A New Channel in Reo with Fuzzy Method

2009 International Conference on Intelligent Networking and Collaborative Systems A New Channel In Reo With Fuzzy Method Reza Tavoli Science and Research branch, Islamic Azad University, Tehran,Iran [email protected] Amir Masoud Rahmani Science and Research branch, Islamic Azad University, Tehran,Iran [email protected] Abstract— Reo is a Coordination Language which is Channel Based that is used in order to connect components and is able to send certain Data through connectors. Most simple Connectors in Reo are Channels. Reo proffers a pattern for composition of Components that can be used as a language for coordinating parallel applications. Reo has been stated crisp for communication and connecting Components and cooperation of Data and non-crisp side of sending Data has not been proffered. Whereas certain applications have non-crisp side and it is not possible to apply them by Reo, therefore in this paper non-crisp side of sending Data by making fuzzy Reo Channel by using fuzzy rules is proposed. that means Data are received by less confidence at the receiver side that by using Reo it is impossible to be applied, therefore there is a need for a mechanism on account of carrying it out that has been pursued in this paper. Method of uncrisped Data by Reo with the help of fuzzy rules is proffered in this paper because fuzzy systems are systems which modulate matters placed in uncrisped space systematically. In next part Reo principles and present channels in Reo has been explained in details. In the third part we will have a summarized review on fuzzy logic. In the forth part making Reo fuzzy by using fuzzy rules has been stated and an application of it will be stated and at the end consequences will be proffered. Keywords-- Reo; Channel Based; Coordination Language; Fuzzy Rules I. II. INTRODUCTION INTRODUCTION OF REO Reo is derived from a Greek word meaning as flow [7]. Reo is a pattern for composition components by using connectors. Composite and complicate connectors in Reo are made using simpler connectors [8]. Simplest connectors in Reo are channels. Channels in Reo are atomic. Each channel has two channel ends; an end is ‘source’ and other one is ‘sink’. Data are flowed in the channel through source end and they are exited from the channel through sink end. A channel can have two ends of source or/and two ends of sink. Reo doesn’t have any limitation upon the behavior of channels. Reo doesn’t have any predefined channels for itself, thus any application is able to define channel, if the behavior of channel is well-defined. There are standard channels in Reo explained further. Reo is a sample for composition of components on the basis of symbol from channels and can send certain Data through connectors between components [1]. Reo emphasizes on a coordinating model channel based that how designers can make more complicated coordinators which are called connectors by simpler connectors. Features of Reo can be Implemented and execute in Reo middle wares. Certain middles wares exist for executing connectors present in Reo such as ReoLite [2], Mocha [3] and Coordination Tool Eclipse [4]. Reo also provides a reconfigurable dynamic for binding connectors at the run time. Connecting web services through connectors and coordinating them are also other applications of Reo [5]. Web services are modular units practically that provides business functionality which is of other practical programs through internet. Several models have been proposed for composition of web services with one another which provides coordination of Static Data related to processes, thus none of these models provides ability of reconfigurable dynamic in participant web services. In fact there is a need for a structure for coordination among web services at the run time which is carried out using Reo. Reo is a mechanism for providing compounding web services [6]. As said before, Reo has been proffered for coordination among components and cooperation of Data among components. But certain applications are present in data communication systems in which Data are sent non-crisp 978-0-7695-3858-7/09 $26.00 © 2009 IEEE DOI 10.1109/INCOS.2009.81 Mohammad Teshnehlab Electrical Eng. K. N. Tossi University Tehran,Iran [email protected] A. Reo Channels The simplest channel in Reo is Sync Channel. In this channel Data are transferred from source end to sink end uniformly, if Data are written simultaneously on the source end. Sync Channel is demonstrated in Figure 1. Figure 1. Sync Channel Sync Drain is a channel that has two source ends. In this channel the act of writing is accomplished on both channels. In this channel Data are red if both ends of Data are written together. Sync Drain channel is demonstrated in Figure 2. 141 one of the sink ends doesn’t do its job properly, the act of writing will be left afloat. Replicator connector is demonstrated in Figure 6.  Figure 2. SyncDrain Channel Sync Spout is opposite to Sync Drain. This channel has two sink ends instead of two source ends. In this channel Data are sent if both ends of Data are red together. Sync Spout channel is demonstrated in Figure 3.  Figure 6. Replicator Connector One of the applications of this connector as a clock is for composite or arrangement orbits that can lead all the clocks at the same time to all Flip-Flops [11]. Merge connector is a connector that is constituted by one sink end and several source ends and sink ends coincide at one point and source end reads the Data, when, all the source ends have had done the act of writing at the same time. Merge connector is demonstrated in Figure 7.  Figure 3. SyncSpout Channel FIFO channel is another channel that has one source end and one sink end and plus these two ends, it enjoys an unbounded buffer too. In this channel, source end stores Data inside the buffer. Sink end receives Data from buffer, if such Data are present in buffer, Data move in buffer in queue order. That means they exist according the way they entered. From another point previous channels defined before are synchronous channels but this channel is asynchronous channel. Synchronous channel is a channel in which Data must be transferred at the same time but in asynchronous channels Data can be transferred not-at-thesame-time order too. FIFO channel is demonstrated in Figure 4. The rectangle in the middle of the Figure is the buffer of the channel.  Figure 7. Merge Connector III. FUZZY LOGIC Theory of fuzzy was introduced in 1965 by an Iranian scientist Professor Latifizadeh [9]. Fuzzy systems are based on rules or knowledge that describes non-determined and non-precise phenomena. Today there are many applications in all grounds for fuzzy theories. Control of industrial process, control of traffic, control of cars, control of the speed of train and etcetera can be taken as examples [10]. 7he heart of a fuzzy system has been a base of knowledge that has been constituted by ‘if-then’ fuzzy rules. An ‘ifthen’ fuzzy rule was an if-then phrase that some of its words have been determined by adjoining membership function. Fuzzy controllers are very simple comprehensively. Designing each fuzzy controller contains three stages. Input stage, Process stage and Output stage. In input stage the operation of fuzzifier is carried out. Input stage contains input variables or control variables. Each variable must enjoy a quantity; for example temperature variable can have certain quantities like cold, warm or hot. Each quantity is written from a quantity to a membership function. Membership functions have been defined once for each of quantities of each of variables. Process stage receives fuzzy ‘if-then’ rules and produces results for each and ultimately compounds the rules related results. There is a relative reason method in this stage for fuzzy act decision-making. This method provides a tool for activating the base of fuzzy Figure 4. FIFO Channel FIFOn channel is a channel from the kind of FIFO which enjoys a buffer constraint by a length of n. Lossy Sync channel is a channel that has a source end and a sink end. This channel is also like Sync Channel but in this channel certain Data are lost. In this channel Data are written at sink end, but Data are transferred to then source end in case sink end accepts the Data, otherwise Data will be lost. Lossy Sync is demonstrated in Figure 5.  Figure 5. LossySync Channel As explained before, composite connectors in Reo are made by simpler connectors. For example Replicator connector is a connector that is constituted by one source end and several sink ends and source ends coincide at one point. The act of writing will be carried out correctly on the source end, when, all the sink ends read the Data. Even if 142 channel. In this paper a new channel is proffered using a fuzzy controller thus to be able to solve this problem. rules. Inference Formula is used for evaluating ‘if-then’ rules in rules base. Interference formula provides a membership function for measuring level of accuracy of comprehensive relation among input and output variables. A membership function which is used a lot is Momdani membership function. We assume that fuzzy rule is as relation number (1): If x is A then y Is N. As said before different barriers exist in the channel and in this paper factor of attenuation is solely discussed. Attenuation means deduction of Current, Voltage and Signal Power, or Data in transferring among two components [12]. A significant parameter in any transferring system is Signal Power. By spreading signal over transferring medium, attenuation will be created on the signal power. Therefore signal power must be increased while it is being deduces. Usually attenuation is by the unit of decibel. Decibel is a criterion on the basis of the proportion of two signal level. Rate of attenuation on decibel is as relation number (3): (1) Momdani membership function is as relation number (2): (2) µ AÆN(x,y)=µ A(x)ŀµ N(y) When µ A(x) is membership rate of x in A. µ N(y) is the membership rate of y in N and µ AÆN(x,y) is the membership rate of comprehensive relation between x and y with Minimum Operator (Ŋ). LdB=10log10(Pin/Pout)  (3) Whereas LdB is attenuation rate on decibel, Pin level of input ability power, Pout level of output ability power. Opposite to attenuation, there is gain. In fact Data increases signal power. Rate of gain on decibel is as relation number (4) : The last stage is Defuzzifier. Defuzzifier converts the amount produced from the previous stage to a crisped amount. Figure 8 demonstrates fuzzy controller with fuzzifier and defuzzifier. GdB=10log10(Pout/Pin) (4) Whereas GdB is the rate of gain on decibel, Pin level of input ability power, Poutlevel of output ability power Designing fuzzy channel is such ways that signal is sent by a level of power from the sender towards the receiver. After that the receiver by a feedback sends the rate of sent signal power rate in the form of control related data, towards the receiver, sender also sends the desired signal by a measured power by fuzzy channel while receiving such data and determining attenuation rate according to relation number (3) by using fuzzy rules that will be explained further. Proposed fuzzy channel is demonstrated in Figure 9.  Figure 8. Fuzzy Controller With Fuzzifier and Defuzzifier Figure 9. Fuzzy Channel . IV. As said before for a fuzzy controller system there is a need for fuzzy input, fuzzy inference engine and at the consequence fuzzy output. In this system input variables is attenuation and signal power that quantities related to attenuation are ‘very low’, ‘low’, ‘medium’, ‘high’, and ‘very high’ and related to power are ‘low’, ‘medium’ and ‘high’. PROPOSED MODEL FOR FUZZY REO In any communication system, the received sent signal is possible to be different than sent signal, according to the presence of different kinds of barriers [12]. For example for Analog Data, such barriers can reduce the Data quality. There is a bit error for digital Data. 1 binary is converted to a zero binary and vice-versa. Attenuation, delay distortion and noise can be pointed as certain barriers transferred in a system. According to the fact that Reo uses channel for transferring the Data among components, channels that until now are present in Reo for transferring there Data among components, are not appropriate, when barriers exist in the Determined Range for attenuation has been assumed from 0 to 10 decibel and determined Range for signal power has been assumed from 0 to 6 watt. A fuzzy input variable has been demonstrated in Figure 10 and 11. 143 Certain fuzzy rules in this system have been defined as below: • If (Attenuation is very low and power is high) then (power is low) • If (Attenuation is low and power is high) then (power is medium) • If (Attenuation is medium and power is low) then (power is medium) • If (Attenuation is high and power is low) then (power is high) • If (Attenuation is very high and power is medium) then (power is very high  Radio sender can be pointed as to be one of the other application of fuzzy channel power in which consumption power is very low; more sending signal power from radio sender, more energy is consumed, whereas signal power can be reduced by using Reo fuzzy channel in such a way the Data is not possible to be received by the receiver properly even as yet.  Figure 10. Fuzzy Input(Attenuation) Advantage of proposed fuzzy channel towards ordinary channels in Reo is that signal of Data are sent towards the receiver by a more appropriate power, that means such Data are received in the receiver by more certainty. Low speed is a disadvantage of this channel because after sending the Data from the sender to the receiver a packet is sent to the receiver which contains receiving power of receiver. Figure 11. Fuzzy Input(Signal Power) In this system, fuzzy output is signal power that this variable also accepts ‘low’, ‘medium’ and ‘high’ amounts. Determined Range for signal power has been assumed from 0 to 6 Watt. V. SIMULATION AND RESULTS In a radio system when sender sends Data to the receiver, distances between two nodes of sender and receiver has direct relation to the attenuation, in such a way that by increasing and decreasing of the distance, attenuation becomes high and low, accordingly. On other side attenuation has direct relation to frequency. Relation number (5) demonstrates the relation among attenuation, distance and frequency[12]: In fact fuzzy controller operates in such a way that for example if attenuation is high or very high in the channel and power of sending signal is low, signal power must also increase and if attenuation in channel is minor and power of sending signal is high, ability on power of signal must be low, such task caused consumption energy not to be wasted and also caused the sending frame on behalf of sender not to be demolished. Fuzzy output variable has been demonstrated in Figure 12. FSL=94.2+20LOG(D)+20LOG(F)  (5) In above formula FSL is the free space loss or attenuation which is calculated on dB, d is the distance between sender and receiver on Km and f is the frequency of sender on GHz. It is assumed that two nodes of sender and receiver in 20 Range of time changes place in ‘Random Way Point’ in such a way that each node of sender/receiver chooses a place randomly and moves towards this place by a steady speed. Then after sending the desired place it stays there for 30 seconds and again it chooses a new random place and moves towards it. At the time of movement no Data is sent on behalf of sender. It has been seen that with input frequency 30 Khz attenuation rate or free space loss for  Figure 12. Fuzzy Output(Signal Power) 144 different seasons between sender and receiver is as shown in Figure 13. According to Figure 14 attenuation rate of 6 to 9 decibel, sender operates by its maximum sending power that is around 6 Watt, but in attenuation of 0 to 2 decibel, it operates by its minimum power that is 0.8 Watt. In this system minimum attenuation rate has been assumed for 0. As demonstrated in Figure 14, sending power increases and decreases according to attenuation rate which caused the energy not to be wasted in time when the attenuation is low. It also causes sending packets not to be demolished when attenuation is high between two nodes. According to observation, rate of energy consumption, when packets are sent by ordinary Reo channels for power of fixed sender of 3,8 Watt is calculated by relation number (6)[13]:  E =P * T = 3.8 * 20*30 = 2280 J Figure 13. Relation Between Distance and Free Space Loss In this system it is assumed that in each second a packet is sent to the receiver from the sender and maximum power of the sender has been assumed to be 6 Watt. Therefore after the desired packet was sent by primary power of P0 through fuzzy channel, attenuation rate on the channel according to the distance between two nodes will be calculated by relation number (5). After calculate attenuation rate on the channel, sending power of the sender at the time left in this Range of time is corrected according the proposed fuzzy controller. For example here primary power has been assumed for P0=3.8w and primary attenuation rate according to the distance between two nodes is 1,03 therefore due to attenuation rate being 'very low' and fuzzy rules, sending power will be 0,772 Watt, that means in the time left (29 seconds) the Data is sent by an power of 0,772 Watt towards the receiver and by the movement of two nodes in next Range of time, attenuation will vary according to the distance between two nodes; continuance of examination procedure is as related above. Figure 14 demonstrates the relation between attenuation and sending power according to the carried out examination in this system. (6) In above relation p=3.8w is the power of each packet and t=600s is 20 Range of 30 seconds. Rate of energy consumption that packets are sent by ordinary Reo channel is calculated by relation number (7) which is 34 percent lower as compare to ordinary Reo channel. E=P × T=(P1+P2+…+P19+P20) × T=(0.772 + 5.24 + … + (7) 5.78 + 0.863) × 30=1512.93 Therefore, by Reo fuzzy channel, less energy has been consumed as compare to ordinary Reo channel, since in cases when attenuation between two nodes is low, the desired packet is sent by a lower power towards the receiver. On other side, plus the rate of energy consumption in Reo fuzzy channel, rate of more healthy packets are sent through his channel because by low and medium sending power and high rate of attentuaion, the sent packet will be demolished, whereas by using fuzzy channel the sending power of the sender of packet increases and less Data will be demolished. In case rate of attenuation is ‘very high’ the packet will be demolished either they are sent through Reo fuzzy cannel or through ordinary channel. Figure 15 demonstrates the rate of proper receiving of the sent Data through Reo fuzzy channel and ordinary Reo channel in 20 Rang Of time   Figure 14. Relation Between Free Space Loss In this System 145 REFERENCES Fuzzy Channel Data transmission Truth Ordinary Channel [1] Data transmission Truth 1.2 1 [2] 0.8 0.6 [3] 0.4 0.2 [4] 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Period Of Time  [5] Figure 15. Data Transmission Truth [6] In 18th Range of time by the help of fuzzy channel, the receiver has received the Data properly but in ordinary channel, due to the sending power being low, Data in receiver has not been received properly. [7] According to observation in this examination in case Data are sent through Reo fuzzy channel 75 percent of the Data will be reached healthy to the sink and through ordinary Reo channel 65 percent will be reached properly. VI. [8] [9] [10] CONCLUSION In this paper, fuzzy Reo channel has been proffered. In this channel, fuzzy method has been used. The advantage of such channel as compare to ordinary Reo channel is that it is able to send Data non-crisped which causes energy not to be wasted and also the amount of sending Data is more as compare to ordinary Reo channel. Disadvantage of such channel is the low speed of it. Radio sending is of the other applications of this channel. [11] [12] [13] . 146 F. Arbab, Reo:" A Channel-based Coordination Model for Component Composition, Mathematical Structures in Computer Science", 14(3):1-38, 2004. D. Clarke, ReoLite: Reo coordination engine, CWI, the Netherlands, http://homepages.cwi.nl/~dave/reolite/. F. Arbab, F.S. deBoer, M.M. Bonsangue and J.V.Guillen-Scholten, “MoCha: a middleware based on mobile channels”, In Proceedings of 26th International ComputerSoftware and Application Conference (COMPSAC02), IEEE Computer Society Press, 2002. C. Koehler, Eclipse Coordination Tools, CWI, the Netherlands, http://homepages.cwi.nl/~koehler/etc. Soheil Saifipoor, Behrouz Tork Ladani, Naser Nematbakhsh, " A Dynamic Reconfigurable Web Service Composition Framework Using Reo Coordination Language " , Fifth European Conference on Web Services: 203-212 , 2007. F. Arbab, and N.K. Diakov,“Compositional construction of Web services using Reo”, In Proceedings of Internationl Workshop on Web Services: Modeling, Architecture and Infrastructure (ICEIS 2004), Porto, Portugal, April 13-14, 2004. F. Arbab and F. Mavaddat, Coordination Through Channel Composition”, In F. Arbab and C. L. Talcott, editors, Proceedings of the 5th International Conference on Coordination Models and Languages (Coordination 2002), York, UK, April 8--11, Lecture Notes in Computer Science, pp. 21-38, Springer-Verlag, 2002. F. Arbab, “Abstract Behavior Types: A foundation model for components and their composition", Lecture Notes in ComputerScience, Springer-Verlag, Vol. 2852, pp. 33-70, 2003. Zadeh, L. A. (1965). Fuzzy sets, ,QI_ DQG FRQWURO _: 338–353. Dr. Mohammed Y. Hassan, Member, IAENG Waleed F. Sharif, M. Sc. "Design of FPGA based PID-like FuzzyController for Industrial Applications",IAENG International Journal of Computer Science, 34:2, IJCS_34_2_05,Advance online publication: 17 November 2007. Niloofar Razavi, Marjan Sirjani," Using Reo for Formal Specification and Verification of System Designs ",2006 4th IEEE/ACM International Conference on Formal Methods and Models for CoDesign,113-122,2006. Stallings, W. SNMP, SNMPv2, SNMPv3, and RMON 1 and 2. Reading,MA:Addison-Wesley,1999. Dieter, C., "Europe Begins the Slow Creep Towards Deregulation", Power Engineering International, pp.22-24, May 1998.