Simulation results and discussions

Suppose that one call request can only connect to one access network at a time here. For each cell, assume the new call arrival rate of conversational, streaming, interactive, and background traffic class calls in the heterogeneous network are

1/ 40

AR , AR1/120, AR1/120, and AR1/ 240 (users/second), respectively, where AR is the equivalent arrival rate. In the simulation, AR is chosen from 1, 3, 5, 7, and 9. Besides, there are three algorithms, BGNAS (proposed), BGNAS^-, and UGT in the simulation. BGNAS is the proposed method which uses bargaining game and the mobility model that is mentioned in chapter 2 has an improvement. BGNAS^- also uses bargaining game but the mobility model is the same as UGT.

41

Fig. 5.1 : New call blocking rate

Fig. 5.1 shows the new call blocking rate. It can be found that BGNAS and BGNAS^- have a little lower new call blocking rate than UGT generally. The difference between BGNAS and BGNAS^- is the improvement of the mobility model.

If the values of x_{th i,} are the same in BGNAS and BGNAS^-, new call blocking rate will be higher in BGNAS^- than in BGNAS because of more precise estimation.

Besides, in BGNAS and BGNAS^-, the design of bargaining power can reinforce the balance of networks loading, which means it can accommodate more call requests for the whole system. By bargaining problem formulation, it can be found an optimal solution according to different bargaining game, and then choose the best network when compare these solutions. Because choosing the maximum value among these optimal solutions means choosing the network which can lead to the maximum benefit for the user and the network.

42 networks. If UGT wants to add the perspective of the user, a linear combination might not be an optimize way, which means the user might choose an inappropriate network for different case.

43

Fig. 5.3 can also be explained by this phenomenon. It can be found that BGNAS and BGNAS^- have higher throughput than UGT especially when equivalent arrival will accept a non-real time services without transmission. By this way, BGNAS and BGNAS^- can achieve higher average throughput than UGT.

1 2 3 4 5 6 7 8 9

Fig. 5.3 : Total throughput of the system and throughput of each network

44

Fig. 5.4 and Fig. 5.5, which depict the number of total handoff calls and the number of failed handoff calls, respectively. We can see BGNAS has much fewer total handoff calls and number of failed handoff calls than BGNAS^- and UGT. This is because BGNAS^- and UGT do not consider the effect if the direction of the MS will change during its holding time. Besides, the number of total handoff calls in BGNAS -is a little lower than UGT. Th-is -is because BGNAS^- considers the dwell time constraint when decide the candidate network set. By this way, BGNAS^- can avoid to choose any network in the candidate network set which those estimated dwell time of candidate networks are all too short.

1 2 3 4 5 6 7 8 9

0 200 400 600 800 1000 1200

Equivalent arrival rate (AR)

Number of total handoff calls

BGNAS BGNAS-UGT

Fig. 5.4 : Number of total handoff calls

45

1 2 3 4 5 6 7 8 9

0 10 20 30 40 50 60 70 80

Equivalent arrival rate (AR)

Number of failed handoff calls

UGT BGNAS

BGNAS-Fig. 5.5 : Number of failed handoff calls

The average delay for voice and video call in the heterogeneous network are shown in Fig 5.6 (a) and 5.6 (b), respectively. It can be found that the average delay for voice call and video call of BGNAS and BGNAS^- are a little higher than UGT in CDMA and WMAN. That is because BGNAS and BGNAS^- will guarantee transmission rate for non-real time services. This will cause the delay of real time services to be higher. But BGNAS and BGNAS^- are a little lower than UGT in WLAN. This is because the number of calls in WLAN with BGNAS and BGNAS^- are fewer than UGT and the probability of each call to get the right to access will increase.

Besides, the average delay for voice and video calls are much lower than the QoS requirements.

46

Fig. 5.6 (a) : Average delay of voice traffic

1 2 3 4 5 6 7 8 9

Fig. 5.6 (b) : Average delay of video traffic

47 Average dropping rate of Voice traffic (%) BGNAS: WCDMA

BGNAS: WMAN

Fig. 5.7 (a): Average dropping rate of voice traffic

1 2 3 4 5 6 7 8 9

Average dropping rate of video traffic (%)

BGNAS: WCDMA

Fig. 5.7 (b) : Average dropping rate of video traffic

48

The average dropping rate for voice and video call are shown in Fig. 5.7 (a) and Fig. 5.7 (b), respectively. We can see the dropping rates are higher in BGNAS and BGNAS^- than that in UGT in CDMA and WMAN. The reason is the same as that for delay. But in WLAN, if the number of calls is higher, the probability that real time services to get the right to transmit the packet will be decrease. This will cause larger delay variance and dropping rate. Besides, the average dropping rate for voice and video call for both of two schemes are much lower than the QoS requirements.

49

Chapter 6 Conclusions

In this thesis, a bargaining game based network access selection (BGNAS) is proposed for heterogeneous wireless environment, which considers conversational, streaming, interactive, and background services. A candidate network set will be found first by checking three constraints which include signal strength constraint, network loading constraint, and dwell time constraint. These candidate networks will form several two-person bargaining games. One player in one bargaining game is the user and the other player is one of the candidate networks. Two preference functions are designed which consider QoS requirements, mobility, and loading balance to represent the degree of preference of one player to the other player. The bargaining power is considered to reinforce the balance of system loading. By the bargaining problem formulation, there exists a unique solution in one bargaining game.

Comparing the solutions of all bargaining games, one network will be selected with maximum value of all bargaining problems.

Simulation results show that BGNAS has higher total throughput than UGT at high arrival rate especially while satisfying the QoS requirements of each traffic class.

This result comes from BGNAS gives more resource for non-real time services as far as possible. By sacrificing a little packet delay and packet dropping rate, BGNAS can achieve more system throughput. Besides, BGNAS reduces the number of handoffs calls more than 50% than UGT and without increasing the number of failed handoff

50

calls. Because if a MS will change the direction when they move, the dwell time estimation will not be the same as the direction would not change. In this case, the overhead during the processing of handoff calls can be avoided significantly. When it comes to the packet delay and packet dropping rate, BGNAS is higher than UGT. But these two schemes are all under the maximum acceptable packet delay and packet dropping rate.

51

Appendix

Followings are proofs of the three axioms mentioned in Chapter 3 for the proposed bargaining problem and the two-person bargaining games.

Axiom 1. Pareto efficiency (PAR):

Let ( , )U d be a bargaining problem, and let ( , )v v_{1 2} and (v v₁ , )₂ be the members of U . If v₁v₁ and v₂ v₂, then the bargaining solutions does not assign

1 2

(v v , ) to ( , )U d . Proof:

Let U be the set of pairs of payoffs to agreements, and (UP NP, ), (UP NP, ) be the members of U . If UP UP and NPNP, then we can obtain that UP^1NP^ (UP)^1-(NP)^ since  is the same in the bargaining game.

Axiom 2. Invariance to equivalent payoff representations (INV):

Let ( , )U d be a bargaining problem, let _i and _i be two numbers with

i 0

  for i1, 2, let U be the set of all pairs (v v₁ , )₂ , where v_i _{i i}v _i for 1, 2

i and ( , )v v is a member of U , and let _{1 2} d (_{1 1}d  ₁, ₂d₂₂). If the bargaining solution assigns ( , )v v to _{1 2} ( , )U d , then it assigns (v v₁ , )₂ to (U d , ). Proof:

Let ( , )U d be the proposed bargaining problem, d0 is the pair of payoffs to disagreement. Let  _u, _n and  _u, _n be four numbers with

52

, 0

u n

   , and U be the set of all pairs (UP NP, ) , where

u u

UP UP and NP _nNP _n. (UP NP, ) is a member of U . Let d (_ud_u  _u, _nd_n_n), then

1-

1-(UPd_u) ^(NPd_n)^ (_uUP _u _ud_u _u) ^(_nNP _n _nd_n_n)^,

so the solution of (U d , ) is the same as ( , )U d .

Axiom 3. Independence of irrelevant alternatives (IIA):

Let ( , )U d and (U d , ) be bargaining problems for which U is a subset of U and dd. If the bargaining solution assigned to ( , )U d is in U, then the bargaining solution assigns the same agreement to (U d , ).

Proof:

Let ( , )U d and (U d , ) be bargaining problems for which U is a subset of U and dd. Since any two of outcomes in U are mutually independent, any reduction of the number of outcomes will not affect the result.

( )1- ( )

u n UP du ^ NP dn ^

 

   

53

Bibliography

[1] F. Bari and V. C. M. Leung, “Automated network selection in a heterogeneous wireless network environment,” IEEE Network Mag., vol. 21, no. 1, pp. 34 - 40, Jan.-Feb. 2007.

[2] Q. Song and J. Abbas, “Network selection in an integrated wireless LAN and UMTS environment using mathematical modeling and computing techniques,”

IEEE Trans. Wireless Commun., vol.12, pp. 42-48, Jun. 2005.

[3] H. J. Wang, R. H. Katz, and J. Giese, “Policy-enabled handoffs across heterogeneous wireless networks,” IEEE WMCSA, pp. 51–60, 1999.

[4] A.Iera, A. Molinaro, C. Campolo, and M. Amadeo, “An access network selection algorithm dynamically adapted to user needs and preferences,” IEEE PIMRC, pp.1 – 5, Sept. 2006.

[5] W. Luo and E. Bodanese, “Radio access network selection in a heterogeneous communication environment,” IEEE WCNC, pp. 1-6, Apr. 2009.

[6] Y. H. Chen, N. Y. Yang, C. J. Chang, and F. C. Ren, “A utility function-based access selection method for heterogeneous WCDMA and WLAN networks,”

IEEE PIMRC, pp. 1-5, Sept. 2007.

[7] G. Ning, G. Zhu, L. Peng, and X. Lu, “Load balancing based on traffic selection in heterogeneous overlapping cellular networks,” IEEE/ IFIP CANET, Sept. 2005.

[8] G. Ning, G. Zhu, Q. Li, and R. Wu, “Dynamic load balancing based on sojourn time in multitier cellular systems,” IEEE VTC, vol. 1, pp. 111-116, 2006.

54

[9] J. Antoniou and A. Pitsillides, “4G converged environment: modeling network selection as a game,” IEEE IST, pp. 1-5, Jul. 2007.

[10] D. Charilas, O. Markaki, and E. Tragos, “A theoretical scheme for applying game theory and network selection mechanisms in access admission control,” IEEE ISWPC, pp. 303-307, May 2008.

[11] M. Cesana, I. Malanchini, and A. Capone, “Modelling network selection and resource allocation in wireless access networks with non-cooperative games,”

IEEE 5th MASS, pp.404 - 409, Sept. 29 -Oct. 2, 2008.

[12] D. Niyato, and E. Hossain , “Dynamics of network selection in heterogeneous wireless networks: an evolutionary game approach,” IEEE Trans. Vehicular Technology, vol. 58, pp. 2008-2017, May 2009.

[13] D. Niyato and E. Hossain, “A noncooperative game-theoretic framework for radio resource management in 4G heterogeneous wireless access networks,”

IEEE Trans. Mobile Computing, vol. 7, no. 3, Mar. 2008.

[14] T. L. Tsai, “Utility function and game-theory based network selection scheme in heterogeneous wireless networks”, 2008.

[15] 3GPP TS 23.271, “Functional stage 2 description of Location Services (LCS),”

3^rd Generation Partnership Project, Tech. Spec., Sept. 2009.

[16] H. Holma and A. Toskala, “WCDMA for UMTS,” John Wiley and Sons, Ltd, 2002.

[17] IEEE Std. 802.16-2009, “IEEE standard for local and metropolitan area networks- part 16: Air interface for broadband wireless access systems,” May 29, 2009.

[18] A. J. Goldsmith and S. G. Chua, “Variable-rate variable-power MQAM for fading channels,” IEEE Trans. Commun., vol. 45, pp. 1218-1230, Oct. 1997.

55

[19] IEEE Std. 802.11, “IEEE standard for local and metropolitan area networks- part 11: Wireless LAN medium access control (MAC) and physical layer (PHY) specifications,” Jun. 2007.

[20] J. Watson, “Strategy: an introduction to game theory,” New York: W.W. Norton, 2008.

[21] M. J. Osborne, “An introduction to game theory,” New York: Oxford University Press, 2004.

[22] 3GPP TR 25.892, “Feasibility study for orthogonal frequency division multiplexing (OFDM) for UTRAN enhancement,” 3^rd Generation Partnership Project, Tech. Rep., Jun. 2004.

[23] G. L. Stuber, “Principles of mobile communication, 2nd Ed.,” Kluwer Academic, 1958

[24] 3GPP TS 23.107, “QoS concept and architecture,” 3^rd Generation Partnership Project, Tech. Spec., Dec. 2008.

[25] UMTS Std. 30.03, “Selection procedures for the choice of radio transmission technologies of the UMTS”, 1998.

[26] R. Guerin, H. Ahmadi, and M. Naghshineh, “Equivalent capacity and its application to bandwidth allocation in high-speed networks,” IEEE J. Select.

Areas Commun., vol. 9, pp. 968-981, Sept. 1991.

[27] M. M. Zonoozi, and P. Dassanayake, “User mobility modeling and characterization of mobility patterns,” IEEE J. Select. Areas Commun., vol. 15, pp. 1239 – 1252, Sept. 1997.

[28] E. Kalai, “Nonsymmetric Nash solutions and replications of 2-person bargaining,” Int. J. Game Theory, vol. 6, pp. 129-133, Sept. 1977.

56

Vita

Rui-Chi Ku was born in 1986 in Taichung, Taiwan. She received the B.E. degree in electrical engineering from Yuan Ze University, Tao-yuan, Taiwan, in 2008, and the M.E. degree in the department of communication engineering, college of electrical and computer engineering from National Chiao Tung University, Hsinchu, Taiwan, in 2010. Her research interests include radio resource management and wireless communication.