Guarantee QoS in WiMAX Networks Using Learning Automata

World Applied Sciences Journal 15 (4): 525-531, 2011 ISSN 1818-4952 © IDOSI Publications, 2011 Bandwidth Allocation in Wimax Networks Using Reinforcement Learning 1 Saeid M. Jafari, 2Majid Taghipour and 3M.R. Meybodi 1 Department of Computer and IT Engineering, Qazvin, Iran University of Applied Science and Technology, Urmia, Iran 3 Department of Electrical and Computer Engineering, Amirkabir, Tehran, Iran 2 Abstract: An important problem for the WiMAX networks is how to provide a guaranteed quality of service for applications. A key aspect of this problem is how base stations should share bandwidth capacity between different classes of traffic. The decision needs to be made for each incoming packet and is known as the packet scheduling problem. A major challenge in packet scheduling is that the behavior of each traffic class may not be known in advance and can vary dynamically. this paper has described how the packet scheduling problem has been modeled as an application for reinforcement learning. We have demonstrated how our reinforcement learning approach could learn scheduling policies that satisfy the quality of service requirements of multiple traffic classes under a variety of conditions. The proposed solution has been designed to have an ability to accommodate integrated traffic in the networks with effective scheduling schemes. A series of simulation experiments have been carried out to evaluate the performance of the proposed scheduling algorithm. Results revealed that the proposed solution performs effectively to the integrated traffic composed of messages with or without time constraints and achieves proportional fairness among different types of traffic. Key words: WiMAX %Scheduling Algorithms %Channel assignment %Learning Automata %Quality of Service INTRODUCTION deliver broadband service in the fixed point-to-point (PTP) or point-to-multipoint (PMP) topologies and it has proposed a framework for the QoS services for four types of traffic. Unsolicited Grant Service (UGS), real-time Polling Service (rtPS), non real-time Polling Service (nrtPS) and Best Effort (BE) QoS classes [5]. UGS supports real-time service flows that have fixed-size data packets on a periodic basis. RtPS supports real-time service flows that generate variable data packets size on a periodic basis. The BS provides unicast grants in an unsolicited manner like UGS where as the UGS allocations are fixed in size. NrtPS is designed to support non real-time service flows that require variable size bursts on a regular basis. BE is used for best effort traffic where no throughput or delay guarantees are provided. Those service classes are defined in order to satisfy different types of Quality of Service (QoS) requirements. However, the IEEE 802.16 standard does not specify the scheduling algorithm to be used [7]. Vendors and operators have to choose the scheduling algorithm(s) to be used. Three types of schedulers must be defined; an uplink and a downlink scheduler both in the Base Station (BS) and just an uplink scheduler for the Subscriber Station (SS) between the different simultaneous connections of the SS. WiMAX technology based on the IEEE 802.16 standard [1] has a very rich set of features [2]. Indeed, it is a very promising Broadband Wireless Access (BWA) technology. The major attractions of WiMAX systems come from their ability to provide broadband wireless access and potential ability to compete with existing wired systems such as fiber optic links, coaxial systems using cable modems and digital subscriber line (DSL) links with much scalability [3]. The WiMAX networks have the capacity to provide flexibility and efficiency to allow coexistence of different types of traffic, such as real-time and multimedia traffic. The IEEE 802.16 standard provides specification for the medium access control (MAC) and physical (PHY) layers for the air interface. The standard includes details about the various flavors of PHY layers supported and characteristics of the MAC layer such as bandwidth request mechanisms and the scheduling services supported [6]. One important issue in the WiMAX networks design is to support QoS services to different types of traffic. IEEE802.16d standard [1] ratified in June 2004, has specified all the techniques of the WiMAX systems to Corresponding Author: Saeid M. Jafari, Department of Computer and IT Engineering, Qazvin, Iran. 525 World Appl. Sci. J., 15 (4): 525-531, 2011 This paper has presented a system for packet scheduling that is based on Learning Automaton. In our approach, Learning Automaton is used to learn a scheduling policy in response to feedback from the network about the delay experienced by each traffic class. Key advantages of our approach are that our system does not require prior knowledge of the statistics of each traffic flow and can adapt to changing traffic requirements and loads. In practice, this helps network providers to deliver a guaranteed QoS to customers, while maximizing network utilization and minimizing the need for manual intervention. We make three key contributions in this paper: (1) we present a model for using RL to address the problem of packet scheduling in Base Station with QoS requirements; (2) we demonstrate the advantages of RL in terms of convergence time in comparison to other scheduling schemes; and (3) we provide an insight into the relative merits of two alternative RL algorithms in the context of this application. We begin by describing the application of packet scheduling. We then describe our solution based on LA and demonstrate its effectiveness in a range of simulated traffic conditions. The Temporary Removal Scheduler (TRS) scheduler [8] involves identifying the packet call power, depending on radio conditions and then temporarily removing them from a scheduling list for a certain adjustable time period TR. The scheduling list contains all the SSs that can be served at the next frame. When TR expires, the temporarily removed packet is checked again. If an improvement is observed in the radio channel, the packet could be topped up in the scheduling list again, otherwise the process is repeated for another TR duration. In poor radio conditions, the whole process could be repeated up to L times at the end of which, the removed packed is added to the scheduling list, independently of the current radio channel condition. The Opportunistic Deficit Round Robin (O-DRR) scheduler [9] is used in the uplink direction. The BS polls subscribers periodically. After each period, the BS determines the set of subscribers that are eligible to transmit and their bandwidth requirements. This set is defined as the eligible set. A number of conditions must be verified by an SS to be in this set: (1) the queue is not empty. (2) The received SIR is above a minimum threshold denoted SIRth. Once these conditions are satisfied, the subscriber is eligible to transmit during a given frame of the current scheduling epoch. The scheduled set changes dynamically depending on the wireless link state of subscribers. At the beginning of each scheduling epoch, the BS resets the eligible and scheduled sets and repeats the above mentioned process. The temporary TRS can be combined with the RR scheduler [7]. The combined scheduler is called TRS+RR. For example, if there are k packet calls and only one of them is temporary removed, each packet call has a portion, equal to 1/(k - 1), of the whole channel resources. The TRS can be combined with the mSIR scheduler. The combined scheduler is called TRS+mSIR [10]. This scheduler assigns the whole channel resources to the packet call that has the maximum value of the Signal to Noise Ration (SNR). The station to be served has to belong to the scheduling list. Previous Work: In this section, we present some schedulers. The simplest scheduling algorithm is the Round Robin (RR) scheduler. It distributes equal channel resources to all the SSs without any priority. The RR scheduler is simple and easy to implement. However, this technique is not suitable for systems with different levels of priority and systems with strongly varying sizes of traffic. There is an extension of the RR scheduler, the Weighted Round Robin (WRR) scheduler, based on static weights. In the same context, we present the Deficit Round Robin (DRR) scheduler. The DRR scheduler associates a fixed quantum (Q i) and a deficit counter (DC i) with each flow i. At the start of each round and for each flow i, DC I is incremented by Q i. The head of the queue i is eligible to be queued if DC i is greater than the length of the packet waiting to be sent (L i). In this case, DC i is decremented by L i. At each round, one packet at most can be sent (and then queued) for each flow. Maximum signal to interference ratio (mSIR) Scheduler is based on the allocation of radio resources to subscriber stations which have the highest Signal to-Interference Ratio (SIR). This scheduler allows a highly efficient utilization of radio resources. However, with the mSIR scheduler, the users with a SIR that is always small may never be served. Problem Definition: This study is based on the model of packet scheduling in cellular network described by Hall and Mars [11]. The RL algorithm has been presented by Taghipoor firstly. But it had high overload because of use of probability matrix, then was not a real automata. Nevertheless in new proposed algorithm use same efficacy. 526 World Appl. Sci. J., 15 (4): 525-531, 2011 The aim of this study is to schedule N classes of traffic, where each traffic class has its own queue qi; i = 1: N. Let qN denote the queue for best-effort traffic, which has no predefined delay requirements. For each remaining queue qi; i = 1:::N - 1, there is a mean delay requirement Ri, which is the maximum acceptable mean queuing delay per packet for the traffic class assigned to qi. Let Mi denote the measured mean queuing delay of packets in qi over the last P packets. The aim is to learn a scheduling policy that minimizes MN while ensuring that Mi _ Ri for i = 1: N - 1. In other words, we want to satisfy the QoS constraints for queues qi; i = 1: N-1 while maximizing the available bandwidth to the best-effort queue qN. In keeping with the model of Hall and Mars [11], all packets in our system have a constant fixed length. This is typical of the internal queues in routers that use a cell switching fabric. We can model this traffic using a discrete-time arrival process, where a fixed length timeslot is required to transmit a packet and at most one packet can be serviced at each timeslot. The arrival of packets is described by a Bernoulli process, where the mean arrival rate li for qi is represented by the probability of a packet arriving for qi in any timeslot. The role of the scheduler is to decide which queue should be serviced at each timeslot (Figure 1). At each timeslot, the scheduler must select an action a g {a1: aN}, where ai is the action of choosing to service the packet at the head of queue qi. The scheduler makes this selection by using a scheduling policy A, which is a function that maps the current state of the system s onto an action a. If the set of possible actions is denoted by A and the set of possible system states is denoted by S, then A: S6A. The second component of the scheduler is a reward function r: S ×A6R. When an action a , A is executed in state s g S, the scheduler receives a reward r(s; a) from the system. This reward provides feedback about the immediate value of executing the action a. The goal is to learn an optimal scheduling by iteratively refining an initial probability vector. Each time we use our current probability vector to select a scheduling action a, we observe the immediate reward R (a) and use this reward as feedback to update our current probability vector (p). This approach is known as reinforcement learning, which has been applied to a variety of scheduling and control tasks. In the next section, we describe our method for using learning automata to optimize the scheduling policy of our queue management system. Fig. 1: Packet scheduling for multiple traffic classes Our Learning Automaton Approach: There are three key components to our application of using Learning Automaton to learn scheduling policies for queue management. First, we require a representation of the state s of our system, which reflects the state of the traffic in our queues. Second, we require a suitable reward function: S×A6R, which reflects the immediate value of our scheduling actions. Finally, we require a learning algorithm to refine our policy function A(s) based on the feedback provided by our reward function. Let us now describe our solution for each of these components of our system. State Representation: The reason for introducing the system state into the policy function is so that the scheduler can learn how to act in different situations. This is in contrast to the approach of using a SLA, which uses a single state in its policy function, i.e., the scheduling policy does not depend on the state of the queues. By introducing a more sophisticated state representation we can potentially gain greater control, albeit at the risk of greater complexity. However, we need to ensure that the state representation is not too complex; otherwise there may be too many parameters to be tuned, which may slow the convergence rate of the algorithm. The aim of scheduling is to use different scheduling policies depending on which queues are not meeting their delay requirements. We represent the state of the system by a set of N -1 binary variables {s1: s-1}, where each variable si indicates whether traffic in the corresponding queue qi is meeting its mean delay requirement Ri, 0 Si =  1 M i ≤ Ri M i > Ri Note that there is no variable corresponding to the best-effort queue qN, since there is no mean delay requirement for that queue. For example, the state {0; 0;::: ; 0} represents that all queues have satisfied 527 World Appl. Sci. J., 15 (4): 525-531, 2011 their mean delay constraint, while (1; 0;::: ; 0} represents that the mean delay requirements are being satisfied for all queues except q1. Thus, if there are N queues in the system including one best-effort queue, then there are 2N-1 possible states. In practice, the number of traffic classes is normally small, e.g., four classes in Cisco routers with priority queuing, in which case the number of states is acceptable. rtime , i = N −1 ∑ wi rtime,i i =1 Where the weights wi depend on which queue was serviced by the last scheduling action. In practice, we have found that suitable weights are wi = 0:3 if qi was the queue serviced by the last action, otherwise wi = 1:0. These weights discourage the scheduler from servicing queues with satisfactory performance if there are other queues experiencing unsatisfactory performance. Although the choice of weight values is not critical, we found that we can significantly improve the convergence rate of our system by using a non-zero weight for queues that were not serviced by the last action. The state reward is positive if a scheduling action causes the system to move to a better state —compared to the previous state s. State —is considered to be better than s if it has more queues whose mean delay requirements are being met, e.g., —= {0; 0;::: ; 0} is better than s = {1; 0;::: ; 0}. Thus Reward Function: The role of the reward function is to provide feedback to the Learning Automaton algorithm about the effect of a scheduling action. Based on this feedback, the learning algorithm can decide how to update the current scheduling policy. The aim of reward function is to provide a positive reward when packets are serviced within their delay requirement and a negative reward when they are late. We also want to provide a positive reward when the system moves to a better state, i.e., when the measured mean delay for a queue falls below the required mean delay. Thus our reward function r comprises a time reward component rtime and a state reward component rstate, where r = rtime + rstate. Every time a scheduling action is executed, the time reward rtime;i for each queue qi is calculated in terms of the mean delay requirement Ri and the measured mean delay Mi. C1 M1 if M i < Ri Ri  rtime , i = C1 if M i = Ri  −C if M i > Ri  2 C tstate =  3 O if S' Is better than S Else Learning Algorithm: The updating algorithm (function) is used to enable the automaton to learn the state of the random environment based on the obtained feedback and choose the best possible action at any point of stage. For a multi-action system, the updating algorithm writes as follows: When a positive response is obtained for an action, its probability is increased and the probabilities of all other actions are decreased. If a negative feedback is received for an action, the probability of that action is decreased and that of others is increased. The time reward is positive when Mi = Ri and negative when Mi 6 Ri. It is maximized when the mean delay requirement is just satisfied. (Figure 2) There is a diminishi0ng reward as Mi approaches zero, since any reduction in Mi below Ri is wasting bandwidth that could be allocated to other queues. In general, it is possible to change the form of the reward function depending on the type of QoS requirements that need to be satisfied. The total time reward is a weighted sum of the rewards for each queue. Pi(n + 1) = pi(n) + a(ri) Pi(n + 1) = pi(n) – ((1 – ri)/3), j i When a negative feedback is obtained for action i, the automaton updates its action probability set based below Eq. Pi(n + 1) = pi(n) – a(ri) Pi(n + 1) = pi(n) + " ((ri)/3), j i In all our simulations, we assume a = 5 and we refer to this as learning parameter. Fig. 2: Time reward function 528 World Appl. Sci. J., 15 (4): 525-531, 2011 Table 1: Slot size for OFDM PH UGS Modulation Channel Coding Slot Size(byte) 64 –QAM 64-QAM QPSK QPSK 3/4 2/3 3/4 1/2 108 96 36 24 Latency (ms) 2000 1500 1000 Table 2: WiMAX parameter Parameter Value PHY Bandwidth Frame per Second Duplexing mode ARQ/CRC OFDM 7MHz 400 TDD OFF 500 0 500 1000 1500 2000 2500 3000 3500 UGS Traffic load (Kbps) Fig. 3: Latency versus traffic UGS Table 3: The scenario used in our simulation Data rate (Kbps) Scenario UGS rtps nrtps BE 1 3 3 3 1 4000 3500 3000 2500 2000 1500 1000 500 0 Evaluation: This section presents the simulation results for the proposed scheduling solution. For testing performance of proposed mechanism, the introduced RL is implemented in the Network Simulator (NS-2) [13] and WiMAX module [7] that is based on the WiMAX NIST module [10]. The MAC implementation contains the main features of the 802.16 standard, such as downlink and uplink transmission. We have also implemented the most important MAC signaling messages, such as UL-MAP and DL-MAP, authentication (PKM), capabilities (SBC), registration (REG), dynamic service addition (DSA) and dynamic service change (DSC). The implemented PHY is OFDM. The current implementation also supports different MCSs. Table 1 shows presents slot size for different modulations and channel coding types. We present a simulation scenario to study thoroughly the proposed scheduling solution. The scenario will present a multi-service case, in which a provider has to support connections with different 802.16 classes and traffic characteristics. The purpose of this scenario is to ensure that the scheduler at the BS takes the service class into account and allocates slots based on the QoS requirements and the request sizes sent by SSs. Another purpose is to test that the scheduler at the BS takes the MAC overhead into account. Table 1 presents information about which applications are active at scenario. Regardless of the simulation scenario, the general parameters of the 802.16 network are the same (Table 2). There is one BS that controls the traffic of the 802.16 network. The physical layer is OFDM. The BS uses the dynamic uplink/downlink slot assignment for 0 500 1000 1500 2000 2500 3000 3500 UGS Traffic Load (Kbps) Fig. 4: Throughput versus traffic Fig. 5: Latency versus simulation time the TDD mode. Both the BS and all Sss use packing and fragmentation in all simulation scenarios. The MAC level uses the largest possible PDU size. ARQ is turned off; neither the BS nor SSs use the CRC field while sending packets. Table 3 presents information number applications are active at different times. 529 World Appl. Sci. J., 15 (4): 525-531, 2011 Fig. 7 shows all the three queues decrease significantly that is due to delay satisfaction. Then rtps value gently ascends until rtps and nrtps numbers tend to a fixed numbers. CONCLUSION In this article, the behavior of some scheduling algorithms and the proposed algorithm based on the delay parameters and simulation were compared. Algorithm proposed algorithms and Maximum signal to interference ratio hvea the best results. RL schemes are able to satisfy QoS requirements. For multiple traffic classes, without starving resources from best-effort traffic. Furthermore, our RL schemes can adapt to changing traffic statistics and QoS requirements. Simulation results show that the proposed Scheduler Find the best solution for nrtPS and rtPS traffic. For future work, we will use our proposed method to build up a flexible and intelligent system based on mSIR scheduler. In this system we will train network with minimize packet latency and signal noise ratio for improving QoS and increasing system performance. Fig. 6: Latency versus simulation time Fig. 7: Automaton Vector variation REFERENCES In this section, we compare five scheduling algorithms: the NIST_RR, mSIR, RR, TRS+RR and TRS+mSIR schedulers with the proposed method. Fig. 3 shows latency packets as a function of the traffic load submitted to the network. The data packets are generated by a streaming multimedia application. UGS scheduling algorithms considered in the round except the standard Rubin worked diagram is linear and its graph throughput is linear with a slope (Figure 4). Because of this the following are the algorithms mentioned above, the traffic request is not the highest priority if package is available in this type of traffic speed data services and will be sent. Figures 5 and 6 show that the RL has a very good behavior in packets mean latency, because this scheduler controls rtPS and nrtps classes latency in both inter and intra-class mechanism by considering latency values of subscribers. TRS+mSIR has good behavior than the RR scheduler because the channel quality of different SSs is not taken into consideration in RR. RR has low efficiency because this scheduler allocates all the symbols to SS even if it has not data to send. 1. 2. 3. 4. 5. 6. 530 IEEE 802.16-2004, IEEE Standard for local and metropolitan area Networks, Air Interface for Fixed Broadband Wireless Access Systems, Oct 2004. Borin, J.F. and N.L.S. Da Fonseca, 2008. Simulator for WiMAX networks. Elsevier journal, Simulation Modelling Practice and Theory, 15(7): 817-833. Ball, C.F., E. Humburg, K. Ivanov and F. Treml, 2005. Comparison of IEEE802.16 WiMAX Scenarios with Fixed and Mobile Subscribers in Tight Reuse, the 14th IST Mobile & Wireless Communication Summit, Dresden, 19-23 June 2005. Borin, J.F. and N.L.S. Da Fonseca, 2008. Simulator for WiMAX networks. Elsevier journal, Simulation Modelling Practice and Theory, 15(7): 817-833. IEEE 802.16e-2005. 2006. IEEE Standard for local and metropolitan area networks, Air Interface for Fixed Broadband Wireless Access Systems Amendment 2: Physical and Medium Access Control Layers for combined Fixed and Mobile Operation in Licensed Bands and Corrigendum. February 2006. Tsai, T., C. Jiang and C. Wang, 2006. CAC and Packet scheduling Using Token Bucket for IEEE 802.16 Networks. J. Communications, 1(2). World Appl. Sci. J., 15 (4): 525-531, 2011 7. 8. 9. Belghith, A. and L. Nuaymi, 2008. Design and Implementation of a QoS-included WiMAX Module for NS-2 Simulator. First International Conference on Simulation Tools and Techniques for Communications Networks and Systems, SIMUTools 2008, Marseille, France, March 3-7,2008. Ball, C.F., F. Treml, X. Gaube and A. Klein, 2005. Performance Analysis of Temporary Removal Scheduling applied to mobile WiMAX Scenarios in Tight Frequency Reuse”, the 16th Annual IEEE International Symposium On Personal Indoor and Mobile Radio Communications, PIMRC’ 2005, Berlin, 11-14 September 2005. Rath, H.K., A. Bhorkar and V. Sharma, 2006. An Opportunistic DRR (O-DRR) Uplink Scheduling Scheme for IEEE 802.16-based Broadband Wireless Networks, IETE, International Conference on Next Generation Networks (ICNGN), Mumbai. 10. Vinay, K., N. Sreenivasulu, D. Jayaram and D. Das, 2006. Performance Evaluation of End-to-end Delay by Hybrid Scheduling Algorithm for QoS in IEEE 802.16 Network, Wireless and Optical Communications Networks, 2006 IFIP International Conference. 11. Hall, J. and P. Mars, 1998. Satisfying QoS with a Learning Based Scheduling Algorithm. School of Engineering, University of Durham. 12. Meybodi, M.R. and S. Lakshmivarahn, 1983. A Learning Approach to Priority Assignment in aTwo Class M/M/1 Queuing System with Unknown Parameters, Proceedings of the Third Yale Workshop on Applications of Adaptive Systems Theory, Center for Systems Science, Yale University, Yale, USA, pp: 106-109. 13. The network simulator ns-2. http:// www.isi.edu/ nsnam/ ns/, September 2007. 531 

IJCSI International Journal of Computer Science Issues, Vol. 9, Issue 1, No 1, January 2012 ISSN (Online): 1694-0814 www.IJCSI.org 109 A Novel Algorithm for Manets using Ant Colony Javad Pashae Barbin `1, Majid Taghipoor2 and Vahid Hosseini 3 1 Islamic Azad University, Naghadeh Branch Uromieh; Iran 2 University of Applied Science Technology Uromieh; Iran 3 Islamic Azad University, Shabestar Branch Uromieh; Iran Abstract Mobile Ad-hoc Networks have recently attracted a lot of attention in the research community as well as the industry. Quality of Service support for MANETs is an exigent task due to dynamic topology and limited resource. Routing, the act of moving information across network from a source to a destination. Conventional routing algorithms are difficult to be applied to a dynamic network topology, therefore modeling and design an efficient routing protocol in such dynamic networks is an important issue. It is important that MANETs should provide QoS support routing, such as acceptable delay, jitter and energy in the case of multimedia and real time applications. One of the meta-heuristic algorithms which are inspired by the behavior of real ants is called Ant Colony Optimization algorithm. In this paper we propose a new on demand QoS routing algorithm “Ant Routing for Mobile Ad Hoc Networks” based on ant colony. The proposed algorithm will be highly adaptive, efficient and scalable and mainly reduces end-to-end delay in high mobility cases. Keywords: MANET, Quality of Service (QoS), Ant Colony, Routing Protocol. 1. Introduction Mobile Ad hoc Networks (MANET) is a communication network of a set of mobile nodes, placed together in an ad hoc manner, without any fixed infrastructure that communicate with one another via wireless links. The devices used to form an Ad Hoc Network possess limited transmission range; therefore, the routes between a source and a destination are often multi hop. As there are no separate routers, nodes that are part of the network need to cooperate with each other for relaying packets of one another towards their ultimate destinations as they do not have central administration, it is easy to deploy and expand. This kind of network is very flexible and suitable for applications such as temporary information sharing in conferences, military actions and disaster rescues. However, multi-hop routing, random movement of mobile nodes and other features unique to MANETs lead to enormous overheads for route discovery and maintenance. Furthermore, compared with the traditional networks, MANETs suffer from the resource constraints in energy, computational capacities and bandwidth. [1] With the increasing needs of QoS provisioning for evolving applications such as real-time audio/video, it is desirable to support these services in ad hoc networking environments. The network is expected to guarantee a set of measurable specified service attributes to the user in terms of end to-end delay, bandwidth, probability of packet loss, energy and delay variance (jitter). The role of a QoS routing strategy is to compute paths that are suitable for different type of traffic generated by various applications while maximizing the utilizations of network resources. [3] The major objectives of QoS routing are: [2] 1. To find a path from source to destination satisfying user’s requirements 2. To optimize network resource usage and 3. To degrade the network performance when unwanted things like congestion, path breaks appear in the network. The main problem to be solved by QoS routing algorithm is the Multi-Constraint Path problem. Algorithms to solve this family of problems are known to be heuristics which can reduce the complexity of the path computation. [2, 4] The path computation algorithm is at the core of QoS routing strategies. Instead of using a shortest path algorithm based on statically configured metrics, as in traditional routing protocols, the algorithm must select several alternative paths that are able to satisfy a set of constraints regarding, for instance, end-to-end delay bounds and bandwidth requirements. However, the algorithms to solve such a problem have been shown to have, in general, high computational complexity. [3] Several approaches have been proposed to address the complexity of multi-constrained path computation problem. The paths that satisfy these multiple constraints Copyright (c) 2012 International Journal of Computer Science Issues. All Rights Reserved. IJCSI International Journal of Computer Science Issues, Vol. 9, Issue 1, No 1, January 2012 ISSN (Online): 1694-0814 www.IJCSI.org are called as feasible paths. The solution of this kind of multi constrained problem requires a path computation algorithm that finds paths satisfying all the constraints. Since the optimal solution of this type of problems with multiple additive and independent metrics is NP-complete, usually heuristics or approximation algorithms can be used to solve such kind of problems. [1] In this paper, I will present a new approach for an ad hoc routing algorithm, which is based on Ant Colony Optimization (ACO) algorithm. We show that for a wide range of different environments and performance metrics, ACO performs better than AODV, AOMDV. The rest of the paper is organized as; section 2 discusses related ant based routing protocol. Section 3 describes the new proposed routing protocol. Section 4 discusses performance evaluation parameters and results of routing protocol followed by conclusions in section 5. 2. Ant colony optimization (ACO) algorithm ACO algorithms have been inspired by the behavior of a real ant colony. The algorithm can find the optimum solution by generating artificial ants. Just as real ants search for food in their environment, the artificial ants search the solution space. The probabilistic movement of ants in the system allows the ants to explore new paths and to re-explore the paths visited earlier. The strength of the pheromone deposit directs the artificial ants toward the best paths and pheromone evaporation allows the system to forget old information and avoid quick convergence to suboptimal solutions. The probabilistic selection of paths allows the artificial ants to search for a large number of solutions. ACO has been applied successfully to discrete optimization problems such as the traveling salesman problem and routing [5, 6]. A number of proofs for the convergence to the optimum path of the ACO can be found in [8, 9]. 2.1. Basic and Background The basic idea of the ant colony optimization metaheuristic is taken from the food searching behavior of real ants. When ants are on the way to search for food, they start from their nest and walk toward the food. When an ant reaches an intersection, it has to decide which branch to take next. While walking, ants deposit pheromone1, which marks the route taken. The concentration of pheromone on a certain path is an indication of its usage. With time the concentration of pheromone decreases due to diffusion effects. This property is important because it is integrating dynamic into the path searching process. [7] 110 Fig. 1 All ants take the shortest path after an initial searching time. [7] Figure 1 shows a scenario with two routes from the nest to the food place. At the intersection, the first ants randomly select the next branch. Since the below route is shorter than the upper one, the ants which take this path will reach the food place first. On their way back to the nest, the ants again have to select a path. After a short time the pheromone concentration on the shorter path will be higher than on the longer path, because the ants using the shorter path will increase the pheromone concentration faster. The shortest path will thus be identified and eventually all ants will only use this one. This behavior of the ants can be used to find the shortest path in networks. Especially, the dynamic component of this method allows a high adaptation to changes in mobile ad-hoc network topology, since in these networks the existence of links are not guaranteed and link changes occur very often.[7] 2.2. Solving Network Routing Using ACO Mobile ad hoc network routing is a difficult problem because network characteristics such as traffic load and network topology may vary stochastically and in a time varying nature. The distributed nature of network routing is well matched by the multi agent nature of ACO algorithms.[8] The given network can be represented as a construction graph where the vertices correspond to set of routers and the links correspond to the connectivity among routers in that network. Now network route finding problem is just finding a set of minimum cost path between nodes present in the corresponding graph representation which can be done easily by the ant algorithms. 3. Proposed Algorithm This paper proposes an on demand QoS routing algorithm. Since the requirements for various applications may vary time to time, the approach for QoS routing may not be proactive. The proposed approach has two phases namely route discovery phase and route maintenance phase. When a source node has to pass data to a destination node with QoS requirements it starts with the route discovery phase. Once the route is found, the data transfer will take place. While data transmission is going on, it is also required to maintain the path to the destination. This is very much desirable and required in mobile ad hoc networks and Copyright (c) 2012 International Journal of Computer Science Issues. All Rights Reserved. IJCSI International Journal of Computer Science Issues, Vol. 9, Issue 1, No 1, January 2012 ISSN (Online): 1694-0814 www.IJCSI.org hence is done in the route maintenance phase. The proposed algorithm differs from other similar algorithms because routing from destination to source is done. 3.1. Route Discovery Phase The outline of this phase is as follows: Step 1: Let the source node S has data to send to a destination D with QoS requirements delay, energy, bandwidth and hop count. Each node has a routing table. Step 2: D initiates a ForwardAnt to source S through all its neighbors which it has learned from periodic hello messages. Step 3: While traveling to the source the ForwardAnt collects transmission delay of each link, processing delay at each node, the available capacity of each link, the number of hops visited and stores in routing table each node. Step 4: When the ForwardAnt reaches the source, it will be converted as BackwardAnt and forwarded towards the original destination. The BackwardAnt will take the same path of the corresponding ForwardAnt but in reverse direction. Step 5: For every BackwardAnt reaching an intermediate node or destination node, the node can find the delay, bandwidth, hop count from the received ant to the respective destination. Now the node can calculate the path preference probability to reach the source. Step 6: If calculated path preference probability value is better than the requirements, the path is accepted and stored in memory. Step 7: The path with higher path preference probability will be considered as the best path and data transmission can be started along that path. 111 Send one ForwardAnt to neighbors end if for each message m in D’s buffer do if (m→type = ForwardAnt) then send m to NextHop if NextHop = S then m →type = Backward end if else if m→type = BackwardAnt then find NextHop in C’s BackRouting table send m to NextHop IncreasePheromone(NextHop,m) if NextHop = D or NextHop = Intemediate node then update averages of packet delays and remaining energy drop m end if end if end for Evaporate() /*calculate higher path preference probability will be considered as the best path and data transmission can be started along that path*/ end for 3.2. Route maintenance phase As mentioned before, the values of pheromone trails are stored in a table at each node. Suppose that a data is currently residing in node N and this node has k neighbors H1, H2... Hk and φi is the amount of pheromone assigned to ei and di. The data will select Hi as the next node with a probability pi (ed) which is calculated using the equation 3. P e ∑e P ⁄d ∑ ⁄d P P ∑P In similar algorithms, Route discovery is achieved by flooding forward ants to the destination then the backward will take the same path of the corresponding forward but in reverse direction. Finally, the data packets sent from source to destination. While the proposed algorithm, the algorithm starts from the destination node and the data packet is sent with the backward. pi (e) measure of energy consumption at node i is compared with the total energy consumption. The above stochastic strategy, described in Algorithm 1, establishes multiple paths between the source and destination. As a result, in contrast to regular position based routing algorithms which usually find a single route to the destination, proposed algorithm is a multipath routing algorithm (i.e. like other ACO routing algorithms). pi (d) measure of routing delay at node i is compared with the total routing delay. Algorithm 1 routing algorithm {S is the source node, D is the destination node and C is the current node} for each clock time do if C = S then Send one BackwardAnt from reverse direction else P P Where: ei: energy level of node i. di: latency of node i. θ : Constant value θ Copyright (c) 2012 International Journal of Computer Science Issues. All Rights Reserved. IJCSI International Journal of Computer Science Issues, Vol. 9, Issue 1, No 1, January 2012 ISSN (Online): 1694-0814 www.IJCSI.org Each node scans pi(ed) and selects the node with the highest preference and forwards the data packet (BackwardAnt) to this node. This process is repeated until the destination node is reached. The ants move from one node to another and carry table 1. Node X Table 1: route information Energy Time Estimated Waiting Ex t t 4. Simulation and Performance analysis VC++ 6.0 program is used to realize the simulation of the algorithm. For our results we assumed 50 mobile nodes communicating via IEEE 802.11. The nodes move inside a simulation area of 1500m*300m. The simulation time is 900 seconds. The nodes move with a maximal velocity of 10 m/s and according to the random waypoint mobility model. Figure 4 depicts the needed number of packets to perform the routing job for all three routing algorithms. In the cases with high mobility it is obvious that MPAC and AOMDV create the least overhead. Especially MPAC shows here a better performance than AOMDV in high dynamic. With less, the overhead of MPAC and AODV are very close. AOMDV shows here again their poor performance by creating large numbers of routing packets. Figure 5 show average remaining energy for all three routing algorithms. Remaining energy of MPAC is high than AODV and AODV. That is because optimum tour is established by the way of flooding, costing more time in MPAC, and lead to increasing consumption energy of network. The number of delivered packets Algorithm 2 Select Next Hop Input: node N for i = 1 to the number of N’s neighbors do P P P ∑ P P Return neighbor i with probability pi end for 112 A MPAC AODV AOMDV 0 200 600 800 1000 Fig. 2 Delivery rate overload of routing 4.1. Simulation and Performance analysis A new routing algorithm should show its performance in comparison with existing and known algorithms. The simulated traffic is Constant Bit Rate (CBR). We compare our protocol with AODV and AOMDV. We evaluate mainly the performance according to the metrics packet delay, overload and packet delivery. Our proposed algorithms are named MPAC (Multi Path Ant Colony). We will first discuss the robustness of the routing protocols. Figure 2 shows the delivery rate, i.e., the part of packets a certain routing protocol was able to deliver properly. This value is important, since it describes the performance which transport protocols will see, i.e., the throughput is restricted by this value. In situations with very high dynamics MPAC shows the best performance. With less dynamic, down to 300 seconds of pause time, AODV has best performance. Because AODV uses particular route, over time this will destroy the path and find a new path will lead to delay. But in the proposed protocol does not caused a problem because use of different routes. Figure 3 shows the overhead of MPAC. All results are very close through all simulation scenarios. This shows that MPAC creates much less routing overhead for all considered mobility scenarios. 400 simulation time(ms) AMPAC AODV AOMDV 0 200 400 600 800 simulation time(ms) Fig. 3 Overhead Copyright (c) 2012 International Journal of Computer Science Issues. All Rights Reserved. 1000 IJCSI International Journal of Computer Science Issues, Vol. 9, Issue 1, No 1, January 2012 ISSN (Online): 1694-0814 www.IJCSI.org 113 The number of delivered packets Information Networking and Applications, AINA 2005 1: 426-431, 2005 MPAC A AODV AOMDV 0 200 400 600 800 [3] Dr. Shuchita Upadhayaya and Charu Gandhi,” QUALITY OF SERVICE ROUTING IN MOBILE AD HOC NETWORKS USING LOCATION AND ENERGY PARAMETERS”, International Journal of Wireless & Mobile Networks (IJWMN), Vol 1, No 2, November 2009. [4] Frederick Ducatelle, ``Adaptive Routing in Ad Hoc Wireless Multi-hop Networks'', Ph.D Thesis, Universitμa della Svizzera italiana Faculty of Informatics. 1000 simulatin time(ms) Fig. 4 Comparison of three protocols by the number of needed routing packets Average remain energy AMPAC AODV [5] A. Sabari, K Duraiswamy, ``Ant Based Adaptive Multicast Routing Protocol (AAMRP) for Mobile Ad Hoc Networks'' International Journal of Computer Science and Information Security Vol 6, No 2, 2009. [6] FJ Arbona Bernat, ``Simulation of Ant Routing Protocol for Ad-hoc networks in NS-2'', Faculty of Electrical Engineering, Mathematics and Computer Science Network Architectures and Services Group. [7] Claudia Lorenz,Patrizia Mottl ``Ant Algorithm - Phd thesis'', Swiss Federal Institute of Technology, Zurick AOMDV [8] P.Deepalakshmi1 , Dr.S.Radhakrishnan2, “Ant Colony Based QoS Routing Algorithm For Mobile Ad Hoc Networks”, in proc International Journal of Recent Trends in Engineering, Vol. 1, No. 1, May 2009. 0 200 400 600 800 1000 simulation time(ms) Fig. 5 Average remaining energy 5. Conclusion In this paper, we have proposed an ant based routing protocol for mobile ad hoc networks. ACO based algorithms have specialized to provide adaptive and efficient solutions to network routing. AMPC provides multiple paths with comparatively less overhead in the network. The simulation results indicate that proposed scheme can perform better than AODV and AOMDV under high mobility because of alternate route maintenance scheme. In future, we have planned to investigate the performance of the algorithm for real-time multimedia data using various mobility models. [9] Chunxue Wu, Fengna Zhang, Hongming Yang,” A Novel QoS Multipath Path Routing in MANET”,International Journal of Digital Content Technology and its Applications Volume 4, Number 3, June 2010. [10] Jafari, Saeid. M., Taghipour, M. and Meybodi, M. R. "Bandwidth Allocation in Wimax Networks Using Reinforcement Learning", World Applied Sciences Journal Vol. 15, No. 4, pp. 525-531, 2011. Javad Pashaei received B.E. degree from Shabestar Azad University in 2006. He received M.S. degree from Shabestar Azad University in 2010. His research interests include network management as well as WiMAX and ad hoc networks. Majid Taghipoor teaches in university of applied science technology. His research interests include network management as well as WiMAX, ad hoc, Manet networks and NOC. Vahid Hosseini received M.S. degree from Shabestar Azad University in 2010. 6. Reference [1] Chlamtac, M. Conti, and J. Liu, “Mobile ad hoc networking: imperatives and challenges”, Ad Hoc Networks, No. 1, 2003. [2] Z.Liu , M. Z. Kwiatkowska and C. Constantinou , 9 A biologically inspired QoS routing algorithm for ad hoc network : , International Conference on Advanced Copyright (c) 2012 International Journal of Computer Science Issues. All Rights Reserved. 

World Applied Sciences Journal 15 (4): 576-583, 2011 ISSN 1818-4952 © IDOSI Publications, 2011 Bandwidth Allocation in WiMAX Networks Using Learning Automaton 1 Saeid M. Jafari, 2Majid Taghipour and 3M.R. Meybodi 1 Department of Computer and IT Engineering Qazvin, Iran University of Applied Science and Technology Urmia, Iran 3 Department of Electrical and Computer Engineering Amirkabir Tehran, Iran 2 Abstract: Recent developments on the wireless communication technology have brought much innovativeness to make wireless access networks, e.g. WiMAX systems, to be able to compete with the wired access networks with much more bandwidth. QoS service provisioning is an important issue for deploying such networks. The IEEE 802.16d standard has specified the services should be provided at the medium access control (MAC) layer in WiMAX networks. However, it has left a wide space for research to develop and implement those specified services. In this paper, the issue of differentiated service provisioning will be addressed with the non-real-time polling service in WiMAX systems. The proposed solution has been designed to have an ability to accommodate integrated traffic in the networks with effective scheduling schemes. A series of simulation experiments have been carried out to evaluate the performance of the proposed scheduling algorithm. In our algorithm we introduce a two-phase Learning Automaton queuing (2PLAQ) algorithm tailored for uplink scheduling in the WiMAX network. It aims to strike the balance between delay requirement and fair bandwidth allocation. The results reveal that the proposed solution performs effectively to the integrated traffic composed of messages with or without time constraints and achieves proportional fairness among different types of traffic. Key words: WiMAX %Scheduling %Learning Automaton %Bandwidth Allocation INTRODUCTION fixed point-to-point (PTP) or point-to-multipoint (PMP) topologies. And it has proposed a framework for the QoS services for four types of traffic. Unsolicited Grant Service (UGS), real time Polling Service (rtPS), non real-time Polling Service (nrtPS) and Best Effort (BE) QoS classes. UGS supports real-time service flows that have fixed-size data packets on a periodic basis. rtPS supports real-time service flows that generate variable data packets size on a periodic basis. The BS provides unicast grants in an unsolicited manner like UGS. Whereas the UGS allocations are fixed in size. nrtPS is designed to support non real-time service flows that require variable size bursts on a regular basis. BE is used for best effort traffic where no throughput or delay guarantees are provided. Those service classes are defined in order to satisfy different types of Quality of Service (QoS) requirements. However, the IEEE 802.16 standard does not specify the scheduling algorithm to be used. Vendors and operators have to choose the scheduling algorithm(s) to be used. Three types of schedulers must be defined; an uplink and a downlink scheduler both in the Base Station (BS) and just an uplink scheduler for the Subscriber Station (SS) between the different simultaneous connections of the SS. WiMAX technology based on the IEEE 802.16 standard [1,2] has a very rich set of features. Indeed, it is a very promising Broadband Wireless Access (BWA) technology. The major attractions of WiMAX systems come from their ability to provide broadband wireless access and potential ability to compete with existing wired systems such as fiber optic links, coaxial systems using cable modems and digital subscriber line (DSL) links with much scalability. The major attractions of WiMAX systems come from their ability to provide broadband wireless access and potential ability to compete with existing wired systems such as fiber optic links, coaxial systems using cable modems and digital subscriber line (DSL) links with much scalability. The WiMAX networks have the capacity to provide flexibility and efficiency to allow coexistence of different types of traffic, such as real-time and multimedia traffic. One important issue in the WiMAX networks design is to support QoS services to different types of traffic. IEEE802.16d standard [1, 2], ratified in June 2004, has specified all the techniques of the WiMAX systems to deliver broadband service in the Corresponding Author: Saeid M. Jafari, Department of Computer and IT Engineering Qazvin, Iran. 576 World Appl. Sci. J., 15 (4): 576-583, 2011 In this paper we present a system for packet scheduling that is based on Reinforcement Learning [16]. In our approach, Reinforcement Learning (RL) is used to learn a scheduling policy in response to feedback from the network about the delay experienced by each traffic class and fairness. In our algorithm we introduce a twophase Learning Automaton Queuing (2PLAQ) algorithm tailored for uplink scheduling in the WiMAX network. It aims to strike the balance between delay requirement and fair bandwidth allocation. In simulations, we consider both popular Poisson traffic and practical burst traffic that is modeled by the Markov Modulated Poisson Process (MMPP). The simulation results verify the correctness of our analytical models and compare 2PLAQ with other scheduling schemes. It achieves a low drop rate and high throughput while maintaining fairness among different connections at the same time. The paper is organized as follows. In Section II we briefly introduce the Related Scheduling Algorithms. In Section III we present the proposed 2PLAQ algorithm. In section IV we analyze and simulate the Two-phase LAQ Scheduling Algorithm. Finally we invest conclusions in Section V. DRR to guarantee the bandwidth allocation of each of the four queues, thus preventing queues with higher priorities from depleting the bandwidth and causing starvation of queues with lower priorities. Analysis model is an important part of the research in scheduling algorithms. In [3], mean delay bound is derived by using the M/G/FQ queuing model, but it is based on WFQ algorithm and does not consider QoS parameters other than MRR. QoS parameters are not considered in [6], where an MMPP model is employed to derive the packet drop rate and characterize uplink rtps and nrtps traffic that share a single First-Come-First-Serve (FCFS) queue. For other scheduling algorithms [4,5,7-12], their performance and effectiveness are demonstrated and compared via simulations only, without in-depth theoretical analysis. Two-Phase Learning Automaton Queuing (2PLAQ) Scheduling Algorithm: In this section, we first present a two-phase Learning Automaton Queuing (2PLAQ) algorithm that addresses the delay and bandwidth requirements while balancing the fairness and efficiency among different connections. Then, an elegant queuing model is established to derive in theory its performance in terms of packet drop rate and throughput. Related Scheduling Algorithms In this Section, We Present Some Schedulers: Several WiMAX scheduling solutions have been proposed. WFQ is proposed and analyzed by an M/G/FQ queuing model in [3]. Although WFQ can guarantee the minimum data rate of the connections, it does not take into consideration the delay constraint. Besides, the time complexity of WFQ scheduling algorithm is high and thus becomes a potential problem for its implementation in the WiMAX network. A two-tier hierarchical architecture is proposed in [4] for WiMAX uplink scheduling. In the higher hierarchy, strict prioritization is used to direct the traffic into the four queues, according to its type. Then, each queue is scheduled according to a particular algorithm, i.e. fixed allocation for UGS, EDF for rtps, WFQ for nrtps and equal division of remaining bandwidth for BE. Although EDF takes care of the delay requirement of the rtps, grouping multiple rtps connections into one queue fails to guarantee the minimum bandwidth requirement of each individual rtps connection. For example, one rtps connection with tight delay budget may dominate the bandwidth allocation, resulting in starvation of other rtps connections. A similar approach is proposed in [7] and it replaces the strict priority algorithm in the higher hierarchy with Deficit Fair Priority Queue (DFPQ). The basic idea of DFPQ is to use Overview of the Two-phase Laq Scheduling Algorithm: The scheduling algorithm aims to meet the QoS requirements of all types of connections. For UGS, the scheduling is straightforward, because a fixed amount of bandwidth is always allocated to each UGS connection during each grant interval, which is determined according to the requested data rate. BE connection has no specific QoS requirement, thus it is not the interest of the study. BE’s scheduling can be done via a simple scheme like equal division of the remaining bandwidth among all BE connections. In the following discussion, we focus on rtps and nrtps traffic that has specific QoS requirements on delay and bandwidth. To strike the balance between delay and bandwidth requirements, the proposed scheduling algorithm decouples them and addresses them separately in two phases. In our discussion so far of LAQ, the bandwidth allocation in Phase 1 is based on MRR. More specifically, each connection is allocated a bandwidth that equals its MRR, unless its total request is lower than its MRR. The remaining bandwidth is then used in Phase 2. Apparently, this is not the only option. In theory, we can make any bandwidth allocation to the two phases. To study the impact of bandwidth allocation between two phases, we introduce a parameter ) , with 0#) #1 and allocate ) i-MRRi to Connection i in Phase 1. 577 World Appl. Sci. J., 15 (4): 576-583, 2011 There is ) for any connection to can provide its requirements. For each connection ) should be determined such that to be able to meet these needs. Therefore, we’ll have a vector of ) . The ) have to be trained using Learning Automaton for each connection so that determining ) as a selective action, indicates delay and bandwidth of the system. The more we concern about selective action, the less packet loss we’ll have and the more fairness we’ll get. So, we are going to have both delay and fairness issues handled. If ) is small, then ) of the reserved bandwidth of each flow is allocated at phase 1 and rest of the bandwidth will remain for next phase. Therefore, small ) allocates less bandwidth to first phase results in an increase in packet loss and also falters fairness, on the other hand large amount of ) , allocates more bandwidth to first phase which leads to reduction in packet loss as well as fairness. At the beginning probability vector contains initial values which are equals. So, chance of selecting any action is equal. After executing ai, it will immediately receive reward and based on that probability vector will be updated. Fig 1: Analytical Model for the Two-phase LAQ Algorithm Unknown Input for Phase 2: The input for Phase 2 is the data not served in Phase 1, which is an unknown parameter to be derived in analysis. In this work, an elegant queuing model as shown in Fig. 1 is constructed to describe the two-phase LAQ algorithm. Our key idea is to creatively use the M/D/1+D (for Markov arrival, deterministic service time, one server, plus deterministic delay budget) queuing model with FCFS and EDF queuing disciplines. We consider a general scenario, where n rtps and/or nrtps connections are established. Connection i has an arrival rate of 8i a delay budget of Ji and a minimum reserved bandwidth of MRRi. For the sake of analytic tractability, we assume that the data arrival forms a Poisson process and all queues have infinite size. Other types of traffic (such as the more practical burst traffic) are studied through simulations. Following the LAQ scheduling algorithm, the queuing model also consists of two phases. In Phase 1, since the connections are served with their reserved bandwidth (i.e. MRRi), each of them can be modeled as a separate and independent queue. Assuming the data packet size is fixed, the service rate is a constant. More specifically, the service rate of Queue i (for Connection i) is µ = MRRi (1). Where L denotes the length of the data L packet. Clearly, the service rate of Queue i is proportional to the MRR of Connection i. Note that, data packets that cannot be served in Phase 1 due to the limited MRR will all be passed to Phase 2. Therefore, we artificially set a delay bound of T (which is the period of one frame) for the queues in Phase I, in order to track the “dropped” data packets of Phase 1, which will be the inputs for Phase 2. Based on the above discussion, we arrive at an M/D/1+D queue for each connection in Phase 1, with an arrival rate of ¸i, a service rate of ¹i and a delay budget of pt +1,i = pt + α * ( M i − Ri ) In general, there is a trade-off between the drop rate and fairness when ) varies from 0 to 1. For example, a small ) indicates more bandwidth allocated to Phase 2, which results in a smaller drop rate but higher unfairness. Analysis of the Two-Phase LAQ Scheduling Algorithm: In this subsection, we analyze the performance of the proposed LAQ algorithm, in order to gain insight into it and to demonstrate in theory its performance. Similar to our earlier discussions, we focus on the QoS performance of rtps and nrtps connections only in our analysis. Note that, we do not explicitly distinguish rtps and nrtps connections, because they only differ in the amount of delay budget. As to be discussed next, our analytic model is generally applicable to any rtps or nrtps connections with given arrival rates, MRR and delay bound. Different Queuing Principles: LAQ involves two phases with different queuing principles, i.e. FCFS in Phase 1 and EDF in Phase 2. Delay Constraint: Each data packet in the queue is associated with a specific delay bound. A data packet is dropped if its waiting time is longer than its delay budget. The delay constraint dramatically increases the analytic complexity. 578 World Appl. Sci. J., 15 (4): 576-583, 2011 ( T. The M/D/1+D queue under FCFS queuing discipline has been well studied. According to [14], the packet drop rate of Queue i is: ( λiτ − ρ ) e− ρ Accordingly, the overall service rate in Phase 2 is i Pid = 1 − 1 + [ ρi + ρi2eλit ρi ∑j r = 0 ( −1) j i j! j i ]−1 (2) µ′ = T and r is the integer satisfying ( µiτ − 1) < r < µiτ The packets “dropped” from Phase 1 become the input of Phase 2. Because such “packet dropping” is random, the “dropped” packets of Queue i also form a Poisson process, with mean arrival rate of λ × P d i As shown in Fig. 1 the input of the queue is the aggregation of the “dropped” packets from all queues in Phase 1. Since the “dropped” packets from each queue are Poisson, the aggregation of them is also Poisson, with a n mean arrival rate of λ ′ = (3) Pd × λ ∑ i =1 i The bandwidth that is reserved but not actually utilized in full by some connections. t = i 1 µi ( ) = λ 1 − Pid × 1 µi i (6) (7) ∑ i =1λi n i Where xi is the normalized throughput of connection i and n is the total number of connections. Here we use the normalized throughput of a connection, i.e. X i = Th i , MRR i (4) with Thi and MRRi stand for the connection i’s actual data rate and reserved data rate, respectively. The Jain’s Fairness Index ranges between 0 and 1. The higher the index, the better the fairness. If Thi = MRRi for all i, or in other words, every connection obtains its reserved data rate, then xi = 1 for all i and Jain’s Fairness Index equals 1. All simulations and analytic calculations are done using NS2 (2.34) simulator. Where t denotes any given time interval and tbusy represents the busy period within t. In the above equation, λ 1 − P d × t is the actual number of arrivals in ( λ′ ∑ ∑ i =1 When a connection is idle, its reserved bandwidth is not used and thus contributing to the bandwidth of Phase 2. For Queue i in Phase 1, the fraction of time when packets are served is: ) n Simulations and Discussion: We have carried out extensive simulations to verify the correctness of the proposed analytical model for the LAQ algorithm and to compare its performance with other scheduling algorithms, under a broad set of simulation parameters. In particular, we consider several comparable scheduling algorithms, including WRR, EDF and DFPQ [7] (which is a representative WiMAX scheduling algorithm and has been patented and well received). Besides packet drop rate and throughput that have been studied in analysis, we are also interested in the fairness performance, which is measured by Jain’s Fairness Index [15] defined as follows: 2  n Xi   i =1   f ( X1, X1,.... X n ) =  n (8) n X2 ∑ i =1 ( ∑ i =1 Pi0µi And the throughput is 1-Pd bandwidth. Thus the remaining bandwidth comes from: ~ the bandwidth that is not reserved by any connections, n i.e. W − MRRi and λ 1 − Pid × t × n i ∑ i =1 tbusy ∑ i =1 MRRi + Pd = Pd′ The service rate in Phase 2 denoted by µ i is estimated as follows. The amount of bandwidth available for Phase 2 is what has been left after Phase 1. We assume that CAC ensures the total reserved bandwidth never greater the total available bandwidth, i.e. n MRRi ≤ W , where W denotes the total available C W− 1 Actually, Equation (4) is a form of Little’s theorem [13] and 1-P0 is also known as the utilization factor. We model Phase 2 as an M/D/1+D queue, with an arrival rate of 8´ and a service rate of µ´. In addition, each packet in the queue is associated with a delay budget, e.g. Ji for the packet from Connection i. The packet drop rate P′ d obtained above is with respect to the input of Phase 2, i.e. the arrivals with a mean rate of 8´ . As shown in Fig 1, the actual inputs of the entire system are the arrivals with means of 81,82,....8n in Phase 1. Therefore, the overall dropping probability is: Where Di equals λi ,τ is the delay bound which is set to µi i ) denoted by P0 , is Pi0 = 1 − λi 1 − Pid × 1 (5) i µi ) t (excluding those being dropped) and 1 is the service µi time for each packet. Thus the fraction of idle time, 579 World Appl. Sci. J., 15 (4): 576-583, 2011 Table 1: Raw Data Rate. SS ID SS1 SS2 SS3 SS4 Modulation type BPSK QPSK 16QAM 64QAM Inner code rate 2-Jan 3-Feb 4-Mar 6-May Bits / symbol 1 2 4 6 Raw data rate Rb(Mbps) 3.716 9.9094 22.2962 37.1603 Bytes / mini slot 1.875 4.999 11.247 18.745 For DFPQ approach in [7], the results are obtained via simulation only.The analysis and simulations are compared in Fig. 2, where X-axis indicates the traffic load of rtps3 (where Dequals arrival rate over MRRi=l) and Yaxis shows the overall packet drop rate of all connections. The simulation result of DFPQ algorithm is also depicted in Fig. 2. We observe that EDF, LAQ and DFPQ have similar packet drop rate, while the drop rate of WRR is much higher, because WRR does not consider the delay budget and is more likely to drop real time data packets. Note that, with the increase of the traffic load of rtps3, the total traffic load becomes significantly higher than the total available bandwidth. Therefore, all four algorithms exhibit high drop rate. Simulation results of the Jain’s Fairness Index are shown in Fig. 3. As we can see, the fairness index of EDF drops dramatically when the traffic load of rtps3 becomes higher than its MRR, because the additional real time traffic (with tight delay budget) aggressively takes bandwidth from the nrtps connections under the EDF algorithm and thus leading to low throughput and unfairness to non-real time traffic. When D of connection rtps3 is greater than 1, the fairness index of both LAQ and DFPQ drops, but DFPQ’s fairness index drops more rapidly, exhibiting worse fairness performance than LAQ. The reason can be two-folded. First, DFPQ always gives real time traffic higher priority than non-real time traffic, resulting in unfairness. The other reason is that DFPQ does not guarantee the minimum reserved rate of each real time connection since EDF is deployed within all the real time connections. Among the four scheduling algorithms, WRR’s fairness index is the stablest when the input traffic varies. In summary, the fairness index of LAQ is better than that of DFPQ and EDF and comparable to WRR. Table 2: MRR and Delay Budget of Different connections. rtps rtps1 rtps2 rtps3 MRR (kbps) 19.2 64 384 rtps4 1024 Delay (ms) 10 30 20 40 nrtps nrtps1 nrtps2 nrtps3 nrtps4 MRR (kbps) 24 48 256 768 Delay (ms) 100 130 170 200 Table 3: Distribution of the 8 Connections in 4 SS. SS ID SS1 rtps1+ nrtps1 s rtps2 + nrtps2 s rtps3 + nrtps3 rtps4 + nrtps4 SS2 SS3 s s s s SS4 s s The simulation parameters are bring in following tables: Each SS establishes a number of connections to the BS in our simulation. We consider four rtps connections, named rtps1, rtps2, rtps3 and rtps4 and four nrtps connections, named nrtps1, nrtps2, nrtps3 and nrtps4. As shown in Table 2, each type of connection is associated with an MRR and a delay budget. Every SS has four connections as shown in Table 3. For example, SS1 has these four connections: rtps1, nrtps1, rtps2 and nrtps2. As a result, there are 16 connections, which request a total MRR of 537 mini slots. We consider two types of traffic in our simulations, i.e. the popular Poisson traffic and the burst traffic generated by the Markov Modulated Poisson Process (MMPP) model, as discussed below. Burst Traffic: To study the performances of the four scheduling algorithms under burst traffic, we use Markov Modulated Poisson Process (MMPP) model to generate data packets. λmax . The burstiness of the traffic is defined as b= λavg The higher the value of b, the more bursty the traffic is. When b equals to 1, the MMPP model is equivalent to the Poisson model. To focus on the impact of burst traffic, we vary b and fix the arrival rate of all connections, i.e. let 8i= 1/5MRRi/L for rtps3 connections and 8i= MRRi/L for other connection. The packet drop rates of the four scheduling algorithms under burst traffic are shown in Fig. 4. As can be seen, WRR is rather vulnerable to the burst traffic. Poisson Traffic: First, we use Poisson traffic, with packet size of l = 800 bytes. The arrival rates of all connections except rtps3 are fixed and match their reserved data rates. For example, the packet arrival rate of Connection i equals MRRi=l. Meanwhile, we vary the arrival rate of rtps3, in order to study the influence on the scheduling results. We have obtained analytic and simulation results for LAQ, as well as the simple WRR and EDF schemes. 580 World Appl. Sci. J., 15 (4): 576-583, 2011 Fig. 2: Drop Rate under Poisson Traffic. Fig. 4: Drop rate under Burst Traffic. Fig. 3: Fairness under Poisson Traffic Fig. 5: Fairness under Burst Traffic. With the increase of b, its drop rate increases dramatically. On the other hand, EDF, LAQ and DFPQ can maintain a reasonably low packet drop rate even when b is large. The fairness of the scheduling algorithms under burst traffic is shown in Fig.4. As we can see, LAQ always maintains a high fairness index, while the fairness of EDF algorithm is the worst among the four algorithms. Notice that when b increases, the fairness index of EDF becomes higher. This is due to the fact that some real time packets of rtps3 connection are dropped under high burstiness and thus the throughput of rtps3 decreases to a value more closer to its reserved throughput. Accordingly, the normalized throughput of rtps3 becomes close to 1, which appears more fair. On the contrary, the fairness of WRR drops with the increase of traffic burstiness. Fig. 6: Drop rate-Impact of Bandwidth Allocation between Two Phases () ). 581 World Appl. Sci. J., 15 (4): 576-583, 2011 Queue (DFPQ). Both analytic and simulation results have clearly shown that LAQ algorithm effectively achieves low packet drop rate and high throughput while maintaining the fairness among different connections. REFERENCES 1. Fig. 7: Fairness-Impact of Bandwidth Allocation between Two Phases () ). In general, there is a trade-off between the drop rate and fairness when ) varies from 0 to 1. For example, a small ) indicates more bandwidth allocated to Phase 2, which results in a smaller drop rate but higher unfairness. However, our simulation results show that the algorithm Achieves higher overall performance under larger) . As can be seen in Fig. 6 and 7, when ) increases from 0 to 1, the drop rate increases only marginally by about 0.1%, while the Jain’s Fairness Index increases from around 0.93 to 0.98, which is close to the perfect fairness. This result justifies our choice of letting ) = 1 in LAQ. CONCLUSION In this article, we have presented a two-phase Learning Automaton Queuing (LAQ) algorithm for uplink scheduling in the WiMAX network. It aims to strike the balance between fair bandwidth allocation and delay requirement. In order to gain deep understanding of and insights into LAQ algorithm, we have established an elegant queuing model to derive in theory the performance metrics in terms of packet drop rate and throughput. The analytic model has been verified by extensive simulations, carried out on the basis of a broad set of parameters according to WiMAX physical layer standards. In simulations, we have considered both popular Poisson traffic and practical burst traffic that is modeled by the Markov Modulated Poisson Process (MMPP). The simulation results have verified the correctness of our analytical models. We have also compared LAQ with WRR, EDF and Deficit Fair Priority 582 IEEE 802.16-2004, 2004. IEEE Standard for local and metropolitan area Networks, Air Interface for Fixed Broadband Wireless Access Systems, 2. "IEEE 802.16e, 2004. IEEE Standard for local and metropolitan area networks, Air Interface for Fixed Broadband Wireless Access System, Amendment", 3. Hawa, M., 2003. “Stochastic Evaluation of Fair Scheduling with Applications to Quality-of-Service in Broadband Wireless Access Networks”, Ph.D. dissertation, University of Kansas. 4. Wongthavarawat, K. and A. Ganz, 2003. “IEEE 802.16 Based Last Mile Broadband Wireless Military Networks with Quality of Service Support” in Proc. Mil. Commun. Conf, pp: 779-784. 5. Liu, Q., X. Wang and G. Giannakis, 2006. “A Cross-Layer Scheduling Algorithm with QoS Support in Wireless Networks,” IEEE Trans. Veh. Tech., 55(3): 839-847. 6. Niyato, D. and E. Hossain, 2006. “Queue-aware Uplink Banwidth Allocation and Rate Control for Polling Service in IEEE 802.16 Broadband Wireless Networks,” IEEE Trans. Mobile Comp., 5(8): 668-679. 7. Chen, J., W. Jiao and H. Wang, 2005. “A Service Flow Management Strategy for 802.16 Broadband Wireless Access System in TDD Mode,” in Proc. IEEE ICC, pp: 3422-3426. 8. Raghu, K.R., S.K. Bose and M. Ma, 2007. “Queue Based Scheduling for IEEE 802.16 Wireless Broadband,” in Proc. 6th IEEE Int. Conf. ICICS, pp: 1-5. 2007. 9. Sharma, V. and N. Vamaney, 2007. “The UniformlyFair Deficit Round-Robin (UF-DRR) Scheduler for Improved QoS Guarantees in IEEE 802.16 WiMAX Networks,” in Proc. Mil. Commun. Conf, pp: 1-7. 10. Shejwal, A. and A. Parhar, 2007. “Service Criticality Based Scheduling for IEEE 802.16 WirelessMAN,” in Proc. 2nd IEEE Int. Conf. AusWireless, pp: 12-18. 11. Salodjar, N. and A. Karandikar, 2008. “An Indexing Scheduler for Delay Constrained Scheduling with Applications to IEEE 802.16,” in Proc. IEEE WCNC, pp: 1471-1476. World Appl. Sci. J., 15 (4): 576-583, 2011 12. Bai, X., A. Shami and Y. Ye, 2008. “Robust QoS Control for Single Carrier PMP Mode IEEE 802.16 Systems,” IEEE Trans. Mobile Comp., 7(4): 416-429. 13. Kleinrock, L., 1979. Queueing Systems, Volume 1: Theory. Hoboken, NJ: John Wiley and Sons, 14. Kok, A.G.D. and H.C. Tijms, 1985. “A Queueing System with Impatient Customers,” J. Appl. Prob., 22(3): 688-696. 583 15. Jain, D.M.C.R. and W. Hawe, 1984. “A Quantitative Measure of Fairness and Discrimination for Resource Allocation in Shared Systems”, dEC Research Report, TR-301. 16. Hall, J. and P. Mars, 1998. Satisfying QoS with Learning Based Scheduling Algorithm. In 6th International Workshop on Quality of Service, pp: 171-176. 

4 Scheduling Algorithm and Bandwidth Allocation in WiMAX Majid Taghipoor1, Saeid MJafari2 and Vahid Hosseini3 1University of Applied Science and Technology Uromieh, of Computer and IT Engineering, Islamic Azad University of Qazvin, Qazvin, 3Department of Computer and IT Engineering, Computer Engineering Deptt., Urmia University, Iran 2Department 1. Introduction The traditional solution to provide high-speed broadband access is to use wired access technologies, such as cable modem, DSL (Digital Subscriber Line), Ethernet, and fiber optic. However, it is too difficult and expensive for carriers to build and maintain wired networks, especially in remote areas. BWA (Broadband Wireless Access) technology is a flexible, efficient, and cost-effective solution to overcome the problems [1]. WiMAX is one of the most popular BWA technologies today, which aims to provide high speed broadband wireless access for WMANs (Wireless Metropolitan Area Network). [2] WiMAX provides an affordable alternative for wireless broadband access supporting a variety of applications of different types including video conferencing, non-real-time large volume data transfer, traditional voice/data traffic throughput E1/T1 connection, and web browsing.[1] Each traffic flow requires different treatment from the network in terms of allocated bandwidth, maximum delay, and jitter and packet loss [3], [5]. Traffic differentiation is thus a crucial feature to provide network-level QoS (Quality of Service). The standard leaves QoS support features specified for WiMAX networks (e.g., traffic policing and shaping, connection admission control and packet scheduling). One of the most critical issues is the design of a very efficient scheduling algorithm which coordinates all other QoS-related functional entities. The key components in WiMAX QoS guarantee are the admission control and the bandwidth allocation in BS. WiMAX standard defines adequate signalling schemes to support admission control and bandwidth allocation, but does not define the algorithms for them. This absence of definition allows more flexibility in the implementation of admission control and bandwidth allocation. In this study, we focus on evaluating scheduling algorithms for the uplink traffic in WiMAX. We evaluate a number of WiMAX uplink scheduling algorithms in a single-hop network, which is referred to as PMP (Point Multi Point) mode of WiMAX. www.intechopen.com 86 Quality of Service and Resource Allocation in WiMAX 2. Overview of WIMAX In this section, we discuss the WiMAX, the uniqueness of WiMAX uplink scheduling. 2.1 WiMAX standard BWA technology promises a large coverage and high throughput. Theoretically, the coverage range can reach 30 miles and the throughput can achieve 75 Mbit/s [1]. Yet, in practice the maximum coverage range observed is about 20 km and the data throughput can reach 9 Mbit/s using UDP (User Datagram Protocol) and 5 Mbit/s using FTP (File Transfer Protocol) over TCP (Transmission Control Protocol)[2]. WiMAX standard has two main variations: one is for fixed wireless applications (covered by IEEE 802.16-2004 standard) and another is for mobile wireless services (covered by IEEE 802.16e standard). The 802.16 standards only specify the PHY (Physical) layer and the MAC (Media Access Control) layer of the air interface while the upper layers are not considered. The IEEE 802.16 standard specifies a system comprising two core components [6]: the SS (Subscribe Station) or CPE (customer premises equipment) and the BS (Base Station). A BS and one or more SS can form a cell with a P2MP structure. Note that the WiMAX standard also can be used in a P2P (Point to Point) or mesh topology. BS acts as a central entity to transfer all the data from MSs (Mobile Station) in a PMP mode. Transmissions take place through two independent channels: downlink channel (from BS to MS) and uplink channel (from MS to BS). Uplink Channel is shared between all MSs while downlink channels is used only by BS. To support the two-way communication, either FDD (Frequency Division Duplex) or TDD (Time Division Duplex) can be adopted. In the following discussion, we focus on the popular TDD. The IEEE 802.16 is connection oriented. Each packet has to be associated with a connection at MAC level. This provides a way for bandwidth request, association of QoS and other traffic parameters and data transfer. All data transmissions are connection-oriented and the connections are classified into four types, namely, UGS (Unsolicited Grant Service), also known as CBR (Constant Bit Rate), rtVR (real-time Variable Bit Rate), nrtVR(non real-time Variable Bit Rate), and BE (Best Effort). Each service related to type of QoS class can has different constraints such as the traffic rate, maximum latency and tolerated jitter. In section 4 we will focus more on QoS the WiMAX technologies. UGS supports real-time service flows that have fixed-size data packets on a periodic basis. RtVR supports real-time service flows that generate variable data packets size on a periodic basis. The BS provides unicast grants in an unsolicited manner like UGS. Whereas the UGS allocations are fixed in size. NrtVR is designed to support non real-time service flows that require variable size bursts on a regular basis. BE is used for best effort traffic where no throughput or delay guarantees are provided. Those service classes are defined in order to satisfy different types of QoS requirements. However, the IEEE 802.16 standard does not specify the scheduling algorithm to be used. Vendors and operators have to choose the scheduling algorithm(s) to be used. www.intechopen.com Scheduling Algorithm and Bandwidth Allocation in WiMAX 87 2.2 WiMAX MAC layer The 802.16 MAC protocol supports transport protocols such as ATM, Ethernet, and IP, and can accommodate future developments using the specific convergence layer. The MAC also accommodates very high data throughput through the physical layer while delivering ATM-compatible QoS, such as UGS, rtVR, nrtVR, and BE. The 802.16 frame structure enables terminals to be dynamically assigned uplink and downlink burst profiles according to the link conditions. This allows for a tradeoff to occur – in real time – between capacity and robustness. It also provides, on average, a 2x increase in capacity when compared to non-adaptive systems. Fig. 1. Wireless MAC protocol classification [25] According to the architecture topology, there are two main wireless MAC protocols: 1. 2. Distributed MAC Protocols: these protocols are founded on principles of CS and CA, excluding the distributed ALOHA protocol. Centralized MAC Protocols: these protocols are based on communicating with a central entity – in case of cellular mobile communications: the base station. Thus all communications is organized and supervised according to the BS MAC management protocol. [25] There are three types of wireless MAC protocol types: 1. 2. Random Access Protocols: according to this access protocol, for a node to be able to access the network it should contend for the medium. Guaranteed Access Protocols: unlike the random access protocols, the communication between nodes is made on some predefined rules. This may be either in the form polling the nodes one by one, or by token exchanging. www.intechopen.com 88 3. Quality of Service and Resource Allocation in WiMAX Hybrid Protocols: these type of protocols are more superior to the previously mentioned other two protocols, since they are made out of the top properties of random access protocols and guaranteed access protocols. [25]. Hybrid protocols can be further subdivided into two categories: a. b. Random Reservation Access protocol: these are the protocols by the MAC where a periodic reservation of the bandwidth is granted on the reception of a successful request from nodes supported by the central node. Demand Assignment protocol: The MAC allocated bandwidth according to the need of application of the node. The hierarchy of the wireless MAC protocols classification could be illustrated as it is shown in figure 1. Therefore, mobile WiMAX MAC protocol could be classified as the demand assignment protocol; knowing that the mobile WiMAX MAC is designed to support QoS and according to the only MAC protocol that guarantees resources is the DA protocol. [25] Fig. 2. The 802.16 protocol stack [6] WiMAX MAC is subdivided into three sub layers with different functionalities. Figure 2 is a basic illustration of the tasks and services that the MAC sub layers are responsible for. www.intechopen.com Scheduling Algorithm and Bandwidth Allocation in WiMAX 89 The upper MAC layer is the CS (Convergence Sub layer). This sub layer is responsible mainly for classification and header suppression of the incoming packets from the network layer. The classification is done according to the QoS parameters of the packet. Then each service flow is assigned a service flow identifier number. The inner layer is called the MAC CSP (common part sub layer) and it is considered the main sub layer of the MAC layer. Finally, the lower MAC layer is called the MAC security sub layer. Functions like support for privacy, user/device authentication are the responsibility of this sub layer. Figure 3 illustrates the PHY layer with the three sub layers of the MAC layer. The Figure shows the data/control plane only and it regarded as the scope of the standard. Fig. 3. WiMAX MAC and PHY layers – data/control plane [25] www.intechopen.com 90 Quality of Service and Resource Allocation in WiMAX The 802.16 MAC uses a variable-length PDU (Protocol Data Unit) and other innovative concepts to greatly increase efficiency. Multiple MAC PDUs, for example, may be concatenated into a single burst to save PHY overhead. Multiple SDUs (Service Data Unit) may also be concatenated into a single MAC PDU, saving on MAC header overhead. Fragmentation allows very large SDUs to be sent across frame boundaries to guarantee the QoS. Payload header suppression can be used to reduce the overhead caused by the redundancy within the SDU headers. The 802.16 MAC uses a self-correcting bandwidth request/grant scheme that eliminates any delay in acknowledgements, while allowing better QoS handling than traditional acknowledgement schemes. Depending on the QoS and traffic parameters of their services, terminals have a variety of options available to them for requesting bandwidth. SAP (Service Access Point) is entities located in between the sub layers in order to convert the SDU to PDU. Basically, when PDUs of an upper layer are passed through the SAP to a lower layer, they are considered as SDU for that particular lower layer. In TDD mode, a WiMAX MAC frame consists of two sub frames, DL-sub frame for downlink transmission and UL-sub frame for uplink transmission, as shown in figure 4. The DL-sub frame comprises a BP (Burst Preamble) and a FCH (Frame Control Header), followed by DL-MAP, UL-MAP, and a number of downlink payload bursts (DL-PL1, …., DL-PLn). As will be discussed later, the UL-MAP contains the uplink scheduling results, i.e., the uplink bandwidth granted to each SS. The UL-sub frame starts with initial ranging contention slots and bandwidth request contention slots, followed by a number of uplink payload bursts (UL-PL1, …, UL-PLn)(Figure 4). The uplink and downlink bursts are not necessarily equal and their length can be adjusted dynamically in order to adapt to the traffic variation. [29] Fig. 4. WiMAX MAC frame structure [29] 3. Problem statement A WiMAX network is designed to incorporate different types of data streams, and it aims at providing QoS guarantee for all the data streams being served by WiMAX. The WiMAX protocol covers physical layer and MAC layer, and there are several challenges for QoS guarantee in WiMAX. www.intechopen.com Scheduling Algorithm and Bandwidth Allocation in WiMAX 91 The IEEE 802.16 standard provides specification for the MAC and PHY layers for WiMAX and there are several challenges for QoS guarantee in WiMAX. In the physical layer, one challenge is the uncertainty of the wireless channel, which makes the guarantee of broadband wireless data service difficult and renders the static resource allocation scheme unsuitable. In the MAC layer, one challenge is the diversified service types, which requires the WiMAX scheduling scheme to be adaptive to the various QoS parameters of different service types. There have been some studies of the WiMAX MAC scheduling problem [3], [4], [5], and [6]. The key components in WiMAX QoS guarantee are the admission control and the bandwidth allocation in BS. WiMAX standard defines adequate signalling schemes to support admission control and bandwidth allocation, but does not define the algorithms for them. This absence of definition allows more flexibility in the implementation of admission control and bandwidth allocation. The research problem being investigated here is, after connections are admitted into the WiMAX network, how to allocate bandwidth resources and perform scheduling services, so that the QoS requirements of the connections can be satisfied. 3.1 What is QoS? QoS refers to the ability of a network to provide improved service to selected network traffic over various underlying technologies including wired-based technologies (Frame Relay, ATM, Ethernet and 802.1 networks, SONET, and IP-routed networks) and wireless-based technologies (802.11, 802.15, 802.16, 802.20, 3G, IMS, etc). In particular, QoS features provide improved and more predictable network service by providing the following services:      Supporting dedicated bandwidth Improving loss characteristics Avoiding and managing network congestion Shaping network traffic Setting traffic priorities across the network Due to the differences in the wired-based and wireless-based access technologies, the detailed QoS implementations for both tend to be different, however they share common roots. What follows next are the common elements shared between wired-based and wireless-based access methods. 3.2 QoS and scheduling in WiMAX A high level of QoS and scheduling support is one of the interesting features of the WiMAX standard. These service-provider features are especially valuable because of their ability to maximize air-link utilization and system throughput, as well as ensuring that SLAs (ServiceLevel Agreements) are met (Figure 5). The infrastructure to support various classes of services comes from the MAC implementation. QoS is enabled by the bandwidth request and grant mechanism between various subscriber stations and base stations. Primarily there are four buckets for the QoS (UGS, rtVR, nrtVR, and BE) to provide the service-class classification for video, audio, and data services, as they all require various levels of QoS www.intechopen.com 92 Quality of Service and Resource Allocation in WiMAX requirements. The packet scheduler provides scheduling for different classes of services for a single user. This would mean meeting SLA requirements at the user level. Users can be classified into various priority levels, such as standard and premium. Fig. 5. Packet scheduling, as specified by 802.16 [6] 3.3 Scheduling algorithm and their characteristic In some cases, separate scheduling algorithms are implemented for the uplink and downlink traffic. Typically, a CAC (Call Admission Control) procedure is also implemented at the BS that ensures the load supplied by the SSs can be handled by the network. A CAC algorithm will admit a SS into the network if it can ensure that the minimum QoS requirements of the SS can be satisfied and the QoS of existing SSs will not deteriorate. The performance of the scheduling algorithm for the uplink traffic strongly depends on the CAC algorithm. Scheduling has also been studied intensively in many disciplines, such as CPU task scheduling in operating systems, service scheduling in a client-server model, and events scheduling in communication and computer networks. Thus a lot of scheduling algorithms have been developed. However, compared with the traditional scheduling problems, the WiMAX MAC layer scheduling problem is unique and worth study for the following reasons. First, the total bandwidth in a WiMAX network is adaptive since AMC (Adaptive Modelling and Coding) is deployed in the physical layer and the number of bytes each time slot can carry depends on the coding and modulation scheme. Second, multiple service types have been defined and their QoS requirements need to be satisfied at the same time. How to satisfy various QoS requirements of different service types simultaneously has not been addressed by any other wireless access standard before. Third, the time complexity of the WiMAX scheduling algorithm must be simple since real-time service demands a fast response from the central controller in BS. www.intechopen.com Scheduling Algorithm and Bandwidth Allocation in WiMAX 93 Fourth, the frame boundary in the WiMAX MAC layer also serves as the scheduling boundary, which makes the WiMAX scheduling problem different from the continuous time scheduling problem. The above four characteristics make the resource allocation in the WiMAX MAC layer a challenging problem. While some similarities to the wired world can be drawn, there are certain characteristics of the wireless environment that make scheduling particularly challenging. Five major issues in wireless scheduling are identified in [9]:      Wireless link variability: Due to characteristics of the channel as well as location of the mobile subscribers. Fairness: Refers to optimizing the channel capacity by giving preference to spectrally efficient modulations while still allowing transmissions with more robust modulations (and hence, consuming a major amount of spectrum) to get their traffic through. QoS: Particularly for WiMAX, QoS support should be built into the scheduling algorithm to guarantee that QoS commitments are meet under normal conditions as well as under network degradation scenarios. Data throughput and channel utilization: Refers to optimizing the channel utilization while at the same time avoiding waste of bandwidth by transmitting over high loss links. Power constrain and simplicity: Be considerate of the terminals’ battery capacity as well as computational limitations both at the BS and MS. 3.4 Classification scheduling algorithms Packet scheduling algorithms are implemented at both the BS and SSs. A scheduling algorithm at the SS is required to distribute the bandwidth allocation from the BS among its connections. The scheduling algorithm at the SS needs to decide on the allocation of bandwidth among its connections. The scheduling algorithm implemented at the SS can be different than that at the BS. The focus of our work is on scheduling algorithms executed at the BS for the uplink traffic in WiMAX i.e. traffic from the SSs to the BS. A scheduling algorithm for the uplink traffic is faced with challenges not faced by an algorithm for the downlink traffic. An uplink scheduling algorithm does not have all the information about the SSs such as the queue size. An uplink algorithm at the BS has to coordinate its decision with all the SSs where as a downlink algorithm is only concerned in communicating the decision locally to the BS. In general, the scheduling algorithms can be classified as frame-based scheduling and sorted-based scheduling. Frame-based scheduling algorithms include WRR (Weighted Round Robin)[7], DRR (Deficit Round Robin)[8], etc. Sorted-based scheduling algorithms include WFQ (Weighted Fair Queue)[9], also known as PGPS (Packet-based Generalized Processor Sharing)[10], and a number of variations of WFQ such as WF2Q (Worst Case Fair Queuing)[11], SCFQ (Self-Clock Faire Queuing)[12]. The advantage of frame-based scheduling algorithms is their low computing complexity, while the disadvantage is the significant worst case delay. On the contrary, scheduling www.intechopen.com 94 Quality of Service and Resource Allocation in WiMAX algorithm in the WFQ family has better performance in worst case delay, but the algorithm complexity is much higher than that of the frame-based scheduling algorithms. 3.5 Uplink scheduling algorithms In the coming subsections the fundamental scheduling algorithms will be briefly described 3.5.1 Round Robin Round Robin as a scheduling algorithm is the most basic and least complex scheduling algorithm. It has a complexity value of O (1) [13]. Basically the algorithm services the backlogged queues in a round robin fashion. Each time the scheduler pointer stop at a particular queue, one packet is dequeued from that queue and then the scheduler pointer goes to the next queue. This is shown in Figure 6. Fig. 6. RR Scheduler It distributes channel resources to all the SSs without any priority. The RR scheduler is simple and easy to implement. However, this technique is not suitable for systems with different levels of priority and systems with strongly varying sizes of traffic. 3.5.2 Weighted Round Robin An extension of the RR scheduler, the WRR scheduler, based on static weights.WRR [14] was designed to differentiate flows or queues to enable various service rates. It operates on the same bases of RR scheduling. However, unlike RR, WRR assigns a weight to each queue. The weight of an individual queue is equal to the relative share of the available system bandwidth. This means that, the number of packets dequeued from a queue varies according to the weight assigned to that queue. Consequently, this differentiation enables prioritization among the queues, and thus the SSes. [15] www.intechopen.com Scheduling Algorithm and Bandwidth Allocation in WiMAX 95 3.5.3 Earliest deadline first It is a work conserving algorithm originally proposed for real-time applications in wide area networks. The algorithm assigns deadline to each packet and allocates bandwidth to the SS that has the packet with the earliest deadline. Deadlines can be assigned to packets of a SS based on the SS’s maximum delay requirement. The EDF algorithm is suitable for SSs belonging to the UGS and rtVR scheduling services, since SSs in this class have stringent delay requirements. Since SSs belonging to the nrtVR service do not have a delay requirement, the EDF algorithm will schedule packets from these SSs only if there are no packets from SSs of UGS or rtVR class. [16] 3.5.4 Weighted fair queue It is a packet-based approximation of the Generalized Processor Sharing (GPS) algorithm. GPS is an idealized algorithm that assumes a packet can be divided into bits and each bit can be scheduled separately. The WFQ algorithm results in superior performance compared to the WRR algorithm in the presence of variable size packets. The finish time of a packet is essentially the time the packet would have finished service under the GPS algorithm. The disadvantage of the WFQ algorithm is that it will service packets even if they wouldn’t have started service under the GPS algorithm. This is because the WFQ algorithm does not consider the start time of a packet. 3.5.5 Temporary removed packet The TRS (Temporary Removal Scheduler) involves identifying the packet call power, depending on radio conditions, and then temporarily removing them from a scheduling list for a certain adjustable time period TR. The scheduling list contains all the SSs that can be served at the next frame. When TR expires, the temporarily removed packet is checked again. If an improvement is observed in the radio channel, the packet can be topped up in the scheduling list again, otherwise the process is repeated for TR duration. In poor radio conditions, the whole process can be repeated up to L times at the end of which, the removed packed is added to the scheduling list, independently of the current radio channel condition [18]. The temporary TRS can be combined with the RR scheduler. The combined scheduler is called TRS+RR. For example, if there are k packet calls and only one of them is temporary removed, each packet call has a portion, equal to 1 , of the k 1 whole channel resources. 3.5.6 Maximum Signal to Interference Ration The scheduler mSIR (Maximum Signal to Interference Ration) is based on the allocation of radio resources to subscriber stations which have the highest SIR. This scheduler allows a highly efficient utilization of radio resources. However, with the mSIR scheduler, the users with a SIR (Signal to Interference Ratio) that is always small may never be served.[18] www.intechopen.com 96 Quality of Service and Resource Allocation in WiMAX The TRS can be combined with the mSIR scheduler. The combined scheduler is called TRS + mSIR. This scheduler assigns the whole channel resources to the packet call that has the maximum value of the SNR (Signal to Noise Ratio). The station to be served has to belong to the scheduling list. 3.5.7 Reinforcement Learning The scheduler RL (Reinforcement Learning) is based on the model of packet scheduling described by Hall and Mars [23]. The aim is to use different scheduling policies depending on which queues are not meeting their delay requirements. The state of the system represented by a set of N -1 binary variables {s1: sn-1}, where each variable si indicates whether traffic in the corresponding queue qi [24]. There is not variable corresponding to the best-effort queue qN, since there is no mean delay requirement for that queue. For example, the state {0; 0; : : : ; 0} represents that all queues have satisfied their mean delay constraint, while (1; 0; : : : ; 0} represents that the mean delay requirements are being satisfied for all queues except q1. Thus, if there are N queues in the system including one best-effort queue, then there are 2N-1 possible states. In practice, the number of traffic classes is normally small, e.g., four classes in Cisco routers with priority queuing, in which case the number of states is acceptable. At each timeslot, the scheduler must select an action a є {a1: aN}, where ai is the action of choosing to service the packet at the head of queue qi . The scheduler makes this selection by using a scheduling policy Π, which is a function that maps the current state of the system s onto an action a. If the set of possible actions is denoted by A, and the set of possible system states is denoted by S, then Π: S→A. 3.5.8 Hierarchical/hybrid algorithms Hierarchical/hybrid algorithms build on the fact that scheduling services have different and sometimes conflicting requirements. UGS services must always have their delay and bandwidth commitment met, so simply reserving enough bandwidth for those services and controlling for oversubscription would be enough; rtVR services have little tolerance for delay and jitter, so an algorithm guaranteeing delay commitments would be more suitable; and finally, BE and nrtVR will always be hungry for bandwidth with no considerations for delay, so a throughput maximizing algorithm might be preferred. While hierarchical refers to two or more levels of decisions to determine what packets to be scheduled, hybrid refers to the combination of several scheduling techniques (EDF for delay sensitive scheduling services such as rtVR and UGS, and WRR for nrtVR and BE for example). There could be hierarchical solutions that are not necessarily hybrid, but hybrid algorithms usually distribute the resources among different service classes, and then different scheduling techniques are used to schedule packets within each scheduling service, making them hierarchical in nature. A two-tier hierarchical architecture is proposed in [24] for WiMAX uplink scheduling. In the higher hierarchy, strict prioritization is used to direct the traffic into the four queues, according to its type. Then, each queue is scheduled according to a particular algorithm, i.e., www.intechopen.com 97 Scheduling Algorithm and Bandwidth Allocation in WiMAX fixed allocation for UGS, EDF for rtVR, WFQ for nrtVR, and equal division of remaining bandwidth for BE. Although EDF takes care of the delay requirement of the rtVR, grouping multiple rtVR connections into one queue fails to guarantee the minimum bandwidth requirement of each individual rtVR connection. For example, one rtVR connection with tight delay budget may dominate the bandwidth allocation, resulting in starvation of other rtVR connections. In [27], the authors use a first level of strict priority to allocate bandwidth to UGS, rtVR, nrtVR and BE services in that order; and then on a second level in the hierarchy, different scheduling techniques are used depending on the scheduling service: UGS, as the highest priority, has pre-allocated bandwidth, EDF is used for rtVR, WFQ for nrtVR, and FIFO for BE. Similarly, explains an algorithm that uses EDF for nrtVR and rtVR classes, and WFQ for nrtVR and BE classes. In [27], the authors implement a two-level hierarchical scheme for the downlink in which an ARA (Aggregate Resource Allocation) component first estimates the amount of bandwidth required per scheduler class (rtVR, nrtVR, BE and UGS) and distributes it accordingly. In [28], a SC (Service Criticality) based scheduling is proposed for the WiMAX network, where an SC index is calculated in every SS for each connection and then sent to BS, and BS sorts the SC of all the connections and assigns bandwidth according to the descending order of SC. SC is derived according to the buffer occupancy and waiting time of each connection. If a malicious connection always reports a high SC, or a connection is generating excessive traffic to occupy its sending buffer, this connection will dominate the available bandwidth and affect other connections. 4. Evaluation This section presents the simulation results for the algorithms scheduling. For testing performance of algorithms, the introduced algorithms are implemented in the NS-2 (Network Simulator) [20] and WiMAX module [21] that is based on the WiMAX NIST module [20].The MAC implementation contains the main features of the 802.16 standard, such as downlink and uplink transmission. We have also implemented the most important MAC signalling messages, such as UL-MAP and DL-MAP, authentication (PKM), capabilities (SBC), REG (Registration), DSA (Dynamic Service Addition), and DSC (Dynamic Service Change). The implemented PHY is OFDM. Lot size(byte) 108 96 36 24 Channel coding 3/4 2/3 3/4 1/2 modulation 64-QAM 64-QAM QPSK QPSK Table 1. Slot size for OFDM PHY The current implementation also supports differencing MCSs (Moulding Code Scheme). Table 1 shows present slot size for different modulations and channel coding types. www.intechopen.com 98 Quality of Service and Resource Allocation in WiMAX We present a simulation scenario to study thoroughly the proposed scheduling solution. The scenario will present a multi-service case, in which a provider has to support connections with different 802.16 classes and traffic characteristics. The purpose of this scenario is to ensure that the scheduler at the BS takes the service class into account and allocates slots based on the QoS requirements and the request sizes sent by SSs. Another purpose is to test that the scheduler at the BS takes the MAC overhead into account. Table 1 presents information about which applications are active at scenario. Regardless of the simulation scenario, the general parameters of the 802.16 network are the same (see Table 2). There is one BS that controls the traffic of the 802.16 network. The physical layer is OFDM. The BS uses the dynamic uplink/downlink slot assignment for the TDD mode. Both the BS and all SSs use packing and fragmentation in all simulation scenarios. The MAC level uses the largest possible PDU size. ARQ is turned off; neither the BS nor SSs use the CRC field while sending packets. Value OFDM 7MHz 400 TDD OFF Parameter PHY Bandwidth Frame per Second Duplexing mode ARQ/CRC Table 2. WiMAX parameter We consider a general scenario, where n rtVR and/or nrtVR connections are established. Connection i has an arrival rate of ¸i, a delay budget of i, and a minimum reserved bandwidth of MRRi. For the sake of analytic tractability, we assume that the data arrival forms a Poisson process and all queues have infinite size. Other types of traffic (such as the more practical bursty traffic) are studied through simulations. The main parameters of the simulation are represented in Table 3. Effects of these parameters are similar over results of all scheduling algorithms. Moreover, producers of this WiMAX module have used these values for testing performance of their simulator. Parameter Frequency band Propagation model Value 5 MHz Two Ray Ground Antenna model Antenna height Transmit power Receive power threshold Frame duration Cyclic prefix (CP) Simulation duration Omni antenna 1.5 m 0.25 205e-12 20 ms 0.25 100 s Table 3. Main parameters of the simulation www.intechopen.com 99 Scheduling Algorithm and Bandwidth Allocation in WiMAX In particular, we consider several comparable scheduling algorithms, including WRR, EDF, and TRS which is a representative WiMAX scheduling algorithm and has been patented and well received). Besides packet drop rate and throughput that have been studied in analysis, we are also interested in the fairness performance, which is measured by Jain’s Fairness Index [22] defined as follows: (  x i )2 n f ( x1 , x2 ,...xn )  i 1 n n  xi i 1 (1) 2 Where xi is the normalized throughput of connection i, and n is the total number of connections. Each SS establishes a number of connections to the BS in our simulation. We consider ten rtVR connections and ten nrtVR connections. Each type of connection is associated with an MRR and a delay budget. Xi  THi MRi (2) ie, with Thi and MRRi stand for the connection i’s actual data rate and reserved data rate, respectively. The Jain’s Fairness Index ranges between 0 and 1. The higher the index, the better the fairness. If Thi = MRRi for all i, or in other words, every connection obtains its reserved data rate, then xi = 1 for all i, and Jain’s Fairness Index equals 1. All simulations and analytic calculations are done using NS2 simulator. UGS 1 Latency 0.8 0.6 0.4 0.2 0 1 2 3 Traffic load 4 5 Fig. 7. Latency versus traffic Figures 7, 8 show delay packets as a function of the traffic load submitted to the network. The data packets are generated by a streaming multimedia application. The diagram of UGS scheduling algorithm by considering delay is linear where its throughput is increasing. As mentioned above, the UGS traffic request is the highest priority. If a packet is available in this type of traffic it will be sent in no time. For accurate performance evaluation, we adopt the WiMAX physical layer standard OFDM_BPSK_1_2 in our simulations. [24] www.intechopen.com 100 Quality of Service and Resource Allocation in WiMAX UGS 3500 3000 Throughput 2500 2000 1500 1000 500 0 300 800 1400 1900 2400 2900 3400 Traffic Load Fig. 8. Throughput versus traffic The fairness of the scheduling algorithms under bursty traffic is shown in figure 9. As we can see, WRR always maintains almost high fairness, while the fairness of EDF algorithm is the worst among the four algorithms. This is due to the fact that some real time packets rtVR connections are dropped under high burstiness, and thus the throughput of rtVR decreases. [30], [31] rtVR 1.02 1 0.98 0.96 Fairness 0.94 EDF 0.92 WRR 0.9 TRS+mSIR 0.88 10 30 50 Simulation Time Fig. 9. Fairness versus Simulation Time www.intechopen.com 70 90 Scheduling Algorithm and Bandwidth Allocation in WiMAX 101 Figure 10 shows the latency as a function of rtVR+nrtVR traffic load. We verify that the TRS scheduler provides a decrease in the latency. Fig. 10. Latency versus Simulation Time Figure 11 shows the latency as a function of rtVR traffic load. We verify that the mSIR scheduler provides a decrease in the latency. Fig. 11. Latency versus Simulation Time In figure 12, the protocols have been compared on the base of throughput. As you see, TRS+RR throughput is greater than all. www.intechopen.com 102 Quality of Service and Resource Allocation in WiMAX NIST_TRS TRS 0.31 Throughput RR 0.6 0.7 0.8 0.7 0.6 0.5 0.4 10 0.3 EDF WRR LTRS 0.2 0.1 30 1 0.998 0.998 0.998 0.99 0.99 0 NIST_TRS TRS RR RR+mSIR Fig. 12. Throughput 5. References [1] IEEE Standard for Local and Metropolitan Area Networks—Part 16: Air Interface for Fixed Broadband Wireless Access Systems, 2004, IEEE802.16. Available from : http://www.ieeefor 802.org/16/. [2] IEEE Standard for Local and Metropolitan Area Networks—Part 16: Air Interface for Fixed Broadband Wireless Access Systems—Amendment 2: Physical and Medium Access Control Layers for Combined Fixed and Mobile Operation in Licensed Bands, 2005, IEEE802.16e.. Available from: http://www.ieee802.org/16/. [3] Aura Ganz, Zvi Ganz, Kitti Wongthavarawat(18 September 2003). Multimedia Wireless Networks: Technologies, Standards, and QoS, Prentice Hall Publisher. [4] Overcoming Barriers to High-Quality Voice over IP Deployments (2003), Intel Whitepaper. [5] DiffServ-The Scalable End-to-End Quality of Service Model (August 2005), Cisco Whitepaper. [6] WiMAX – Delivering on the Promise of Wireless Broadband (Second Quarter 2006), Xcell Journal - Issue 57. [7] M. Katavenis, S. Sidiropoulos, and C. Courcoubetis. Weighted Round-Robin Cell Multiplexing in A General-Purpose ATM Switch Chip, IEEE J. Sel. Areas Commun., vol. 9, no. 8, pp. 1265–1279, Jan. 1991. [8] M. Shreedhar and G. Varghese (1995), Efficient Fair Queueing Using Deficit Round Robin, in Proc. IEEE SIGCOMM, pp. 231–242. 135. [9] A. Demeres, S. Keshav, and S. Shenker(1989). Analysis and Simulation of A Fair Queueing Algorithm, in Proc. IEEE SIGCOMM, pp. 1–12. www.intechopen.com Scheduling Algorithm and Bandwidth Allocation in WiMAX 103 [10] A. Parekh and R. Gallager(1992). A Qeneralized Processor Sharing Approach to Flow Control: The Single Node Case, in Proc. IEEE INFOCOM , pp. 915–924. [11] J. Bennet and H. Zhang(1996). WF2Q: Worst-Case Fair Weighted Fair Queueing, Procceding of IEEE INFOCOM, 1996, pp. 120–128. [12] S. Golestani (1994). A Self-Clocked Fair Queueing Scheme for Broadband Applications, Procceding of IEEE INFOCOM, pp. 636–646. [13] R. Jain, lecture notes ( 2007), A Survey of Scheduling Methods, University of Ohio. [14] M. Katevenis, S. Sidiropoulos and C. Courcoubetis(1991). Weighted round robin cell multiplexing in a general purpose ATM switch chip, Selected Areas in Communications, IEEE Journal on 9(8), pp. 1265_1279. [15] S. Belenki (2000). Traffic management in QoS networks: Overview and suggested improvements, Tech. Rep. [16] M.Shreedhar and G.Varghese(June 1996). Efficient Fair Queuing using Deficit Round R bin, IEEE/ACM Transactions on Networking, vol. 1, pp. 375‐385. [17] T. Al_Khasib, H. Alnuweiri, H. Fattah and V. C. M. Leung(2005). Mini round robin: enhanced frame_based scheduling algorithm for multimedia networks, IEEE Cmmunications, IEEE International Conference on ICC, pp. 363_368 Vol. 1. [18] Nortel Networks,Introduction to quality of service (QoS)(September 2008), Nortel NetworksWebsite, 2003. [Online]. Accessed on 1st of September 2008. [19] C.F. Ball, F. Treml, X. Gaube, and A. Klein(September 2005). Performance Analysis of Temporary Removal Scheduling applied to mobile WiMAX Scenarios in Tight Frequency Reuse, the 16th Annual IEEE International Symposium On Personal Indoor and Mobile Radio Communications, PIMRC 2005, Berlin, 11 – 14. [20] QoS-included WiMAX Module for NS-2 Simulator. First International Conference on Simulation Tools and Techniques for Communications Networks and Systems, SIMUTools 2008, Marseille,France, March 3-7,2008. [21] The network simulator ns-2(September 2007). Available from : http://www.isi.edu/nsnam/ns/. [22] D. M. C. R. Jain andW. Hawe(1984). A Quantitative Measure of Fairness and Discriminationfor Resource Allocation in Shared Systems, dEC Research Report, TR-301. [23] J. Hall , P. Mars(December 1998). Satisfying QoS with a Learning Based Scheduling Algorithm, School of Engineering, University of Durham,. [24] M.Taghipoor,G Tavassoli and V.Hosseini(April 2010). Gurantee QoS in WiMAX Networks with learning automata, ITNG 2010 Las Vegas, Nevada, USA. 12-14 [25] Ajay Chandra V. Gummalla, John o. Limb.Wireless Medium Access Control Protocols, IEEE Communications Surveys, 2000. [26] Q. Liu, X. Wang, and G. Giannakis(May 2006). A Cross-Layer Scheduling Algorithm with QoS Support in Wireless Networks, IEEE Trans. Veh. Tech., vol. 55, no. 3, pp. 839–847. [27] D. Niyato and E. Hossain ( Dec. 2006). Queue-aware Uplink Banwidth Allocation and Rate Control for Polling Service in IEEE 802.16 Broadband Wireless Networks, IEEE Trans. Mobile Comp., vol. 5, no. 8, pp. 668–679. [28] A. Shejwal and A. Parhar(2007). Service Criticality Based Scheduling for IEEE 802.16 WirelessMAN, in Proc. 2nd IEEE Int. Conf. AusWreless, , pp. 12–18. www.intechopen.com 104 Quality of Service and Resource Allocation in WiMAX [29] H. Chen, thesis (spring 2008). Scheduling and Resource Optimization in Next Generation Hetergeneous Wireless Networks, University of Luoisiana. [30] Jafari, Saeid. M., Taghipour, M. and Meybodi, M. R.(2011). Bandwidth Allocation in Wimax Networks Using Reinforcement Learning, World Applied Sciences Journal Vol. 15, No. 4, pp. 525-531. [31] Jafari, Saeid. M., Taghipour, M. and Meybodi, M. R. (2011).Bandwidth Allocation in Wimax Networks Using Learning Automata, World Applied Sciences Journal Vol. 15, No. 4, pp. 576-583. www.intechopen.com Quality of Service and Resource Allocation in WiMAX Edited by Dr. Roberto Hincapie ISBN 978-953-307-956-1 Hard cover, 376 pages Publisher InTech Published online 03, February, 2012 Published in print edition February, 2012 This book has been prepared to present state of the art on WiMAX Technology. It has been constructed with the support of many researchers around the world, working on resource allocation, quality of service and WiMAX applications. Such many different works on WiMAX, show the great worldwide importance of WiMAX as a wireless broadband access technology. This book is intended for readers interested in resource allocation and quality of service in wireless environments, which is known to be a complex problem. All chapters include both theoretical and technical information, which provides an in depth review of the most recent advances in the field for engineers and researchers, and other readers interested in WiMAX. How to reference In order to correctly reference this scholarly work, feel free to copy and paste the following: Majid Taghipoor, Saeid MJafari and Vahid Hosseini (2012). Scheduling Algorithm and Bandwidth Allocation in WiMAX, Quality of Service and Resource Allocation in WiMAX, Dr. Roberto Hincapie (Ed.), ISBN: 978-953-307956-1, InTech, Available from: http://www.intechopen.com/books/quality-of-service-and-resource-allocation-inwimax/scheduling-algorithm-and-bandwidth-allocation-in-wimax InTech Europe University Campus STeP Ri Slavka Krautzeka 83/A 51000 Rijeka, Croatia Phone: +385 (51) 770 447 Fax: +385 (51) 686 166 www.intechopen.com InTech China Unit 405, Office Block, Hotel Equatorial Shanghai No.65, Yan An Road (West), Shanghai, 200040, China Phone: +86-21-62489820 Fax: +86-21-62489821