Académique Documents
Professionnel Documents
Culture Documents
Abstracto
Durante la ltima dcada, la resolucin de problemas complejos de
optimizacin con algoritmos metaheursticos ha recibido considerable atencin
entre los profesionales e investigadores. Por lo tanto, muchos algoritmos
metaheursticos se han desarrollado en los ltimos aos. Muchos de estos
algoritmos estn inspirados en diversos fenmenos de la naturaleza. En este
trabajo, se introduce un nuevo algoritmo basado en la poblacin, el Len
optimizacin del algoritmo (LOA),. estilo de vida propio de los leones y sus
caractersticas de cooperacin ha sido la motivacin bsica para el desarrollo
de este algoritmo de optimizacin. Algunos problemas de referencia se
seleccionan de la literatura, y la solucin del algoritmo propuesto se han
comparado con los de algunos bien conocido y ms nuevos meta-heurstica
para estos problemas. Los resultados obtenidos confirman el alto rendimiento
del algoritmo propuesto en comparacin con los otros algoritmos utilizados en
este trabajo. Y 2015 de la Sociedad de CAD / CAM Ingenieros. Produccin y
hospedaje por Elsevier. Todos los derechos reservados. Este es un artculo de
acceso
abierto
bajo
la
licencia
CC
BY-NC-ND
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
2.1. leones inspiracin son las ms inclinaciones sociales de todas las especies
salvajes que muestran altos niveles de cooperacin y el antagonismo [55]. Los
leones son de particular inters debido a su fuerte dimorfismo sexual tanto en
el comportamiento social y la apariencia. El len es un felino salvaje con dos
tipos de organizacin social: residentes y nmadas. Los residentes viven en
grupos, llamados orgullo [56]. Una manada de leones incluye tpicamente
alrededor de cinco hembras, sus cachorros de ambos sexos, y uno o ms de los
varones adultos. Los machos jvenes son excluidos de su orgullo el nacimiento,
cuando alcanzan la madurez sexual [56]. Como se menciona anteriormente, el
segundo comportamiento organizacional se llama nmadas, que se mueven de
forma espordica, en parejas o en singularmente. Las parejas se ven ms en
los hombres relacionados que han sido excluidos de su orgullo maternal.
Observe que un len puede cambiar los estilos de vida; residentes pueden
llegar a ser nmadas y viceversa [56]. A diferencia de todos los otros gatos,
leones suelen cazar juntos con otros miembros de su orgullo. Varias leonas
trabajan juntos y dan la vuelta al presas desde diferentes puntos y atrapar a la
vctima con un ataque rpido.
caza grupo coordinado aporta una mayor probabilidad de xito en la caza de
len. Los leones machos y algunas leonas suelen quedarse y descansar a la
espera de las leonas cazadores para volver de la caza [57]. Los leones se
aparean en cualquier poca del ao, y las hembras son poliestro (cuando las
hembras no cran de sus cachorros son receptivos) [58]. Una leona puede
aparearse con diferentes parejas cuando ella est en celo [59]. En la
naturaleza, los leones machos y hembras marcan su territorio y en otras
partes, que parece un buen lugar con la orina.
Para cada len, la mejor solucin obtenida en el pasado que se llama posicin
opera- mejor visitado, y durante el proceso de optimizacin se actualiza
progresivamente. En carta de acuerdo, un territorio es un orgullo son una que
consiste en que cada miembro de mejor posicin visitado.
En cada orgullo, algunas hembras que se seleccionan al azar van de cacera.
Los cazadores se mueven a las salas de la presa para rodear y atraparlo. Hay
tela de las hembras se mueven para evitar diferentes posiciones del territorio.
Los leones machos en el orgullo, deambulan en el territorio. Las hembras se
aparean en manadas con uno o algunos machos residentes. En cada orgullo,
los hombres jvenes son excluidos de su orgullo maternal y convertirse en
nmada cuando alcanzan la madurez y, su potencia es menor que los machos
residentes. Tambin, un len del nmada (tanto hombres como mujeres) se
mueve al azar en el espacio de bsqueda para encontrar un lugar mejor
(solucin). Si el macho fuerte nmada invaden el macho residente, hay macho
dente es expulsado de la manada por el len nmada. El macho se vuelve
nmada len residente. En la evolucin, algunas hembras residentes emigran
de un orgullo a otro o cambian sus estilos de vida y se convierten en nmadas
y viceversa algunos leones nmada femenino se unen las manadas. Debido a
muchos factores, como la falta de alimentos y la competencia, ms dbil len
va a morir o ser matado. Por encima de proceso contina hasta que se
satisface la condicin de parada.
2.2. algoritmo propuesto
2.2.2. En cada caza orgullo cierta mirada femenina para una presa en un grupo
para proporcionar alimentos para su orgullo. Estos cazadores tienen estrategias
especficas para rodear la presa y atraparlo. En general, los leones siguieron
aproximadamente los mismos patrones cuando la caza [60]. Stander [60]
divide los leones en siete funciones diferentes acecho, que se muestra en la
Fig. 1, la agrupacin de estos papeles a Extremo izquierdo, central y derecho
ala posiciones. Durante la caza, cada una leona corrige su posicin en funcin
de su propia posicin y las posiciones de los miembros del grupo. Debido a este
hecho de que durante la caza de algunos de estos cazadores de presas rodear
y atacar desde la posicin opuesta, utilizamos oposicin - Aprendizaje basado
en (OBL). El concepto bsico de oposicin con base en el aprendizaje (OBL) fue
propuesto por Tizhoosh [61] y se ha demostrado ser un mtodo eficaz para
resolver problemas de optimizacin.
Un alto nmero de xitos indica que los leones han convergido a un punto que
est lejos de ser el punto ptimo. Del mismo modo, un nmero bajo de xito
demuestra que los leones estn haciendo pivotar alrededor de la solucin
ptima sin mejora significativa. As que este factor se puede utilizar como una
serie de elementos tiles para el tamao de un torneo.
by beha-vior of ants foraging for food. Particle Swarm Optimization [25] mimics the
social behavior of a flock of migrating birds trying to reach an unknown destination.
Marriage in Honey Bee Optimization Algorithm (MBO) was proposed by Abbass [26],
and mimics processes of reproduction in the honey bee colony. Bacterial Foraging
Algorithm [27] simulates search and optimal foraging of bacteria. The Shuffled Frog
Leaping algorithm [28] was inspired by a frog population searching for food. The Cat
Swarm algorithm [29] was developed based on the behavior of cats. Invasive weed
optimization was proposed by Mehrabian and Lucas [30], and mimics the ecological
behavior of colonizing weeds. Monkey Search [31] simulates a monkey in search for
food resources. Water flow-like algo-rithm [32] was inspired by water flowing from
higher to lower levels. Biogeography-based optimization algorithm was intro-duced by
Simon [33], and inspired by biogeography which refers to the study of biological
organisms in terms of geographical distribution (over time and space). The Fish
School Search [34] was proposedbasedonthegregarious behaviorofoceanic fish.
CuckooSearch [35] and Cuckoo optimizationalgorithm [36] are
basedonreproductionstrategy of cuckoos.Bat-inspiredAlgorithm [37] was inspiredbythe
echolocationbehaviorofbats.Firefly algorithm [38] simulates the socialbehaviorof
fireflies basedontheir flashingchar- acteristics.DolphinPartnerOptimization [39] and
Dolphin echolocationalgorithm [40] wereinspiredbydolphins'beha- viors.
Flowerpollinationalgorithm [41] mimicsthepollination characteristicsof
floweringplantsandtheassociated flower consistencyofsomepollinatinginsects.Krillherd
[42] inspired by the herdingbehaviorofkrillindividuals.Wolf search [43] and
GreyWolfOptimizer [44] are inspiredby behaviorsofwolves.Watercyclealgorithm [45] was
basedon the observationofwatercycleprocessandhowriversand streams flow
totheseaintherealworld.TheSocialspider
optimization,inspiredbythesocialbehaviorofakindof spider,hasbeenproposedrecently
[46]. Forest Optimization Algorithm [47] was inspiredbyfewtreesintheforestswhich can
survive for many years,whileothertreescouldlivefora short time.
Aforementionedalgorithmsarewidelyappliedbyresearch- ers inmanydifferentareas [48
51]. However,thereisno particularalgorithmtogainthemostappropriatesolutionfor all
optimizationproblems.Somealgorithmsprovidebetter solution
forsomeparticularproblemscomparedwithothers.
Therefore,pursuingfornewoptimizationtechniquesisan open problem [52]. In
thispaper,anoptimizationalgorithmbasedonlion's
behaviorandsocialorganization,namelyLionOptimization
Algorithm(LOA)isproposed.Intheliterature,Wang [53] and Rajakumar [54]
proposedtwoalgorithmsinspiredbyfew charactersoflions.Rajakumar [54]
describedthemain operatorofLion'sAlgorithmas Mating thatreferstoderiving new
solutionsandTerritorialDefenseandTerritorialTakeover intend to find
andreplacetheworstsolutionbynewthebest solution.
LikeLion'sAlgorithm,Lionprideoptimizer [53] is based on fighting
betweenindividualandmating.Butlionsin addition ofmatingand fighting
exhibitotherbehaviorssuchas specialstyleofpreycapturing,territorialmarking,migration,
difference betweenlifestyleofnomadandresidentlions.So,
proposedalgorithmisinspiredbysimulationofthesolitary and
cooperativebehaviorsoflionswhicharecompletely different fromthepreviousalgorithm.
After thisintroduction,theremainderofthispaperis structuredasfollows:In Section2 the
proposedLion OptimizationAlgorithm(LOA)isoutlined,anditsimplemen- tation
stepsareexplainedindetails.Comparativestudyand experimentalresultsarepresentedin
Section 3 to verifythe efficiency oftheproposedalgorithm.Finally,conclusionsare
presentedinthelastsection. 2. LionOptimizationAlgorithm(LOA) In
thissection,theinspirationoftheproposedmeta-heuristic is first
discussed.Then,LionOptimizationAlgorithm(LOA)is presented. 2.1. Inspiration
Lionsarethemostsociallyinclinedofallwildcatspecies which
displayhighlevelsofcooperationandantagonism [55]. Lions
areofparticularinterestbecauseoftheirstrongsexual
dimorphisminbothsocialbehaviorandappearance.Thelion is
awildfelidwithtwotypesofsocialorganization:residents and
nomads.Residentslivesingroups,calledpride [56]. A pride oflionstypicallyincludesabout
five females,theircubs of bothsexes,andoneormorethanoneadultmales.Young males
areexcludedfromtheirbirthpridewhentheybecome sexually mature [56].
Asmentionbefore,thesecondorganiza- tional
behavioriscallednomads,whomoveaboutsporadi- cally,
eitherinpairsorsingularly.Pairsaremoreseenamong related
maleswhohavebeenexcludedfromtheirmaternal pride.
Noticethatalionmayswitchlifestyles;residentsmay becomenomadsandviceversa [56].
Unlikeallothercats,Lionstypicallyhunttogetherwith other
membersoftheirpride.Severallionessesworktogether and
encirclethepreyfromdifferentpointsandcatchthevictim with
afastattack.Coordinatedgrouphuntingbringsagreater
probabilityofsuccessinlionhunts.Themalelionsandsome lionesses
usuallystayandrestwhilewaitingforthehunter lionesses toreturnfromthehunt [57].
Lionsdomateatany time oftheyear,andthefemalesarepolyestrous(when females
notrearingtheircubsarereceptive) [58]. Alioness may
matewithmultiplepartnerswhensheisinheat [59]. In nature,
maleandfemalelionsmarktheirterritoryandelse- where,
whichseemsagoodplacewithurine. In thiswork,somecharactersoflionsaremathematically
modeled inordertodesignanoptimizationalgorithm.Inthe
proposedalgorithm,LionOptimizationAlgorithm(LOA),an initial
populationisformedbyasetofrandomlygenerated
solutionscalledLions.Someofthelionsintheinitial population(%N)
areselectedasnomadlionsandrestpopula- tion (residentlions)israndomlypartitionedinto P
subsets called prides. S percent ofthepride'smembersareconsidered as
femaleandrestareconsideredasmale,whilethisrate(sex rate %S)
innomadlionsisviceversa. Foreachlion,thebestobtainedsolutioninpassediterations is
calledbestvisitedposition,andduringtheoptimization
processisupdatedprogressively.InLOA,aprideterritoryis an
areathatconsistsofeachmemberbestvisitedposition.In each
pride,somefemaleswhichareselectedrandomlygo hunting.
Huntersmovetowardsthepreytoencircleandcatch it.
Therestofthefemalesmovetowarddifferentpositionsof territory.
Malelionsinpride,roaminterritory.Femalesin prides
matewithoneorsomeresidentmales.Ineachpride, young
malesareexcludedfromtheirmaternalprideand
becomenomadwhentheyreachmaturityand,theirpoweris less thanresidentmales.
Also,anomadlion(bothmaleandfemale)movesrandomly in thesearchspaceto find
abetterplace(solution).Ifthe strong nomadmaleinvadetheresidentmale,theresidentmale
is drivenoutofthepridebythenomadlion.Thenomadmale
becomestheresidentlion.Intheevolution,someresident females
immigratefromonepridetoanotherorswitchtheir
lifestyles andbecomenomadandviceversasomenomad
femalelionsjoinprides.Duetomanyfactorssuchaslackof food
andcompetition,weakestlionwilldieorbekilled.Above
processcontinuesuntilthestoppingconditionissatisfied.
2.2. Proposedalgorithm
2.2.1. Initialization The LOAisapopulation-basedmeta-heuristicalgorithmin which the
first stepistorandomlygeneratethepopulationover the
solutionspace.Inthisalgorithm,everysinglesolutionis called Lion. Ina Nvar
dimensionaloptimizationproblem,a Lion isrepresentedasfollows: Lion x1; x2; x3;;
xNvar _ _ 1 Cost (fitness value)ofeachLioniscomputedbyevaluating the costfunction,as:
fitness valueoflion f Lion f x1; x2; x3;; xNvar _ _ 2 In first step, Npop
solutionsaregeneratedrandomlyinsearch space. %N of
generatedsolutionsarerandomlychosenas nomad
lions.Therestofthepopulationwillberandomly divided into P
prides.Everysolutioninthisalgorithmhasa specific
genderandremainedconstantduringtheoptimization
process.Toemulatingthisfact,ineachpride%S (%75%90) of
entirepopulationformedinthelaststepareknownas
femalesandtherestasmales.Fornomadlions,thisratiois vice versa % (1_S).
Overthesearchingprocesseverylion
marksitsbestvisitedposition.Accordingtothesemarked
positions,everypride'sterritoryisformed.So,foreachpride,
markedpositions(bestvisitedpositions)byitsmembersform that pride'sterritory. 2.2.2.
Hunting In eachpridesomefemalelookforapreyinagroupto
providefoodfortheirpride.Thesehuntershavespecific
strategiestoencirclethepreyandcatchit.Ingeneral,lions
followedapproximatelythesamepatternswhenhunting [60]. Stander [60] divided
thelionsintosevendifferentstalking roles, shownin Fig. 1,
groupingtheserolesintoLeftWing, Centre
andRightWingpositions.Duringhunting,eachlioness corrects
itspositionbasedonitsownpositionandthepositions of membersofthegroup. Due
tothisfactthatduringhuntingsomeofthesehunters
encirclepreyandattackfromoppositeposition,weutilize OppositionBasedLearning(OBL).Thebasicconceptof OppositionBasedLearning(OBL)wasproposedbyTizhoosh [61] and
hasbeenproventobeaneffectivemethodforsolving optimizationproblems.
The principleofOpposition-BasedLearning (OBL) isgivenin Fig. 2.
Accordingtoaforementionedfacts,huntersaredividedinto three
subgroupsrandomly.Groupwithhighestcumulative members'
finessesisconsideredasCenterandtheothertwo groups
considerastwowings.Adummyprey(PREY) is considered incenterofhunters(PREY
Phuntersx1; x2; x3;; xNvar =number ofhunters).Duringhunting,hunters are
selectedoneafteranotherrandomly,andeachselected hunter
attackondummypreywhichthisprocedurewillbe defined
lateraccordingtogroupthatselectedlionisbelongto that.
Throughouthunting,ifahunterimprovesitsown finesses, PREY will
escapefromhunterandnewpositionof PREY is obtainedasfollows: PREY0 PREYrand 0;
1_PI _ PREY_Hunter 3 where PREYiscurrentpositionofprey,Hunterisnew position
hunterwhoattacktopreyandPIisthepercentage of improvementin fitness ofhunter(see
2.2.3. Moving Toward Safe Place As mentioned in last subsection, in each pride some
females go hunting. Remained females go toward one of the areas of territory. Since
territory of each prides consist of personal best so far positions of each member, and
assists Lion Optimization Algorithm (LOA) to save the best solutions obtained so far
over the course of iteration, it can be used as valuable and reliable information to
improve solutions in LOA. Therefore, the new position for a female lion may be given
as: Female Lion0 Female Lion 2D
where Female Lion is current position of female lion, D shows the distance between the
female lion's position and the selected point chosen by tournament selection among
the pride's territory. R1fg is a vector which its start point is the previous location of the
female lion, and its direction is toward the selected position.{R2} is perpendicular to
{R1}. Now we describe our tournament section strategy. First, we define the success of
a lion if it improves his or her best position at last iteration of the LOA.
A high number of successes indicate that the lions have converged to a point that is far
from the optimum point. Similarly, a low number of success shows that the lions are
swinging around the optimum solution without significant improvement. So this factor
can be used as a useful elements for size of a tournament.
2.2.4. Roaming Each male lion in a pride roams in that pride's territory due to some
reasons. To emulate this behavior of resident males, % R of pride territory are selected
randomly and are visited by that lion. Along roaming, if resident male visits a new
position which is better than its current best position, update his best visited solution.
This roaming is a strong local search and assists Lion Optimization Algorithm (LOA) to
search around of a solution to improve it. This progress is shown in Fig. 6. As shown in
Fig. 6, lion moves toward the selected area of territory by x units, wherein x is a
random number with uniform distribution.
2.2.4. Roaming Each male lion in a pride roams in that pride's territory due to some
reasons. To emulate this behavior of resident males, % R of pride territory are selected
randomly and are visited by that lion. Along roaming, if resident male visits a new
position which is better than its current best position, update his best visited solution.
This roaming is a strong local search and assists Lion Optimization Algorithm (LOA) to
search around of a solution to improve it. This progress is shown in Fig. 6. As shown in
Fig. 6, lion moves toward the selected area of territory by x units, wherein x is a
random number with uniform distribution.
where d shows the distance between the male lion's position and the selected area of
territory. The vector from the male lion's position to the selected area of territory shows
the original direction of movement. To provide a chance for searching a wider area
around current solution and adding intensification property to the method and to
search for a wider area around current solution, the angle is added to this direction. It
seems an angle which selected by uniform distribution among _/6 (rad) and /6 (rad)
is adequate for this goal. In the Pseudo-code 3, the behavior of the male lion is shown:
2.2.5. Mating Mating is an essential process that assures the lions' survival, as well as
providing an opportunity for information exchange among members. In every pride,
%Ma of female lions mate with one or several resident males. These males are selected
randomly from the same pride as the female to produce offspring. For nomad lions it's
different in that a nomad female only mates with one of the males which are selected
randomly.