IJCSI Iteratioal Joural of Computer Sciece Issues, Vol. 10, Issue 4, No 1, July 2013 ISSN (Prit): 1694-0814 ISSN (Olie): 1694-0784 www.ijcsi.org 11 A New Eergy Cosumptio Algorithm with Active Sesor Selectio Usig GELS i Target Coverage WSN Ali Bagrezai 1, Seyed Vahab AL-Di Makki 2, Ali Shokouhi Rostami 3 1 Departmet of Commuicatio, Kermashah Sciece ad Research brach, Islamic Azad uiversity, Kermashah, Ira 2 Electrical Departmet, Egieerig Faculty, Razi Uiversity, Kermashah, Ira. 3 Departmet of Computer, Islamic Azad Uiversity, Behshahr Brach, Ira Abstract I wireless sesor etwork, due to impossibility of replacig battery, the problem of eergy ad etwork lifetime is oe of the importat parameters. I asymmetric sesor etworks, due to limited rage of ormal sesors it is ot possible to commuicate directly with cetral statio by these sesors. I oted etwork, maager odes are used which have more eergy, processig power ad broader telecommuicatio rage. Coectivity ad sedig iformatio to cetral statio are doe through them. The optimal selectio ad cosiderig the eergy of itermediate odes to select ad trasmit data ad also icreasig etwork lifetime is oe of the most importat parts of wireless etwork desig. I this paper, a gravitatioal force algorithm is used to solve the problem that is a power aware Selectio algorithm i sesor etwork. Keywords: asymmetric sesor etwork, poit coverage, the etwork eergy, gravitatio, velocity, Newto's law 1. Itroductio Recet techology developmets i micro-electromechaical systems ad i itegrated circuits led to developmet of small sesors with high processig power of iformatio ad low power cosumptio. These sesors have umerous applicatios such as multimedia, medical, surveillace, military telecommuicatios ad home applicatios. Pocket PC, pager ad cell phoes are amog them. A set of these sesors make a powerful etwork as wireless sesor etwork that is able to sample from local values,process ad sed them to other sesors ad fially to mai observer (user). Service quality is a versatile combiatio with multiple meaigs ad is oe of etwork desiger's goals.i order to achieve such desig, may mechaisms are desiged ad evaluated [1,2]. The mai challege i wireless ad mobile systems desig Origiate from two mai sources of these systems i.e. telecommuicatio badwidth ad eergy. To solve these limitatios it is eeded to Desig telecommuicatio techiques to icrease badwidth eeded for each user ad desig powerful protocol for efficiet use of eergy. Desigs will be differet Depedig o expected capabilities of system ad i various applicatios. For example, i may applicatios, the optimal umber of odes ad cosumed eergy i executive rouds, ad maximizig etwork lifetime are basic requiremets of etwork. (I etworks classificatio, Time iterval of etwork activity is divided ito certai parts so that each iterval ust after choosig selected category is activated i size of that time ad other odes of etwork will tur off. This part is called a roud). Clusterig is a solutio for this. To collect ad aggregate data i a sesor etwork, the odes ca be orgaized i small groups called cluster. Each cluster cotais a cetral ode called cluster heads ad some member odes. A twolevel hierarchy of cluster heads (i high level) ad member odes (i low level) is costructed through clusterig [3]. As replacig battery i may applicatios is ot appropriate, low eergy cosumptio is oe of the basic eeds i these etworks ad lifetime of each sesor ca be effectively icreased by optimizig eergy cosumptio [4]. Schemes that are efficiet I terms of power have applied more i these etworks. These schemes are beig ivestigated i all layers of etwork i two aspects of hardware desigig ad algorithm ad protocol desigig. Oe way to reduce eergy cosumptio is to decrease the umber of sesors i sesig area to esure idetificatio of each target i the area. If the etwork is scalable, the Algorithms to decrease umber of sesors ca be efficietly implemeted [5]. Copyright (c) 2013 Iteratioal Joural of Computer Sciece Issues. All Rights Reserved.
IJCSI Iteratioal Joural of Computer Sciece Issues, Vol. 10, Issue 4, No 1, July 2013 ISSN (Prit): 1694-0814 ISSN (Olie): 1694-0784 www.ijcsi.org 12 Fig. 1.Poit coverage [6] I poit coverage, the aim is to produce coverage i set of poits. A set of sesors is show i figure 1 that are radomly arraged to cover a set of targets (square odes which are gree). The coected black odes produce a set of active sesors that are result from a timig mechaism [6]. Poit coverage scearios have may applicatios. I this sceario, a umber of targets with certai positio are cosidered that should be cotrolled. A large umber of sesors are radomly distributed Very close to targets. These sesors sed collected iformatio to cetral processig ode. Based o this method, each target should be cotrolled by at least oe sesor at ay momet, assumig that each sesor is able to cotrol all targets i its sesig rag. Oe way to reduce eergy cosumptio is to decrease umber of active sesors i Coverage area. A method for icreasig sesor etwork lifetime through savig eergy is to divide a set of sesors ito several separated sets. This classificatio should be i a way that each set covers all targets completely. These separated sets are activated cosecutively so that oly oe set is active at ay momet [7]. 2. Related works May researches have bee ivestigated i the field of power aware algorithm ad optimizatio of power cosumptio. Carbuar ad et al preseted a way to save eergy cosumptio by detectig positio of sesors ad decreasig their overlappig [8]. I [6], a method is preseted to save eergy cosumptio. Based o this method, each area of sesors limit is divided ito two sets. Oly a set of them are active at ay time ad they will be activated alteratively. Based o the method preseted i [7], at first, odes are active or iactive distributive to obtai cosidered coverage rage ad remai uchaged. I the etworks that the odes distributed statistically, there is problem of heterogeeous distributio of eergy i odes. I fact, as the sesor is closer to target, its eergy cosumptio is more. So, coectivity ad etwork coverage will ot be esured completely [9]. I [10], a method is preseted to achieve a scalable coverage. This method is used to ehace eergy efficiecy whe there is high computatioal complexity ad slag. I method proposed i [11], at first the sesors are radomly distributed the a self-healig algorithm is used to produce a complete coverage. Eergy optimizatio is obtaied based o eergy optimizer algorithm. Numerical simulatio verifies lower eergy cosumptio i this etwork i compariso with iitial radomly distributed etwork. I [12], it is oted that oe way to reduce eergy cosumptio is to decrease eergy cosumptio i boudaries of covered regios. I [13], a method is preseted to reduce the umber of sesors ad eergy cosumptio based o biological algorithms. The advatage of this method compared to other methods is uiform distributio of sesors. I [14], a method is preseted to icrease etwork lifetime which is based o maximizig the umber of sesor classes. I this method, a ode is allowed to be a member of more tha oe group which will icrease etwork lifetime. I [14], a relatio is preseted for siks velocity, eergy optimizatio ad reductio of data packets Failure possibility. I aother research, i paper [15] Usig data trasmissio i multiple paths the etwork is resistat to ode Failure. Here, each ode determies its ext hop based o a ode which has the highest residual eergy. I [16], a method is preseted for clusterig. This method is based o maximum delay ad wasted eergy by itermediate odes ad cluster size. This algorithm is based o creatig a spaig tree whose root is ode of cluster head. I aother research, i [17], ulike other papers about sesig rage of ode, a disk with fixed radius aroud the ode is ot cosidered. I this paper, a scheme is preseted to reduce etwork power cosumptio by establishig cooperatio betwee odes. I article [18], Data Compressio Problem i wireless etworks is formulated accordig to eergy. I this model, a percetage of each sesor data does't sed; based o data correlatio, therefore, result i reduced eergy cosumptio. It is also show that the greedy method is the most optimal method i terms of eergy cosumptio. I the paper [19], a umber of itermediate odes are distributed i the etwork to save eergy. This paper aims to obtai distributio of itermediate odes i etwork So that the etwork lifetime icreases. I [20], hierarchical clusterig is used i wireless etworks to achieve lower eergy cosumptio. Cosiderig the cluster head, eergy cosumptio eeded for commuicatio of each ode with processig ceter will decrease. I may studies icludig refereces [5-8], clusterig methods are preseted to reduce the umber of clusters. I [21], methods are proposed to decrease power cosumptio ad icrease Copyright (c) 2013 Iteratioal Joural of Computer Sciece Issues. All Rights Reserved.
IJCSI Iteratioal Joural of Computer Sciece Issues, Vol. 10, Issue 4, No 1, July 2013 ISSN (Prit): 1694-0814 ISSN (Olie): 1694-0784 www.ijcsi.org 13 lifetime of target coverage etwork. I this paper, the advatage of usig a algorithm based o greedy protocol is preseted. I [22], to select relay stage, two algorithms based o clusterig -the shortest distace ad greedy algorithm are compared. I preseted greedy algorithm, Classes of odes that do ot have maager ode are merged with other classes of odes. This process cotiues util all classes will obtai maager odes. Although this method has low computatioal ad telecommuicatio complexity, it is ot a optimal method to decrease etwork eergy cosumptio. Aother method preseted i [22] is the shortest distace algorithm. I [23], a method is preseted to icrease etwork lifetime based o gravitatioal algorithm. This algorithm aims to icrease etwork lifetime by optimizig ad decreasig eergy cosumptio ad icreasig productivity of moitorig etwork. I this paper, each ode that is active i curret roud fids its shortest distace to closest maager ode. The shortest distace Selectio decreases etwork eergy cosumptio. 3. Eergy model This algorithm is based o timig protocol of activity duratio etworks. Timig protocol is oe of the groupig protocols which are placed i sesor etworks. It uses twostep mechaism (iitiative ad executive) ad works o the basis of data commuicatio i shape of sigle-hop or multi-hop Icludig some super odes ad relay ad moitorig sesor. I this protocol, group selectio is doe by usig size fuctio desiged i the protocol. I iitiative phase, some odes which is called "sesor" sed their propagatio messages to their eighbors. I secod phase (executive) which is kow as stable phase, data receptio or trasmissio is doe from sesor odes to relay odes ad from relay odes to destiatio. Figure 2 shows the pla of protocol operatio. Some odes of super odes trasmit data carefully, like LEACH algorithm [15], Accordig to the pla The eergy is saved by groupig i remaiig time of iactive odes. I groupig protocols, eergy cosumptio is costat i whole etwork due to periodic circulatio of active sesors. Hece, we used this feature i our paper. As show i Figure 2 each roud icludes two phases: iitiative phase ad executive phase. The iitiative phase icludes two parts. The former is devoted to moitorig sesors selectio. The latter is for relay sesors selectio. It is obvious that usig super odes icreases etwork lifetime. Fig. 2. Timelie of proposed protocol Performace 3.1 Eergy Model The eergy model is cosidered for trasmittig ad receivig oe of data i accordace with LEACH eergy model. Assume that the distace betwee a trasmitter ad a receiver is d i eergy model metioed above. If d is more tha d0, the multi-path model (with less path coefficiet 4) is used; otherwise ope space model (with less path coefficiet 2) is used. E ( l, d) E Tx Tx elec Eelect mp ad fs ( l) E Tx amp le ( l, d) le elec elec 2 l fsd d d0 (1) 4 l d d d Is required eergy to activate the electrical circuit fs are activatio eergies for power amplifiers i multi-path ad ope space modes, respectively. Its geeral form is represeted: (1) With costat coefficiets p ad q (2) I receiver case E Tx ( l, d) p qd (2) The cosumed eergy is received with oe of data sizes (3). E Rx ( l) E ( l) le p (3) Rx elec elec I preseted asymmetrical etworks, it is assumed that iitial eergy of super odes is several times greater tha iitial eergy of ormal sesors. The cosumptio eergy of a relay ad moitorig ode are deoted by Es 1 ad i each roud respectively. Ec 1 0 Copyright (c) 2013 Iteratioal Joural of Computer Sciece Issues. All Rights Reserved.
IJCSI Iteratioal Joural of Computer Sciece Issues, Vol. 10, Issue 4, No 1, July 2013 ISSN (Prit): 1694-0814 ISSN (Olie): 1694-0784 www.ijcsi.org 14 3.2 Sesor Network Protocol Desig The problem is emphasized o how to desig a protocol to icrease etwork lifetime ad decrease eergy cosumptio i available odes. The bechmarks are tryig to use more from usual eergy of sesors. I coverig etworks, the physical positios of odes ad times of usig them should be cosidered i desiged protocol. The times of usig sesor ad also the distace betwee selected ode (i fact i relay path) ad super odes have crucial role for eergy cosumptio of that group. Therefore, we should seek for a relatio betwee these two parameters ad their eergy cosumptio. At first, we state the problem ad cosidered situatios. The, similar parameters that iclude timig algorithm based o the super odes (for poit coverage) will be explaied below. Our etwork cotais N sesors amed S 1 to have M super odes amed S u1 to um S N. We S (M<N). The proposed timig algorithm is divided ito time itervals with certai rouds ad idetical itervalst r. Selected group is oly active durig time of Tr ad other odes are off durig a roud. durig roud a, T r ca be computed by cosidered groupig time, the groups of eergy estimate physical parameters of lifetime ad types of ormal sesors is used i etwork. 3.3 domiatig Provisios etwork The provisios of etwork are listed below: There are K targets with defied positios i etwork composed of sesor odes ad super odes. I cosidered sceario, sesor odes ad super odes are radomly distributed. This pla of sesor odes activities must be guarateed accordig to followig coditios after ruig algorithm for etwork lifetime: TargetsTa 1toT ak must be covered. There are odes S 1 to S N which perform moitorig task ad are deployed radomly. The super odes S to S are deployed. u 1 A set of odes C 1 to C should be selected. Each C is set of active odes ad is geerated by protocol i each roud. Each set of C is ecessary ad sufficiet to cover k targets. I fact, the obective is to divide sesor odes ito active ad iactive groups. Active sesors must be able to u M commuicate ad cover. The obective is to use this algorithm for maximizig the groups, reducig eergy cosumptio ad icreasig etwork lifetime. I each executive roud, it should be checked whether a ode is active as a sesor ode or a relay ode. Each ormal sesor has iitial Ei ad high processig power. Commo sesors Dissimilar to super odes have higher eergy, greater lifetime ad higher processig power. All super odes are coected to each other by a path betwee two super odes. Each active sesor exists i oe of C groups ad coected to a super ode by relay odes. Each sesor is coected to oe of super odes through Data trasmissio path. Sesor odes possess iitial eergy E i, commuicatio rage R c, ad sesig rage R R R ). s ( c s This selectio must be located ad distributed. Decisio makig is doe for usig data i eighborig ode with fixed multi-hop distace. Defiitio 1: I defied poit coverage, it should be said that whe the Euclidea distace betwee odes ad target is Less tha or equal R, the target is covered. s Defiitio 2: Sesors ca coect to each other or super odes if the Euclidea distace is less tha R. Defiitio 3: Network lifetime is defied as time iterval i which all k targets will be covered by a set of active sesor odes that are coected to super odes. 3.4 Sesor Nodes Selectio Algorithm As idicated before, desiged groupig algorithm which are executed at the begiig of each performace roud, icludes two sectios. The first sectio is selected active odes. The secod sectio is attributed to data collectio from odes ad data trasmissio through relay odes. I the first sectio, oe of C groups is formed i a way that must be satisfied i above provisios. Whe this group is active, all other odes are iactive (Sleep Mode) ad cosume little eergy. They should be evaluated i ext phase. This evaluatio is doe by cosiderig a series of physical factors of sesors durig a roud. s Copyright (c) 2013 Iteratioal Joural of Computer Sciece Issues. All Rights Reserved.
IJCSI Iteratioal Joural of Computer Sciece Issues, Vol. 10, Issue 4, No 1, July 2013 ISSN (Prit): 1694-0814 ISSN (Olie): 1694-0784 www.ijcsi.org 15 3.5 System Specificatios The etwork is offered a squared eviromet. There are T targets i the eviromet that are covered with a k coectio of coverig etwork. Tars Icludes all targets i sesig domai S. They are covered by odes. The umber of targets is located i sesig rage of ode S 1 which is show by m 1. The iitial eergy of commo sesors is Ei ad iitial eergy of super odes is three times greater tha E. The eergy cosumed i each roud is called Es1 ad the cosumed eergy of a relay i each roud is called Ec 1. The first sectio iclude sesor ode selectio, checkig size fuctio for evaluatio ad selectig active moitorig odes that are w time uits (the Secod is the time uit here). The waitig time of ode S is computed by a fuctio measurig physical parameters of sesor S. Waitig time is stated as a multiple coefficiet for total time of a roud by usig the parameters of a ode: remaiig eergy, iitial eergy ad umber of targets see i the rage of a sesor. A sesor decides to sleep or awake after passig remaiig time. If E Es 1 Ec 1 ( E is remaiig eergy of sesor ode S ) the the ode caot be coverted to a sesor ode. So waitig time is ot computed ad t is waitig time of ode which is equivalet to w. It meas that the ode is ot a sesor. Otherwise, whe E Es 1 Ec 1 t is computed ad, ispected. Whe t is fiished Tars, S itroduces itself as a sesor ode ad ois active odes i the group. The, ew selected ode shows the positio of two-hop eighborig odes. If there is a ode such as S at the ed of the roud that Tars ad E E E, the s 1 c 1 ode seds the o coverage message to super ode. It meas that etwork lifetime is termiated. At this time, a message cotaiig o completed coverage is set to super odes ad the etwork seds this message to fial moitorig destiatio. 4. Gravitatioal force I 1995 for the first time vadoris ad Tsag [24] proposed GLS algorithm to search ad solve NP-complete problems i ad i 2004 Barry Webster [25] preseted it as a robust algorithm ad amed it GELS. The idea of this algorithm is based o gravitatioal force priciple that causes obects are attracted to each other i the ature, So that the heavier obect has more gravitatioal force ad imposes it o other obects ad attracts the obects with lower weight toward itself. However, the Distace of two obects is very effective o size of this force; cosider two obects with same weight ad differet Distace compared to obect with less weight, the Obect which has less distace to low weight obect ca impose more gravitatioal force o it. I GELS, Newto's law of gravity formula betwee two obects is: F= ) 4( Where m 1 ad m 2 are mass of first obect ad secod obect respectively. G Equals to gravitatioal costat value 6.672, R is the radius parameter ad the distace betwee two obects. Also GELS imitates this process of ature to search through a search space. So that search space, world ad obects i this world are possible resposes to search. Each obect has a weight, the weight of each obect is the performace or the search criteria, i which the best respose has maximum weight ad oe of obects, caot hold a zero weight [26-28]. I this way, the possible resposes i search space based o the criteria that depeds o type of problem are divided ito categories that each category is kow as a dimesio of problem respose ad a value called iitial velocity is cosidered for each dimesio of problem respose which will be explaied below. GELS iclude a vector whose size specifies the umber of respose dimesios. The values of this vector represet relative velocity i each dimesio. The algorithm starts with a iitial respose, iitial velocity vector ad movemet directio. For each dimesio i velocity vector, a radom umber betwee oe ad maximum speed is selected ad it is the value of each elemet i each dimesio. The iitial respose is geerated by user or radomly as curret respose. For each dimesio i iitial velocity vector, accordig to iitial velocity vector of respose dimesios, a directio is selected to move which is equals to respose dimesio that has maximum iitial velocity i iitial velocity vector. The algorithm cosists of a poiter obect that ca move i search space ad the weight cosidered for obect poiter is fixed i all calculatios ad the obect always refers to a respose with maximum weight. The algorithm is completed with occurrece of oe of two coditios: All Copyright (c) 2013 Iteratioal Joural of Computer Sciece Issues. All Rights Reserved.
IJCSI Iteratioal Joural of Computer Sciece Issues, Vol. 10, Issue 4, No 1, July 2013 ISSN (Prit): 1694-0814 ISSN (Olie): 1694-0784 www.ijcsi.org 16 compoets of iitial velocity vector are zero, or umber of algorithm iteratios reaches its maximum. I used Newto s formula, by replacig two mass i umerator of equatio ad replacig with differece betwee cost of cadidate respose ad curret respose, the gravitatioal force betwee two obects is calculated usig the followig equatio: G( CU CA) f 2 R (5). Where CU ad CA are cost of curret respose ad cadidate respose respectively.this formula has a positive value if cost of curret respose is greater tha cost of cadidate respose ad has a egative value if cost of cadidate respose is larger. The value of this force, positive or egative, is added to velocity vector i status of curret path. If this actio causes the value of velocity parameter exceeds the maximum settig, it takes a maximum value. If the update results i egative value, it takes zero. The available Parameters i GELS: Maximum Velocity: The maximum value that ca be allocated to each elemet of iitial velocity vector ad this parameter prevets from gettig too big. Radius: the Radius which is used i the formula to calculate the gravitatioal force. Iteratio: Defies the maximum umber of algorithm iteratios which Esures that the algorithm is termiated [26]. 5. The proposed algorithm I ew proposed method, gravitatioal emulatio local search algorithm (GELS) is used as a strategy to select optimal sesors. The choice is doe for moitorig i Poit coverage wireless sesor etwork. The goal of this algorithm is to icrease etwork lifetime by optimizatio ad reducig power cosumptio ad icreasig moitorig etwork efficiecy. At first For each executive rouds i etwork, a umber of sesors are activated to moitor, they Will lose some of their eergy For each activatio ad These sesors must be chose i a way To esure that these sesors will cover all poits eeded for moitorig ad Also there will be a distace from these odes to the sik which has a cost. To solve the problem of optimal sesor selectio for moitorig i Poit coverage wireless sesor etwork,at first we cosider three distace matrix, iitial velocity matrix ad time matrix which distace matrix, iitial velocity matrix are radomly produced. I velocity matrix, a iitial velocity will be give to each available sesor i the etwork which is cosidered as a mass ad the i later stages the speed will chage besides the time matrix is obtaied from followig equatio Based o distace matrix ad velocity matrix: ) 6( The After creatio three metioed matrices, Sesors will be placed radomly withi a array ad For each executive rouds i etwork, the sesors which cover limitatios of problem ad have greater mass ad speed ad shorter distace compared to targets of etwork ad also have less frequet tha other sesors ad they Have bee selected with this coditio that they perform Moitorig i etwork, will be activated. At this time, a solutio has created for problem ad the the Suitability of solutio will be calculated ad will be recogized as a mass of that solutio. Accordig to law of gravity, the best solutio must be the largest mass. Next solutio will be created from curret solutio based o problem limitatio that the suitability of this solutio is calculated ad will be kow as mass of solutio ad I the case of optimizatio to curret solutio, created Solutio is chose as curret solutio I.e. the problem will go toward optimizatio. If created solutio is ot optimal compared to curret solutio, The algorithm does ot cosider created solutio ad the makes aother solutio from curret solutio ad examies it Ad this actio is repeated util the algorithm retur the optimal solutio. The algorithm is completed with occurrece oe of two coditios: All compoets of iitial velocity vector are zero, or the umber of algorithm iteratios reaches its maximum. 6. Simulatio results To evaluate the performace of proposed algorithm, the software C# is used. Sesor etwork is simulated i various modes. Selectio methods to simulate ad compare iterface distace Selectio, iclude a method based o greedy algorithm for distace selectio, a method based o clusterig -the shortest distace Selectio ad the method proposed i this paper. Table 1 is used for above simulatios ad proposed algorithm is compared with sources [22] ad [23] ad [29] ad [30]. It should be oted that i proposed algorithm the ru time of each roud of proposed algorithm due to algorithm simplicity ad low slag Is very low ad very substatial Copyright (c) 2013 Iteratioal Joural of Computer Sciece Issues. All Rights Reserved.
IJCSI Iteratioal Joural of Computer Sciece Issues, Vol. 10, Issue 4, No 1, July 2013 ISSN (Prit): 1694-0814 ISSN (Olie): 1694-0784 www.ijcsi.org 17 compared to other algorithms that are simulated,i additio it reduces eergy cosumptio Ad lifetime that ca be see i the simulatio. Table 1: The values used i simulatio Parameter Network Size SNodes Locatio Nodes Locatio Nodes Iitial Eergy SuperNode Iitial Eergy Commuicatio Rage Sesig Rage Number of Nodes 300 Number of SNodes 25 Value 500 * 500 m Radom Radom 0.1 J 0.5 J 90 m 60 m Number of Target 20 Eelec 50 J/bit Fig. 5. Results of the gravity algorithm for 300 Normal sesor, 25 Maager Sesor ad 20 o target 7. Coclusios Fig. 3. Results of gravity algorithms for 200 Normal sesors, 25 Maager Sesor ad 20 targets Fig. 4. Results of the gravity algorithm for 300 Normal sesors, 15 Maager Sesor ad 30 targets I this paper, a gravitatioal emulatio local search algorithm is used to solve Optimal sesor selectio for moitorig i poit coverage wireless sesor etwork.as Well as a New Method is proposed To Calculatio suitability Ad evaluate preseted Solutios To solve Optimal sesor selectio problem for moitorig i poit coverage wireless sesor etwork. The advatages of this algorithm are speed, low ru time, icrease lifetime of etwork by optimizatio ad reduce eergy cosumptio ad icrease moitorig etwork efficiecy. The results shows improvemet ad superiority of proposed algorithms compared to sources [22] ad [23] ad [29] ad [30]. This improvemet is more apparet i large-scale systems. Refereces [1] E.H. Callaway, Wireless Sesor Networks, Architectures ad Protocols, Aurebach Publicatios, vol 1, pp. 20-28,2004. [2] M. Cardei, D. MacCallum, X. Cheg, M. Mi, X. Jia, D. Li, ad D.-Z. Du, Wireless Sesor Networks with Eergy Efficiet Orgaizatio, Joural of Itercoectio Networks, Vol 3, No 3-4, pp 213-229, Dec 2002. [3] G. Gupta ad M. Youis, Load-Balaced Clusterig i Wireless Sesor Networks, IEEE Iteratioal coferece o commuicatios, Achorage, Alaska, 2003. [4] M. Cardei M. T. Thai, Y. Li, W. Wu, Eergy-Efficiet Target Coverage i Wireless Sesor Networks,,IEEE INFOCOM, 2005. [5] J. Carle, D. Simplot, Eergy efficiet area moitorig by sesor etworks, IEEE Computer 37 (2) pp 40 46, 2004. [6] S. Sliepcevic ad M. Potkoak, Power efficiet orgaizatio of wireless sesor etworks, Proc. IEEE Copyright (c) 2013 Iteratioal Joural of Computer Sciece Issues. All Rights Reserved.
IJCSI Iteratioal Joural of Computer Sciece Issues, Vol. 10, Issue 4, No 1, July 2013 ISSN (Prit): 1694-0814 ISSN (Olie): 1694-0784 www.ijcsi.org 18 Iteratioal Coferece o Commuicatios, pp. 472 476, 2001. [7] D. Tia ad N. D. Georgaas, A ode schedulig scheme for eergy coservatio i large wireless sesor etworks, Wireless Comm. Mob. Comput., vol. 3, pp. 271 290, 2003. [8] B. Carbuar, A. Grama ad J. Vitek. Distributed ad Dyamic Vorooi Overlays for Coverage Detectio ad Distributed Hash Tables i Ad-hoc Networks, (ICPADS 2004), Newport Beach, CA, pp. 549-559, July 2004. [9] C. Che & J. Ma, Desigig Eergy-Efficiet Wireless Sesor Networks with. Mobile Siks, ACM Iteratioal Workshop o (WSNA), pp. 343 349, 2006. [10] J. Lu, J. Wag, ad T. Suda, Scalable Coverage Maiteace for Dese Wireless Sesor Networks, EURASIP Joural o Wireless Commuicatios ad Networkig Volume, 2007. [11] S. Commuri, M.Watfa, Coverage Strategies i 3D Wireless Sesor Networks, Iteratioal Joural of Distributed Sesor Networks, pp. :333-353,October 2006. [12] M. Watfa, S. Commuri, Boudary Coverage & Coverage Boudary Problems i Wireless Sesor Networks, Iteratioal Joural of Sesor Networks, IJSNET, Volume: 2, Number: 3/4, Pages:273-283, 2007. [13] Y. Xu ad X. Yao, "A GA Approach to the Optimal Placemet of Sesors i Wireless Sesor Networks with Obstacles ad Prefereces, IEEE CCNC, 2006. [14] A. S. Rostami, M. R. Tahatalab, H. M. Bereti, "Novel Algorithm of Eergy-Aware i Asymmetric Wireless Sesor Networks Routig for I-Poit Coverage", CICN, pp 308-313,IEEE 2010. [15] J. Kim, D. Kim, Eergy-efficiet Data Gatherig Techiques Usig Multiple Paths for Providig Resiliece to Node Failures i Wireless Sesor Networks,, JOURNAL OF COMMUNICATIONS, VOL. 1, NO. 3, JUNE 2006. [16] B. Aou, R. Boutaba, Clusterig i WSN with Latecy ad Eergy Cosumptio Costraits,, Joural of Network ad Systems Maagemet, Vol. 14, No. 3, September 2006. [17] A. S. Rostami, M. R. Tahatalab, H. M. Bereti, "Decreasig the Eergy Cosumptio by a Distributio Classificatio New Algorithm i Choosig the Best Sesor Node i Wireless Sesor Network with Poit Coverage", CICN, pp 269-274, IEEE 2010. [18] S. J. Baek, G. d. Veciaa, X. Su, Miimizig Eergy Cosumptio i Large-Scale Sesor Networks Through Distributed Data Compressio ad Hierarchical Aggregatio,, IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS, VOL. 22, NO. 6, AUGUST 2004. [19] Y. T. Hou, Y. Sh. Haif, D. Sherali, S. F. Midkiff, O Eergy Provisioig ad Relay Node Placemet for Wireless Sesor Networks,,IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS, VOL. 4, NO. 5, 2005. [20] S. Badyopadhyay, E. J. Coyle, A Eergy Efficiet Hierarchical Clusterig Algorithm for Wireless Sesor Networks,, IEEE, 2003. [21] M. Cardei, M. T. Thai, Y. Li, Eergy -Efficiet Target Coverage i Wireless Sesor Networks, IEEE, 2005. [22] W.Awada ad M. Cardei, Eergy Efficiet data Gaterig i heterogeeous Wireless Sesor etworks,, IEEE, WiMob,2006. [23] A. S. Rostami, H. M.Berety ad A. R. Hosseiabadi, A Novel ad Optimized Algorithm to Select Moitorig Sesors by GSA, ICCIA, 829 834, 2011. [24] C. Voudouris ad E. Tsag, Guided Local Search, Departmet of Computer Sciece, Uiversity of Essex, UK, August 1995. [25] B. Webster, Solvig Combiatorial Optimizatio Problems Usig a New Algorithm Based o Gravitatioal Attractio,Ph.D., thesis, Melboure, Florida Istitute of Techology, 2004. [26] A. R. Hosseiabadi, A. B. Farahabadi, M. S. Rostami, A. F. Latera, Presetatio of a New ad Beeficial Method Through Problem Solvig Timig of Ope Shop by Radom Algorithm Gravitatioal Emulatio Local Search, Iteratioal Joural of Computer Sciece Issues, Vol. 10, Issue 1,745-752, 2013. [27] A. R. Hosseiabadi, M. Yazdapaah ad A. S. Rostami, A New Search Algorithm for Solvig Symmetric Travelig Salesma Problem Based o Gravity, World Applied Scieces Joural 16 (10): 1387-1392, 2012. [28] A. R. Hosseiabadi, M. R. Ghaleh, S. E. Hashemi, Applicatio of Modified Gravitatioal Search Algorithm to Solve the Problem of Teachig Hidde Markov Model, Iteratioal Joural of Computer Sciece Issues, Vol. 10, Issue 3, 2013, 1-8. [29] Zh. Liu, Maximizig Network Lifetime for Target Coverage Problem i Heterogeeous Wireless Sesor Networks, MSN 2007, LNCS 4864, pp. 457 468, 2007. [30] Sh. Rostami, M.H. Tahatalab, H. M. Berety, S. E. Naghibi, Decreasig the Eergy Cosumptio by a New Algorithm i Choosig the Best Sesor Node i Wireless Sesor Network with Poit Coverage, IEEE CICN, 2010. Copyright (c) 2013 Iteratioal Joural of Computer Sciece Issues. All Rights Reserved.