrow channels from those pre-allocated to video calls temporarily. Although the CBP for the voice calls is reduced, the issue of improving the CDP during the handoff is not addressed. The work proposed by Ei-Kadi et al. borrowed band-width from multimedia connections for supporting the new calls or handoff connections because multimedia connections can toler-ate and gracefully adapt themselves to transient fluctuations in QoS (Ei-Kadi et al., 2002). The borrowed bandwidth is returned to the original connections as soon as possible to satisfy the QoS requirements. There is 15% of bandwidth reserved exclusively for multimedia handoff connections. Thus, if a new call or handoff re-quests for bandwidth, the scheme inEi-Kadi et al. (2002)tries to borrow bandwidth from other existing connections first. If the bor-rowed bandwidth is insufficient for the request, the connection will try to use the bandwidth in the reservation pool. If there is no enough bandwidth, the connection will be dropped.
Although the above-mentioned resource reservation algorithms more or less decrease the handoff dropping probability, the issue of unfairness is still unresolved. This is unacceptable in the 4G system since some new metrics, such as class of traffic, monetary cost, should be considered in 4G handoff process (McNair & Fang, 2004). The handoff process in the packet-switched 4G system is appar-ently a more critical challenge task than in traditional wireless net-works due to the existence of more bandwidth intensive multimedia applications, client mobility and other new metrics such as monetary cost. In Choi, Kim, Kim, and Kim (2001), Choi proposed a two-tier cell structure which reserved the bandwidth for class 1 traffic only when the MH located in ‘‘Tier-2” as shown inFig 1, and the boundary for the two-tier structure is half of the radius of base station. However, the fixed boundary used for the two-tier structure does not fit the volatile behavior of wireless mo-bile networks. We thus employ two renowned machine learning techniques in this work, i.e., grey prediction theory and particle swarm optimization (PSO), to decide the dynamic boundary of the two-tier structure that suits each individual mobile host and the expected amount of bandwidth used in the neighboring cells for the handoff calls, respectively, in the packet-switched 4G sys-tem so that the CDP of handoff calls can be effectively lowered and the resource can be reserved more efficiently. The motivation of choosing the grey prediction theory and PSO to implement dy-namic two-tier structure and bandwidth reservation scheme in this work is that the grey prediction theory and swarm intelligence have been successfully applied in many areas, such as time series prediction (Chi et al., 1999; Sheu and Wu, 2000; Wen, Lee, & Cho, 2005; Wen et al., 2001), Internet traffic prediction (Hasegawa, Wu, & Mizuno, 2001), decentralized-based routing (
Mavromousta-kis & Karatza, 2004), cell assignment in personal communication services networks (Shyu, Lin, & Hsiao, 2004), and earning predic-tion in the investment decision (Ko & Lin, 2004).
The proposed approaches are compared with the representative bandwidth reservation schemes in the literature, such as Fixed Reservation scheme (FR) and Rate-Based Borrowing scheme (RBB) (Ei-Kadi et al., 2002). The experimental results reveal that our approaches can achieve better performance than that of other bandwidth reservation schemes in terms of CDP and CBP.
The remainder of the paper is organized as follows. A primitive bandwidth reservation scheme for 4G system is introduced in Sec-tion2. In Section3we state how to use grey prediction theory to compute the dynamic boundary of two-tier structure. Then in Sec-tion4we state how to incorporate the particle swarm optimization techniques into the bandwidth-reserving estimator given in Sec-tion2for better performance achievement. Section5is the simu-lation results, which compare the proposed approach with FR and RBB schemes. Conclusions are given in Section6.
2. A primitive adaptive resource reservation scheme
The traffic in cellular networks is usually categorized into the following two classes in the literature (Oliveira et al., 1998). Class I traffic denotes real-time multimedia traffic, such as interactive audio and video, while Class II is non-real-time data traffic, such as images and text. The representative bandwidth reservation schemes in the literature (Ei-Kadiet al., 2002; Kuo et al., 2001; Lee et al., 2000; Levine et al., 1997; Malla et al., 2001; Oliveira et al., 1998; Wu et al., 2000) anticipate that a Class I connection re-quest will make a handoff into one of its neighboring cells in the future and thus try to reserve some bandwidth in surrounding cells before the connection request is admitted. The Class I connection is forced to dropped during handoff if its minimum acceptable band-width requirement cannot be satisfied in the entering cell. As for Class II traffic, a handoff is always accepted as long as there is any free bandwidth available. Although the above-mentioned bandwidth reservation schemes can effectively lower the CDP in traditional macrocell wireless networks, whether they fit the requirement of the new metrics defined for processing 4G multi-media handoff is doubtful. We thus propose a primitive resource reservation scheme in this section to aim at reducing overheads among the BSs and reserving bandwidth in an effective manner, effectively decreasing the CDP for the 4G multimedia handoffs, while keeping bandwidth utilization at a reasonable level.
The amount of the reserved bandwidth is determined by the fol-lowing three factors:
The probability that the MH will move to a neighboring cell will be larger if the neighboring cell is a hot cell.
Tier 2
Tier 1
MS
The current reserved bandwidth for the six neighboring cells. The probability of moving to a neighboring cell is propor-tional to the bandwidth that the neighboring cell reserves. The distance of the MH’s position to the neighboring cells.
There are more chances that the MH will move to the neigh-boring cell that the MH is closer to.
Based upon the above considerations, the bandwidth reserved in cell B for the MH located at cell A when the new connection is accepted as shown inFig. 2, can be derived as follows:
BRB¼ C PB BWMH 1 DB
; ð1Þ
where C is a constant, BWMHdenotes the minimum bandwidth re-quested by a MH at cell A, PBrepresents the probability that the MH moves to cell B, and DBstands for the distance of the MH’s posi-tion to cell B. Notably, The parameters BWMHand PBcan be obtained easily by a simple computation in the base station, and DBcan be acquired based on the location management implemented in the 4G system (Saha, Mukherjee, Misra, Chakraborty, & Subhash, 2004; Zahariadis, 2003).
3. Adaptive two-tier structure using grey prediction theory Grey theory was initiated by J.L. Deng in 1982 (Chen, Lin, Hsu, Ku, & Liu, 2003; Deng, 1989; Fan et al., 2004; Gudmuundson, 1991; Zhao, 2004). Here we call the system ‘‘white system” if all the information of a system is known. On the other hand, if we do not have any information about a system, the system is called ‘‘black system”. Therefore, a grey system is a system which we have only a few information about it. Grey theory points at the sys-tem with uncertain information and provides the relation analysis and model construction. It has been widely applied to various fields including control system, random variable problem. In the grey theory, the grey prediction model is mostly used to predict the behavior of a Grey system. Through the prediction, we can make a decision with the grey system. Since the processing only need a few data to get predictive value with high accuracy, it is very suitable for system with a real-time requirement.
In this work, we adopt the following mobility profile of the MH: The speed of the MH.
The distance between the MH and the radius of base station. Probability to neighbor cell is larger if it is a hot cell.
Since the above-mentioned metrics are all time series, we can use them as the parameters of the grey system, and apply the pre-dictive values of these parameters as the inputs to the following equation so that the boundary for each individual MH can be decided as shown inFig. 3:
Bounadary ¼ Radius Ratio Speed Direction; ð2Þ where Radius is the radius of the base station, Ratio denotes the dis-tance between the MH and the radius of base station, Speed repre-sents the speed of the MH, and Direction stands for the probability to the neighbor cells.
The grey system can be briefly reviewed two import methods: (i) GM (1, 1) Model Construction and (ii) Rolling checking. 3.1. GM (1, 1) Model Construction
Grey prediction based on grey model (GM) has three basic oper-ations. They are the accumulated generating operation (AGO), in-verse accumulated generating operation (IAGO) and grey modeling. The GM(1, l) model is the most commonly used model. According to the Grey Theory, an irregular raw data can be trans-formed to the regular data by using AGO (Gudmuundson,1991). Being processed through AGO, the generated data can be used to construct system model in grey differential equation. GM(1, l) model construction procedure is as described as follow:
Suppose there is one parameter of the system that is intended to be predicted, and the non-negative data sequence are denoted as xð0Þ¼ ðxð0Þð1Þ; xð0Þð2Þ; xð0Þð3Þ; . . . ; xð0ÞðnÞÞ;
8
n 4 ð3Þ and the GM(1, l) model isxð0ÞðkÞ þ aZð1Þ
ðkÞ ¼ b; k ¼ 1; 2; 3; . . . ; n; ð4Þ where a is named as ‘‘develop parameter”, and b is the ‘‘grey input”.
Fig. 2. Hexagonal cellular architecture with a cluster size equal to seven.
A
B
C
D
E
F
Mobile HostStep I: We can get the first order AGO sequence by taking AGO on x(0) and denoted as xð1ÞðkÞ ¼ AGOðxð0ÞðkÞÞ ¼X k i¼0 xð0ÞðiÞ ð5Þ
and the rth order AGO is defined as xðrÞðkÞ ¼X
k
m¼1
xðr1ÞðmÞ; k ¼ 1; 2; 3; . . . ; n; ð6Þ xðrÞðkÞ ¼ xðrÞðk 1Þ þ xðr1ÞðkÞ: ð7Þ Step II: Find Z(1)(k)
Zð1ÞðkÞ ¼ 0:5½xð1ÞðkÞ þ xð1Þðk 1Þ: ð8Þ Step III: Use ‘‘Least Square Method” to find matrix B and vector yn, B ¼ Zð1Þð2Þ 1 Zð1Þð3Þ 1 : : : : Zð1ÞðkÞ 1 2 6 6 6 6 6 6 4 3 7 7 7 7 7 7 5 ; ð9Þ yn¼ ½xð0Þð2Þ; xð0Þð3Þ; xð0Þð4Þ; . . . ; xð0ÞðnÞ: ð10Þ Step IV: The parameters a and b can be derived through follow operation: ^ a ¼ ½a; bT¼ ðBTBÞ1BTyn; ð11Þ where a ¼ Pn k¼2z1ðkÞ Pn k¼2xð0ÞðkÞ ðn 1Þ Pn k¼2zð1ÞðkÞxð0ÞðkÞ ðn 1ÞPnk¼2½zð1ÞðkÞ2 Pn k¼2zð1ÞðkÞ 2 ; ð12Þ b ¼ Pn k¼2½z 1ðkÞ2Pn k¼2xð0ÞðkÞ Pn k¼2zð1ÞðkÞ Pn k¼2zð1ÞðkÞxð0ÞðkÞ ðn 1ÞPnk¼2½zð1ÞðkÞ 2 Pnk¼2zð1ÞðkÞ 2 : ð13Þ Step V: The response equation can be expressed by:
xð1Þðk þ 1Þ ¼ xð0Þð1Þ b a
eakþb
a: ð14Þ
Step VI: x(0)(k + 1) is obtained by xð0Þðk þ 1Þ ¼ ð1 ea Þ xð0Þð1Þ b a eak: ð15Þ 3.2. Rolling checking
Rolling checking method is mostly used to check the precision of the grey model. The procedure is described as follow:
Initially, use the first n data to construct a prediction model, then use the prediction model to predict the value of the (n+1)th data and compare the predicted data to the original (n+1)th data. After that, use the next n original data (from 2 to n+1) to construct another model to predict the (n+2)th data and compare it to the original (n+2)th data. Continue the operation until all the data are predicted. There is an error check equation in the grey system theory. When i = 1, k = 4, 5, 6, . . . , n1, the traditional error equa-tion is
eðk þ 1Þ ¼x
ð0Þðk þ 1Þ ^xð0Þðk þ 1Þ
xð0Þðk þ 1Þ 100%; ð16Þ
where e(k + 1) is the error of the (k + 1) instant of GM(1, l) model, and k+16n. The rolling checking average error (RCAE) of GM(1, l); is defined as: e ¼ 1 n 4 Xn k¼4 keðk þ 1Þk 100%: ð17Þ
Therefore, the rolling checking average precision of GM(1, l) is de-fined as
r
ð1 eÞ 100%: ð18Þ4. Bandwidth reservation mechanism using particle swarm optimization
Particle swarm optimization (PSO) is a computational intelli-gence approach to optimization that is based in the behavior of swarming or flocking animals, such as birds or fish. In the PSO, every individual moves from a given point to a new one which is a weighted combination of the individual’s best position ever found, and of the group’s best position. The PSO algorithm itself is simple and involves adjusting a few parameters. With little mod-ification, it can be applied to a wide range of applications. Because of this, PSO has received growing interest from researchers in var-ious fields.
In this work, we assume that each base station executes its indi-vidual PSO algorithm, and each swarm consists of seven particles (cells) as shown inFig. 1. Recall that a simple relationship between the expected reserved bandwidth and the three input parameters was derived as shown in Eq.(1)in Section2. To improve the accu-racy of this equation, we assume the input/output parameters have the following nonlinear relationship:
BRBðtÞ ¼ C ðPBðtÞÞx1 ðBWMHðtÞÞx2 1 ðDBðtÞÞx3
; ð19Þ
where the value of x1, x2and x3is expected to be determined by the PSO technique. Meanwhile, the fitness function used in the PSO is the handoff call dropping probability (CDP) for multimedia class (Class I) traffic, since the achievement of the low CDP is the main goal of this work.
A summary of the PSO algorithm used in this work is given below:
(1) Initialize the swarm of the particles such that the position ~xijðt ¼ 0Þ of each particle is random within the hyperspace, where j = 1, 2, 3, denote the values of the three exponent parameters as given in Eq.(19), respectively.
(2) Compare the fitness function of each particle, Fð~xijðtÞÞ, which is the CDP of Class I traffic of each individual during current time period, to its best performance thus far, pbestij; if Fð~xijðtÞÞ < pbestijthen
ðiÞ pbestij¼ Fð~xijðtÞÞ; ð20Þ
ðiiÞ ~xpbestij¼ ~xijðtÞ: ð21Þ
(3) Compare Fð~xijðtÞÞ to the global best particle, gbestj if Fð~xijðtÞÞ < gbestjthen
ðiÞ gbestj¼ Fð~xijðtÞÞ; ð22Þ
ðiiÞ ~xgbestj¼ ~xijðtÞ: ð23Þ
~
v
ijðtÞ ¼ ~v
ijðt 1Þ þ c1 r1 ð~xpbestijðtÞ ~xijðtÞÞ þ c2 r2 ð~xgbetjðtÞ ~xijðtÞÞ; ð24Þ where r1and r2are random numbers between 0 and 1, and c1 and c2are positive acceleration constants, which should sat-isfy c1+ c264 as reported in Kennedy (1998).(5) Move each particle to a new position:
ðiÞ ~xijðtÞ ¼ ~xijðt 1Þ þ ~
v
ijðtÞ; ð25ÞðiiÞ t ¼ t þ 1: ð26Þ
(6) Repeat steps(2)–(6)until convergence.
5. Simulation results
A series of simulations are conducted to compare the proposed bandwidth reservation scheme (DTBR) schemes, with the fixed res-ervation scheme (FR), and the scheme without bandwidth reserva-tion (NR). Meanwhile, the rate-based borrowing scheme (RBB) Ei-Kadi et al. (2002)is also compared with the proposed work because it was reported in Ei-Kadi et al. (2002) that the RBB scheme achieves better performance than other representative bandwidth allocation and reservation schemes in the literature, such as the well-known bandwidth reservation scheme presented inOliveira et al. (1998). The RBB scheme not only allows the new calls and handoff connections to borrow bandwidth form existing multime-dia connections, but also reserves 15% of bandwidth exclusively for Class I handoff connections.
In the NR scheme, no bandwidth is reserved for handoff connec-tions in each cell. If there is no bandwidth available when the MH moves to the new coverage area, the handoff call is disconnected and a forced termination occurs. As for the FR approach, a set of channels called guard channels are preserved in each cell to pro-vide a way of prioritizing handing off calls on new call originations by setting aside a fixed bandwidth to support handing off users. New call originations cannot be assigned bandwidth from the guard channel pool. The guard channels are set to 20% of the whole bandwidth for the FR scheme in our simulations. Meanwhile, the
acceleration constants c1and c2 are both set to two in the PSO scheme.
The connections in the simulations are divided into two classes. A Class I traffic, which is a multimedia connection, is allowed to move to a neighboring cell only when the unallocated bandwidth in the target cell exceeds the requested bandwidth. A data connec-tion of Class II traffic can be granted to switch to a neighboring cell as long as the target cell possesses any unused bandwidth. Addi-tionally, a new connection of Class I real-time traffic is allowed to borrow bandwidth from Class II non real-time connections in the same cell if the unallocated bandwidth in the current cell is smaller than the minimum bandwidth that the new Class I traffic requests in the proposed scheme. Similar approach was taken in Wu et al. (2000) and Ei-Kadi et al. (2002)to effectively reduce the new call blocking probability of real-time traffic.
There are 36 cells included in the simulation environment as shown inFig. 4. A total of 30 Mbps bandwidth is allocated in each cell. Both classes of the connections are listed inTable 1. The band-width requirement for each connection is randomly selected with-in the range of the maximum and the mwith-inimum bandwidth requirement listed inTable 1. Both the class and the location of each MH are randomly selected at the initial state. Each MH is gi-ven a speed characteristic, which decides the time spent in a cell, in order to simulate handoffs. If a hot cell neighbors with the cell that a MH is located at, then the MH has a probability of 0.5 to move to the neighboring hot cell, and a probability of 0.1 to one of other neighboring cells. On the other hand, each MH will move to one of the six neighboring cells with equal probability if no neighboring hot cell exists.
35 30 31 32 33 34 35 29 24 25 26 27 28 29 23 18 19 20 21 22 23 17 12 13 14 15 16 17 11 6 7 8 9 10 11 5 0 1 2 3 4 5 35 30 31 32 33 34 35
Fig. 4. Cellular topology with 36 cells in the simulations. Table 1
Multimedia traffic used in the simulations. Traffic class Bandwidth requirement Average call duration (min) Example Class I 30 Kbps 3 Voice service Class I 256 Kbps 5 Video-phone Class I 1–4 Mbps 10 Video service Class II 5–20 Kbps 0.5 E-mail, paging
Class II 64–512 Kbps 3 Remote login & Data on demand
Fig. 5shows the comparison of call dropping probability (CDP) for multimedia handoffs (Class I), andFig. 5illustrates the CDP for combined Class I and II traffic. We can see fromFig. 5that the CDP for multimedia handoffs is the lowest for the proposed scheme (DTBR). Besides, the CDP for combined Class I and II traffic in pro-posed scheme is still lower than the other three schemes as shown inFig. 6. The NR has the worst performance as expected since it does not reserve bandwidth for the handoffs at all. As for the
RBB scheme, although it uses bandwidth borrowing technique to lower down the CDP for handoffs, its fixed bandwidth reservation mechanism is still inferior to the approach of adaptive bandwidth reservation based on the dynamic change of mobile node condi-tions as taken in this work.
Fig. 7shows the CBP for the new multimedia connections in the four schemes, and Fig. 8illustrates the CBP for combined Class I and II traffic. The call blocking probability for the new connections in the PSOBR and the RBB schemes is apparently improved by means of the channel borrowing technique. Meanwhile, the effec-tiveness of adaptive bandwidth reservation contributes to the bet-ter performance achieved in the proposed scheme as illustrated in Figs. 7 and 8. The FR scheme has the highest CBP for new connec-tions because it reserves fixed bandwidth for multimedia handoff connections and results in lessened bandwidth available for new connections.
The bandwidth utilization of various mechanisms is given in Fig. 9. Note that the bandwidth utilization is defined as:
Bandwidth Utilization ¼
P
for each cellUsed bandwidth of each cell P
for each cellMaximum bandwidth of each cell
: ð27Þ
The proposed scheme achieves better performance than the other three schemes in bandwidth utilization due to the efficient usage of adaptive bandwidth reservation mechanism. The RBB scheme uses bandwidth borrowing technique to achieve higher bandwidth utilization than the NR and FR schemes. Bandwidth utilization is the
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.02 0.04 0.1 0.2 0.4 1 2
Call arrival rate (call/sec)
NR FR RBB DTRB
Fig. 5. Call dropping probability for Class I traffic in the four schemes.
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.02 0.04 0.1 0.2 0.4 1 2
Call arrival rate (call/sec)
NR FR RBB DTRB
Fig. 6. Call dropping probability for combined Class I and II handoffs in the four schemes. 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.02 0.04 0.1 0.2 0.4 1
Call arrival rate (call/sec) Class1 CBP
NR FR RBB DTRB
Fig. 7. Call blocking probability for Class I traffic in four schemes.
0 0.1 0.2 0.3 0.4 0.5 0.6 0.02 0.04 0.1 0.2 0.4 1 2
Call arrival rate (call/sec)
NR FR RBB DTRB
Fig. 8. Call blocking probability for combined class I and II traffics in the four schemes. 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0.02 0.04 0.1 0.2 0.4 1 2
Call arrival rate (call/sec)
NR FR RBB DTRB
poorest in the FR scheme since it always reserves fixed bandwidth in each cell which is not necessarily used by the handoffs. 6. Conclusion
In this paper, an adaptive bandwidth reservation scheme is pro-posed to reduce forced termination of multimedia handoffs in the packet-switched 4G systems. Grey prediction theory and particle swarm optimization techniques are employed to compute the dy-namic boundary of a two-tier structure and the amount of reserved bandwidth for the handoffs in the candidate target cells. This work also tries to decrease the call blocking probability of new connec-tions by using a channel borrowing technique. The simulation re-sults show that the proposed scheme performs better then the fixed reservation (FR) scheme, the scheme without reservation (NR), and the rate-based borrowing scheme (RBB) when call block-ing probability for new connections, call droppblock-ing probability for the handoffs, and bandwidth utilization are compared. The pro-posed scheme is proved to be a good choice for the bandwidth res-ervation scheme used for processing multimedia handoffs in 4G systems where the increasing amounts of multimedia connections are expected and the QoS requirements of multimedia traffic need to be maintained persistently during connection time. Subsequent research will investigate the feasibility of applying other intelligent tools such as neuro-fuzzy and genetic algorithms into the proposed scheme to improve the accuracy of the motion prediction for the MH.
Acknowledgment
The authors would like to thank the National Science Council of the Republic of China, Taiwan for financially supporting this re-search under Contract No. NSC 96-2628-E-259-022-MY3 and NSC 97-2218-E-259-005.
References
Boumerdassi, S., & Beylot, A.-L. (1999). Adaptive channel allocation for wireless PCN. Mobile Networks and Applications, 4, 111–116.
Chi, S. C., Chen, H. P., & Cheng, C. H. (1999). A forecasting approach for stock index future using grey theory and neural networks. In International joint conference on neural networks (Vol. 6, pp. 3850–3855).
Chen, L. R., Lin, C. H., Hsu, R. C., Ku, B. G., & Liu, C. S. (2003). A study of Li-ion battery charge forecasting using Grey theory. In The 25th international telecommunications energy conference (pp. 744–749).
Choi, C. H., Kim, M. I., Kim, T. J., & Kim, S. J. (2001). A call admission control mechanism using MPP and 2-tier cell structure for mobile multimedia computing. In: Tenth international conference on computer communications and networks (pp. 581–584).
Deng, J. L. (1989). Introduction to grey system theory. The Journal of Grey System, 1(l), 1–24.
Ei-Kadi, M., Olariu, S., & Abdel-Wahab, H. (2002). Rate-based borrowing scheme for QoS provisioning in multimedia wireless networks. IEEE Transactions on Parallel and Distributed Systems, 13(2), 156–166.
Fan, S. H., Xiao, T. Y., Guo, C., & Wen, J. (2004). A model of the expanded grey theory. In IEEE international conference on networking, sensing and control (Vol. 1, pp. 339–342).
Gudmuundson, M. (1991). Analysis of handover algorithms. In Proceedings of Vehcular Tech. Conf. ‘91 (pp.537–542). St. Louis, MO.
Hasegawa, M., Wu, G., & Mizuno, M. (2001). Applications of nonlinear prediction methods to the internet traffic. In The 2001 IEEE international symposium on circuits and systems (Vol. 2, pp. III-169III-172).
Huber, J. F. (2004). Mobile next-generation networks. IEEE Multimedia, 11(1), 72–83. Hui, S. Y., & Yeung, K. H. (2003). Challenges in the migration to 4G mobile systems.
IEEE Communications Magazine, 41(12), 54–59.
Jiang, H., & Zhuang, W. (2004). Quality-of-service provisioning in future 4G CDMA cellular networks. IEEE Wireless Communications, 11(2), 48–54.
Kennedy, J. (1998). The behavior of particles. In Seventh international conference on evolutionary programming (pp. 581–589).
Ko, P. C., & Lin, P. C. (2004). A hybrid swarm intelligence based mechanism for earning forecast. In Second international conference on information technology for application (pp. 193–198).
Kuo, G.-S., Ko, P.-C., & Kuo, M.-L. (2001). A probabilistic resource estimation and semi-reservation scheme for flow-oriented multimedia wireless networks. IEEE Communications Magazine, 135–141.
Lee, J.-H., Jung, T.-H., Yoon, S.-U., Youm, S.-K., & Kang, C.-H. (2000). An adaptive resource allocation mechanism including fast and reliable handoff in IP-Based 3Gwireless Networks. IEEE Personal Communications, 42–47.
Lee, G.-C., Wang, T.-P., & Tseng, C.-C. (2001). Resource reservation with pointer forwarding schemes for mobile RSVP. IEEE Communications Letters, 5(7), 298–300.
Levine, D. A., Akyildiz, I. F., & Naghshineh, M. (1997). A resource estimation and call admission algorithm for wireless multimedia networks using the shadow cluster concept. IEEE/ACM Transactions Networking, 5(1), 1–12.
Liu, T., Bahl, P., & Chlamtac, I. (1998). Mobility modeling, location tracking, and trajectory prediction in wireless ATM networks. IEEE Journal of Selected Areas in Communications, 16(6), 922–936.
Malla, A., El-Kadi, M., & Todorova, P. (2001). A fair resource allocation protocol for multimedia wireless networks. In IEEE international conference parallel processing (pp. 437–443).
Mavromoustakis, C. X., & Karatza, H. D. (2004). Agent-based throughput response in presence of node and/or link failure (on demand) for circuit switched telecommunication networks. Computer Communications, 27(3), 230–238. McNair, J., & Fang, Z. (2004). Vertical handoffs in fourth-generation multinetwork
environments. IEEE Wireless Communications, 11(3), 8–15.
Oliveira, C., Kim, J. B., & Suda, T. (1998). An adaptive bandwidth reservation scheme for high-speed multimedia wireless networks. IEEE Journal of Selected Areas in Communications, 16(6), 858–874.
Saha, D., Mukherjee, A., Misra, I. S., Chakraborty, M., & Subhash, N. (2004). Mobility support in IP: A survey of related protocols. IEEE Network, 18(6), 34–40. Schollmeier, G., & Winkler, C. (2004). Providing sustainable QoS in next-generation
networks. IEEE Communications Magazine, 42(6), 102–107.
Sheu, S. T., Wu, C. C. (2000). Using grey prediction theory to reduce handoff overhead in cellular communication systems. In The 11th IEEE international symposium on personal, indoor and mobile radio communications (Vol. 2, pp. 782– 786).
Shyu, S. J., Lin, B. M. T., & Hsiao, T. S. (2004). An ant algorithm for cell assignment in PCS networks. In Networking, IEEE international conference on sensing and control (Vol. 2, pp. 1081–1086).
Wen, K. L., Dai, J. W., Chang, T. C., & Tong, C. C. (2001). The discussions of class ratio for AGO algorithm in grey theory. In IEEE international conference on systems, man, and cybernetics (Vol. 1, pp. 7–11).
Wen, Y. H., Lee, T. T., & Cho, H. J. (2005). Hybrid models toward traffic detector data treatment and data fusion. IEEE Networking, Sensing and Control, 525–530. Wu, X., Yeung, K. L., & Hu, J. (2000). Efficient channel borrowing strategy for
real-time services in mulreal-timedia wireless networks. IEEE Transactions in Vehicular Technology, 49(4), 1273–1284.
Zahariadis, T. (2003). Trends in the path to 4G. Communications Engineer, 1(1), 12–15.
Zhao, Q., Gao, J., Wu, T., & Lu, L. (2004). The grey theory and the preliminary probe into information acquisition technology. In International conference on information acquisition (pp. 402–404).
