Atpresent, theideaofwireless sensor network (WSN) hasatMcted agreat dealOfresearch attention. A WSN coinsists ofalarge number oftiny sensing nodes deployed in a target area.Eachnodehasprocessing anvd xireless commllumication capabilities. Duetothelimited capabilities andthevulnerable tureofanindividual sensor, the sensors inWSN aredeployed inhiLgh density (high npto 20nodes/m3 [1]). Thesensor distribution hasreceived considerable attention recently whentheenvirorment is unknown orhostile, whichcamotbeperformed manually To scatter sensors byaircraft isa possible solution. However, bythis means, theactual landing position cannot becontrolled duetotheexistence ofwindandobstacles. Consequently, thecoverage may be inerior tothe application requirements nomatter howmanysensors are dropped. Insuchcases, itisnecessary tomakeuseof mobile sensors, whichcanmovetotheapproprate places to provide therequired coverage Inthis paper, wepropose aself-deployment algorithm in sensornetworks. Ourgoalistonmaximize thecoverage usingarelative small quantity ofsensornodes inagiven target area. Theaboveproblem canbedealt withan evolutionary selection algorithm. Mainguseofthe advantage of genetic algoritm (GA) in solving multi-objective optimization problems (MOPs)and combining itwiththemobility ofthesensor nodes, wecan getthebest solution. Therest ofthepaper isorgaized asfollows. InSection II we review therelated workintheliterature. Section III gives somedefinitions ofthecoverage problem inWSN.In Section IV,wepresent theproposed optimal distribution algorithm. Following that, wepresent oursimulation results inSection Vandconclude thepaper inSection VI. II.RELATED WORK Theconcept ofcoverage asaparadigm forevaluating many-robot systems wasintroduced byGage[2]. Gage defines three basic typesofcoverage thatareblanket coverage, barrier coverage andsweepcoverage. The objective ofblanket coverage istoachivea static arrangement ofnodesthatmaximizes thetotal detection area. Theobjective ofbarrier coverage istominimize the probability ofnmndetected penetration tirough thebarrier. Andsweepcoverage ismore-or-less equivalent toamoving barrier. According tothistaxonomy, thedistribution problem described inthispaperisa blanket coverage problem. In[3], a probing-based density control algorithm is proposed toensure long-lived. robust sensing coverage by leveraging unconstrained network scale. Inthis algorithml onlyasubset ofnodes aremaintained inwoiking modeto ensure desired sensing coverage, andother rednndant nodes areallowed tofall asleep forextending system lifetime. The working nodes continue working -until they n ontoftheir energy oraredestroyed. Andthesleeping nodewakesnp occasionally toprobe itslocal neighborhood andstarts working whenthere isnoworkingL nodewithin itsprobing range.