[attachment=14366]
INTRODUCTION
Traditionally, mathematical models have been usedto solve many real world problems. In conventional mathematical models, it is not possible to capture all aspects of the real world phenomena. In particular, lack of certain types of information, errors in measurement while collecting data, and at times the presence of incorrect information, all give rise to uncertainty and undermine the confidence in the results obtained from the models. Proper representation of uncertainty and its incorporation into the models could be of genuine use in constructing more realistic models of real world problems.The mathematical representation of uncertainty has received considerable attention since theintroduction of Fuzzy Set Theory by Zadeh in the year of 1965.The concept of Fuzzy Sets is based on the grouping of the elements into classes that do not have sharply defined boundaries. These classes are used to describe ambiguityand vagueness in mathematical models of empirical empirical phenomena. A binary system is inadequate to represent these systems because certain aspects of reality always escape these models. A statement can bepartially true and its truth value is treated in a relativistic sense rather than the absolute sense. This allows large flexibility in inferring information and in modeling problems that are too complex or ill-defined to be modeled otherwise. Fuzzy Set Theory provides for a mathematical representation of ambiguity and level of perception of the truth value of information.
FUZZY SYSTEMS
Fuzzy logic is a computational paradigm that provides a mathematical tool for representing and manipulating information in a way that resembles human communication and reasoning processes . A fuzzy system is a rule-based system that uses fuzzy logic, rather than Boolean logic, to reason about data Its basic structure includes four main components, as depicted in Fig. 1: (1) a Fuzzifier, which translates crisp (realvalued) inputs into fuzzy values; (2) an inference engine that applies a fuzzy reasoning mechanism to obtain a fuzzy output; (3) a Defuzzifier, which translates this latter output into a crisp value; and (4) a knowledge base, which contains both an ensemble of fuzzy rules, known as the rule base, and an ensemble of membership functions, known as the database.
ABSTRACT:
Recently, the number of internet users is increasing therefore transferred data is also increased and routers should transfer much more packets. Congestion occurs when arrival rate to the router is greater than its departure rate; furthermore for solving this problem, it is necessary to design an effective queue management algorithm. According to dynamic of input packets and different packet priorities queue management becomes very complex. For this reason, it is better to use an intelligent algorithm. In this paper we use a fuzzy controller called Fuzzy-Multiple Queue Management (MQM) for solving the mentioned problem. In this controller sender and receiver routers communicate with each other for balancing the rate of transferred packets. The Fuzzy-MQM algorithm is also compared with the Multiple Queue Management algorithm and simulation results show that proposed algorithm has better performance, throughput and also low packet loss.
Congestion control in current Internet still is a critical issue. The number of Internet users is rapidly increasing and therefore the amount of data to be carried also
increases. Furthermore to support new Internet applications such as voice over IP, it is necessary to design effective Quality of Service approaches. Congestion control provides quality of service over the best effort networks. Congestion occurred when arrival rate to a router is greater than its departure rate. Each router in the network uses queue management (QM) and scheduling as two classes of algorithms that are related to congestion control. The QM algorithms try to control the length of packet queues by dropping packets when appropriate. Scheduling algorithms on the other hand, determine which packet to drop next and which is to send and also used to manage the allocation of bandwidth among flows. Network congestion induced by traffic leads to wasting all the resources that the packet consumed on its way from source to destination. Recently, many QM schemes have been proposed to provide high network utilization with low loss and delay by regulating queues at the bottleneck links in TCP/IP best-effort networks, including random early detection (RED), adaptive RED (A-RED) [2], random exponential marking (REM) [3], and core-stateless label fairness . The QM approach can be contrasted with the “Tail Drop” (TD) QM approach, employed by commoninternet routers, where the discard policy of arriving packets is based on the overflow of the output port buffer. Contrary to TD, QM mechanisms [5] start dropping packets earlier in order to be able to notify traffic sources about the incipient stages of congestion.QM allows the router to separate policies of dropping packets from the policies for indicating congestion. The use of Explicit Congestion Notification (ECN) [6] was proposed in order to provide TCP an alternative to packet drops as a mechanism for detecting incipient congestion in the network. A QM-enabled gateway can mark a packet either by dropping it or by setting a bit in the packet’s header if the transport protocol is capable of reacting to ECN. The use of ECN for notification of congestion to the end-nodes generally prevents unnecessary packet drops.
In [7], is proposed a QM algorithm that considers both the average queue length and the estimated packet arrival rate together in order to detect incipient congestion. It algorithm predicts the average queue length and controls it to maintain a certain reference value to achieve high page link utilization and low queuing delay. It also has low complexity and easy configuration. In communication networks that employ TD queuing and additive-increase multiplicative-decrease (AIMD) congestion control algorithms are discussed.In this paper, we design a new fuzzy controller called Fuzzy-MQM for avoiding the congestion problem to get higher performance. Fuzzy controller is a widely used computational intelligence technique for dealing with information processing. It mechanism provides a system for designing feedback control algorithms in such cases where the system to be controlled is too complex to employ classical control methods. This algorithm is represented typically in the form of linguistic rules, which is gained through ex perience by human operators or other researchers who may be well familiar with the system to be controlled.
this paper, we design a new fuzzy controller called Fuzzy-MQM for avoiding the congestion problem to get higher performance. Fuzzy controller is a widely used computational intelligence technique for dealing with information processing. It mechanism provides a system for designing feedback control algorithms in such cases where the system to be controlled is too complex to employ classical control methods. This algorithm is represented typically in the form of linguistic rules, which is gained through experience by human operators or other researchers who may be well familiar with the system to be controlled.
QM mechanisms
QM mechanisms aim to provide high page link utilization with low loss rate and queuing delay, while responding quickly to load changes. Several schemes have been proposed to provide congestion control in TCP/IP networks. RED [1], which was the first QM algorithm proposed, simply sets some minimum and maximum marking thresholds in the router queues. In case the average queue size exceeds the minimum threshold, RED starts randomly marking packets based on a probability depending on the average queue length, whereas if it exceeds the maximum threshold every packet is dropped.Recently, new proposed QM mechanisms have appeared to give alternative solutions, and approached the problem of congestion control differently than RED, due to the difficulties of appropriately setting RED parameters based on dynamic network conditions. Specifically, REM algorithm uses the instantaneous queue size and its difference from a target value to calculate the mark probability based on an exponential law. A-RED adjusts the value of the maximum mark probability to keep the average queue size within a target range half way between the minimum and maximum thresholds. A-RED also specifies a procedure for automatically setting the RED parameter of queue weight as a function of the page link capacity.
The Fuzzy-MQM algorithm
The Fuzzy-MQM algorithm in each router is developed based on multiple queues that flows the incoming packets to one of the queues with respect to the priority and in order to increase performance efficiency, a fuzzy controller algorithm has been used. This structure has multiple queues that are designed for different packet priorities. Three queues are used in each input page link attached to a router to keep packets with high, medium and low priority. The duty of the classifier is to separate the packets arriving into the router based on their priority and to send them to the relevant queue. The scheduler is responsible to selecting packets that must be sent to the output link. When an input queue is full, the newly arrived packets are dropped. On the other hand, each one of the packets has a specific deadline and the packet shall be sent before the deadline expires; otherwise, it will be dropped.
The Fuzzy-MQM controls the arrival and departure rates of packets with different priorities in order to save the router of congestion and overload subsequently prevent
the network from congestion. In this algorithm if the arrival rate of packets with specific priority to a queue is high, since the queue quickly becomes full, the controller shall send a message to the sender router concerning a decrease in the rate of sending packets with that specific priority. In addition, if a queue with specific priority be in threshold of filling, the process of sending packets from that queue shall be performed more than other queues.The fuzzy controller used in this algorithm, as Fig. 3shows, has three inputs, namely L1, L2 and L3 which are the status of high, medium and low priority queues.According to status of the inputs the controller produces two types of outputs; I or O. Each one of the inputs and outputs mentioned above use some membership functions, as shown in Fig. 4 and 5. The fuzzy controller
has a rule base, according to table 1, which makes decisions on the basis of the rules in it. For example if the high priority queue is Full, the medium priority queue is Empty and the low priority queue is Full therefore inference engine searches the rule base and finds the Rule 13 of Table 1. It is clear that the high and low priority queues are in critical situations and according to the Rule13 more packets must be sent from the high priority queue. On the other hand because the medium priority queue is empty we notify the sender router to send more packets from this priority.