An Adaptive Virtual Machine Location Selection Mechanism in Distributed Cloud

Size: px
Start display at page:

Download "An Adaptive Virtual Machine Location Selection Mechanism in Distributed Cloud"

Transcription

1 KSII TRANSACTIONS ON INTERNET AND INFORMATION SYSTEMS VOL. 9, NO. 12, Dec Copyrght c2015 KSII An Adaptve Vrtual Machne Locaton Selecton Mechansm n Dstrbuted Cloud Shukun Lu 1, Wea Ja 2 1 School of Informaton Scence and Engneerng, Central South Unversty, Changsha, Chna [e-mal: lu_shukun@csu.edu.cn] 2 Department of Computer Scence and Engneerng, Shangha Jao Tong Unversty, Shangha, Chna [e-mal: wea@gmal.com] *Correspondng author: Shukun Lu Receved July 7, 2015; revsed September 22, 2015; accepted October 18, 2015; publshed December 31, 2015 Abstract The locaton selecton of vrtual machnes n dstrbuted cloud s dffcult because of the physcal resource dstrbuton, allocaton of mult-dmensonal resources, and resource unt cost. In ths study, we propose a mult-obect vrtual machne locaton selecton algorthm (MOVMLSA) based on group nformaton, doubly lnked lst structure and genetc algorthm. On the bass of the collaboraton of mult-dmensonal resources, a ftness functon s desgned usng fuzzy logc control parameters, whch can be used to optmze search space solutons. In the locaton selecton process, an orderly nformaton code based on group and resource nformaton can be generated by adoptng the memory mechansm of bologcal mmune systems. Ths approach, along wth the domnant elte strategy, enables the updatng of the populaton. The tournament selecton method s used to optmze the operator mechansms of the sngle-pont crossover and X-pont mutaton durng the populaton selecton. Such a method can be used to obtan an optmal soluton for the rapd locaton selecton of vrtual machnes. Expermental results show that the proposed algorthm s effectve n reducng the number of used physcal machnes and n mprovng the resource utlzaton of physcal machnes. The algorthm mproves the utlzaton degree of mult-dmensonal resource synergy and reduces the comprehensve unt cost of resources. Keywords: Dstrbuted cloud, locaton selecton, mult-dmenson, mmune memory Ths work s supported n part by the Natonal Natural Scence Foundaton of Chna under Grant Numbers and , the Internatonal Scence & Technology Cooperaton Program of Chna under Grant Number 2013DFB10070, the Chna Hunan Provncal Scence & Technology Program under Grant Number 2012GK4106, and the "Moble Health" Mnstry of Educaton - Chna Moble Jont Laboratory (MOE-DST No. [2012]311). We express our thanks to the edtor and all the revewers for our manuscrpt. ISSN :

2 KSII TRANSACTIONS ON INTERNET AND INFORMATION SYSTEMS VOL. 9, NO. 12, December Introducton Wth the growth of network applcaton servces, nformaton technology archtecture and a varety of resources should be effectvely ntegrated to manage physcal resources effectvely, mprove the utlzaton rate of resources, and reduce resource unt costs [1]. Vrtual machne technology s a key to vrtualzaton, and t s wdely appled to dstrbuted cloud [2]. Wth the rapd popularty of cloud computng, the unlmted use of lmted resources can be acheved by users n the future. Under ths condton, users can obtan the physcal resources they actually reure, smlar to how people purchase fuel or natural gas for ther daly routnes. However, users need to select the most approprate nterface to obtan such resources; otherwse, a consderable amount of resources wll be wasted. The selecton of the proper purchase wndow has thus become a key and fundamental problem that necesstates urgent solutons. Nowadays, the resources needed by cloud users are manly emboded n the form of a vrtual machne. To make a vrtual machne perform effcently, the host resource s mapped to the applcaton layer, and the resource schedulng process s encapsulated n the search process of the vrtual machne [3, 4]. Thus, the key problem n the resource allocaton process s the rapd and proper selecton of a vrtual machne for the correspondng physcal nodes under the premse of satsfyng all servce-level targets for dfferent applcatons [5]. Physcal resource utlzaton and user satsfacton can be greatly mproved wth an adaptve selecton mechansm that allows users to automatcally select a vrtual machne accordng to comprehensve factors, whch depend on a physcal machne, durng the allocaton of vrtual machne resources. Therefore, as the key ssue n vrtual machne deployment, the locaton selecton of vrtual machnes needs to be solved urgently. In ths study, we frst formulate our optmzaton problem as a bn packng based on mult-dmensonal resource utlty to determne the optmal locaton selecton of vrtual machnes to physcal machnes, wth consderaton of the reurements for dependablty. Ths method s dfferent from the tradtonal way. Second, ths study proposes a new populaton updatng method based on mmune memory and on a new mult-obect genetc algorthm for the locaton selecton of vrtual machnes based on the structure of a doubly lnked lst. Fnally, durng the codng process of vrtual and physcal machnes, we used a mechansm of nformaton groupng mode, through whch the smlar scale vrtual machnes can be allocated to the proper locatons n shorter tme compared wth the tradtonal way. The rest of ths paper s organzed as follows. Secton 2 surveys related works. Secton 3 states the problem that we wll address and presents related defntons. Secton 4 descrbes the detal of gene encodng and evaluaton functon desgn. Secton 5 descrbes the MOVMLSA algorthm. Secton 6 presents the expermental evaluaton and analyss. Fnally, Secton 7 concludes ths paper and suggests a future research drecton. 2. Related Work The mappng problem of vrtual machnes to physcal nodes can be regarded as a mult-dmensonal vector packng problem [6 8] and as a NP-hard problem. At present, ths problem s manly solved wth a heurstc algorthm [8, 9]. However, the current research on the locaton selecton of vrtual machnes n a cloud platform s manly amed at the

3 4778 Lu et al.: An Adaptve Vrtual Machne Locaton Selecton Mechansm n Dstrbuted Cloud optmzaton of only one partcular dmenson. For example, exstng research only ams to guarantee servce-level obectves, mnmze the number of physcal nodes, and reduce vrtual machne mgraton and energy consumpton [10]. The optmzaton goals n some cases are nconsstent and contradctng. For example, a vrtual machne s placed on a mnmal number of physcal nodes to reduce the number of used nodes; n ths way, dle nodes can be saved, and energy consumpton and management cost can be reduced [11]. Nevertheless, extensve vrtual machne mgraton occurs. If the goal s to reduce such mgraton, the number of used physcal nodes s lkely to ncrease. In [12, 13], a genetc algorthm was used to tackle the problem of the statc placement of vrtual machnes wthout consderaton of the overhead costs nvolved n the vrtual machne mgraton. In [14], node vrtualzaton ntegraton was descrbed as a packng and random optmzaton problem n the cloud data center, but the consderatons focused only on processor resources and not on other dmensons, such as memory and nput/output (I/O). In [15, 16], the authors proposed a management framework for vrtual machne placement n cloud computng, but they faled to consder resource costs and system energy consumpton. In [17, 18], the authors proposed a schedulng strategy for vrtual machnes based on a genetc algorthm, the hstorcal data of the cloud computng system, and the current state of the system [19]. Ths method acheves an deal load balance and ncurs mnmal overhead for vrtual machne mgraton; however, t gnores resource utlzaton and energy consumpton n the data center. In [13, 15], the problem of the locaton selecton of vrtual machnes was dvded nto a mult-obectve optmzaton problem and a bn packng problem but wthout the consderaton of the cost of vrtual machne mgraton; the study focused only on the statc placement of vrtual machnes and dsregarded dynamc deployment based on vrtual machne mgraton. In [18, 20], the problem was consdered as a combnatoral optmzaton problem based on bn packng; smlarly, the authors only consdered the statc placement of vrtual machnes and dsregarded dynamc placement based on vrtual machne mgraton. Most optmzaton methods for vrtual machne placement are mplemented n several phases to solve a sngle-obectve optmzaton problem. Multple targets are rarely optmzed smultaneously. Conseuently, only a local rather than a global optmzaton soluton s obtaned. In sum, extensve research results have been obtaned n the cloud placement of vrtual machnes, but several serous problems, ncludng those enumerated below, have yet to be solved. (1) Most studes determne data centers by selectng the reured physcal nodes and not the most approprate nodes from the cloud data center. A complete strategy for the locaton selecton of vrtual machnes should be dvded nto two levels, namely, data center and physcal machne. (2) Many of the exstng studes on the vrtual machne placement strategy are manly based on a sngle dmenson of target optmzaton under certan rules. Conseuently, the generated optmal placement method may be based on only one certan condton. An effcent strategy for the locaton selecton of vrtual machnes should consder the dependences between dmensonal constrants and comprehensve balance. (3) Many new challenges need to be addressed to dynamcally allocate and manage the shared physcal and vrtual resources of data centers. For example, an adaptve framework for the locaton selecton of effectve vrtual machnes s lackng. In locaton selecton, the cost of resources, system performance, energy consumpton, and other factors must be consdered. An effcent algorthm for the locaton selecton of vrtual machnes must also be desgned such that t s adaptable to dfferent user goals and busness needs.

4 KSII TRANSACTIONS ON INTERNET AND INFORMATION SYSTEMS VOL. 9, NO. 12, December Problem Statement and Related Defntons The total amount of resources utlzed by users n a cloud computng platform s euvalent to the partcular resources of a vrtual machne. Each user applcaton runs n ts own, ndependent vrtual machne. The effcent use of cloud computng resources and the reducton n the cost of such resources are key academc research ssues. 3.1 Problem Statement The mult-dmensonal collaboratve problem of the locaton selecton of vrtual machnes can be consdered a mult-obectve combnatoral optmzaton problem. The avalable resources of each physcal machne, such as the central processng unt (CPU), memory, dsk, and I/O devces, can be used as mult-dmensonal vectors. Every dmenson s a physcal resource for physcal machnes. Each resource reured by a vrtual machne corresponds to a mult-dmensonal vector. The goal of the locaton selecton of vrtual machnes s to place dfferent vrtual machnes to multple physcal nodes accordng to the dfferent needs of multple users. The locaton selecton process must be based on an effectve adaptve framework. The problem of the locaton selecton of vrtual machnes, whch s based on multple targets, can be descrbed as follows. In order to descrbe the problem and defntons convenently and clearly, we desgned two notaton tables n whch the basc meanngs of the man symbols are defned accordngly. SYMBOL H V v M m t R r r L S s Tabel 1. Notaton table (1) DESCRIPTION number of data centres vrtual machne set the th vrtual machne physcal machne set the th physcal machne task set of a vrtual machne resource set of a data centre resource set of th physcal machne the th dmenson resource of the th physcal node resource servce perod useful resource servce vector the useful servce rato of r t the th dmenson resource of kth task on k W w task generaton rato set of vrtual machne task generaton rato of v p n m probablty dstrbuton matrx of data source s n m cloud resource schedulng matrx µ load balance varance of all physcal nodes v

5 4780 Lu et al.: An Adaptve Vrtual Machne Locaton Selecton Mechansm n Dstrbuted Cloud A cloud platform comprses a number of data centers. In ths study, we suppose that H data H centers exst n a cloud platform and denote all resources n the data centers as R. One data center comprses n physcal nodes. R = { r1, r2, r3... r..., rn} denotes the total resources of one data center. r = { r,,,..., r r r } denotes a set of dfferent dmensons resource of m. Durng an entre servce perod, whch s denoted as L, the useful servce tme obeys the Posson dstrbuton [21], and all the resources are ndependent of one another, n.e., r R, I r = φ. In the resource set of a data center, whch s denoted as R, the useful = 0 servce vector s denoted as S = { s, s,..., s } 1 2 n, and the useful servce rate of r s denoted as s (0 n). A vrtual machne set s denoted as V = { v1, v2, v,..., v } 3 m, whch s placed on physcal nodes. The set comprses m vrtual machnes. In ths study, we suppose that v only has one task t at any tme. The kth task, whch s denoted as t = { tk, tk, tk,..., tk},0 m, k ncludes dfferent attrbutes. The task generaton rate of a vrtual machne can be descrbed as W = { w1, w2, w3, w4,... w..., wm}, 0 m, where w denotes the task producton rate of the th m vrtual machne. All ndependent tasks satsfy t V, I = 0t = φ, r Rt, V. If r can satsfy all the reurements of t, then t can be assgned to r, and all operatons can be run. In some cases, many r can satsfy the reurements of t. To acheve an effectve balance among all resources, t s allocated to r wth a probablty of matrx of data resources s denoted as Pn m ( P ) = 0 Table 2. Notaton table (2) SYMBOL DESCRIPTION U, th dmenson resource utlty of m U, utlty of th resource of v whch s assgned to m v the th dmenson resource of v whch s assgned to m m the th dmenson resource of m remanng resource bggest matrx of physcal node E F vrtual machne resource matrx f the reured resource for v f the reured resource number of th dmenson for v C resource blled vector of m c the cost of obtan th dmenson from m mappng matrx of v to m G N Y Q mgraton number of vrtual machne overhead of placng v on n m h= 1 P. The probablty dstrbuton =, wth the condton that P 0 and n P = 1. The cloud resource schedulng matrx s denoted as Sn m = ( S ). If v s assgned m the max number of resource dmenson n m h

6 KSII TRANSACTIONS ON INTERNET AND INFORMATION SYSTEMS VOL. 9, NO. 12, December to m, then S =1 ; otherwse, S =0. Durng all the allocaton processes, the followng condtons must be fulflled: n = 0 n = 0 P S = 1 P S t s n m n P S t s = 0 = 0 = 0. (1) In ths work, we dscuss the proper placement of vrtual machnes accordng to the usage nformaton of mult-dmensonal vectors of such vrtual machnes n the locaton selecton process. We evaluate the resultng performance by referrng to Mcrosoft s 2008 report on vrtual machne management technues n the evaluaton standards for physcal server resources,.e., accordng to the CPU, memory, network bandwdth, and dsk I/O [15]. The problem s descrbed as follows: (1) In n physcal machne nodes ( M= { m 1, m2, m3,..., m n} ), n refers to the number of physcal nodes n the physcal cluster. The man resources nclude memory, CPU, dsk, bandwdth, and I/O. (2) In m vrtual machne nodes ( V = { v1, v2, v3,..., v m } ), m refers to the number of vrtual machnes. The man resource reurements of a vrtual machne nclude memory, CPU, dsk, bandwdth, and I/O. (3) A mappng between vrtual and physcal machnes must be establshed to satsfy all the reurements of vrtual machnes and to reduce the physcal nodes assgned to these vrtual machnes. Durng the mappng process, the sum of the assgned vrtual machne resources cannot exceed the amount of resources of the physcal machne. 3.2 Related Defntons Defnton 1: The largest resource vector of m can supply {,,,..., T r = r r r r }, where r denotes the maxmum number of dmenson resource suppled by m,1 n. The remanng resource matrx of the physcal machne node set can be defned as follows: r1 r1 r1... r r2 r2 r2... r E = r3 r3 r3... r rn rn rn... r n Defnton 2: Gven an adeuate amount of physcal resources, a vrtual machne runs approprately. The reured amount of resources of v can be denoted as {,,,..., } T f = f f f f. f denotes the reured number of th dmenson resource, 1 m. The reured resource matrx of a vrtual machne s denoted as follows:

7 4782 Lu et al.: An Adaptve Vrtual Machne Locaton Selecton Mechansm n Dstrbuted Cloud f1 f1 f1... f f2 f2 f2... f F = f3 f3 f3... f fm fm fm... f m Defnton 3: Resource cost The reured amount of resources s blled. The blled vector can be marked wth T C = {c, c, c,..., c }, where c denotes the cost of obtanng th dmenson resource from m. Defnton 4: Mappng matrx of v to G g11 g 12 g1... g1 m g g g... g = 1 2 m gn 1 gn2 gn... g nm m Here, 1 n, 1 m, and g {0,1}. g denotes whether v s placed on ndcates that V s placed on the m ; otherwse, g =0. m. g = 1 Defnton 5: Let x {0,1}. x = 1 ndcates that a vrtual machne mgrates from one physcal node to another physcal node. Otherwse, the vrtual machne does not mgrate. We suppose that S denotes the number of mgratons undertaken by a vrtual machne and n that S= x, x {0,1}. Accordng to ths defnton, the overhead of placng a vrtual machne = 1 T on a physcal node can be denoted as Y, where Y=F G C. Defnton 6: n P = p, p {0,1}. p = 1 ndcates that at least one vrtual machne s assgned = 1 to a physcal node. Otherwse, no vrtual machne s assgned to a physcal node. Defnton 7: µ = Mn( D / ) denotes the balanced load varance of all the physcal machnes. ^2 D denotes the th dmenson varance, and denotes the number of dmensons. D = (p p ) / n, where n denotes the number of physcal nodes. p denotes the average n value of the performance of the th dmenson for all physcal nodes. p s the th performance value of the m. All the performance values are normalzed. In ths study, the optmzaton goal for vrtual machne placement s as follows:

8 KSII TRANSACTIONS ON INTERNET AND INFORMATION SYSTEMS VOL. 9, NO. 12, December mn ( m) mn( Y ) mn(p) N max Q = 1 = 1 U,. (2) Constraned condtons: At any tme, the total amount of all types of resources reured by all vrtual machnes cannot exceed the total amount of resources of all physcal machnes. The amount of resources reured for each vrtual machne should also not exceed that reured for each physcal machne. The concrete constraned condtons can be formalzed as follows: ( a) n r R, I r = = 0 f ( b) U, T, n m ( c), g f r = 1 = 1 m ( d ) f R = 1 m ( e), v x m = 1 T, denotes the threshold of the resource for m, and U, denotes the useful rato of the resource n v, whch s assgned to m. The neualty (e) above ndcates that the th dmenson resource (.e. CPU, memory, dsk space, bandwdth, and I/O) of the vrtual machne that are assgned to the physcal machne m cannot exceed those of the physcal machne m. If a user applcaton can be placed on X vrtual machnes, then the performance factors of the vrtual machnes suppled to the applcaton P must satsfy servce-level agreements (SLAs). (3) 4. Chromosome Gene Encodng and Evaluaton Functon Desgn In ths part, we wll ntroduce the process of chromosome gene encodng and evaluaton functon desgn. Usng the GA (Genetc Algorthm) technology, we can explan the codng nformaton of vrtual machne and physcal machne clearly and easly. In addton, evaluaton functon can be expressed combned wth the codng nformaton. 4.1 Encodng Encodng s generally the frst factor that can affect the search effcency of evolutonary algorthms. Encodng can reflect the mappng relatons of solutons to chromosomes [16]. For the locaton selecton of vrtual machnes, the soluton codng of the physcal nodes to whch vrtual machnes are assgned can be consdered a chromosome. The assgned vrtual machne s the gene value. The advantage of ths codng s that the number of chromosome genes s determned by the total number of physcal servers [22]. Therefore, ths stuaton cannot

9 4784 Lu et al.: An Adaptve Vrtual Machne Locaton Selecton Mechansm n Dstrbuted Cloud reduce computng speed because of the emergence of extensve codng. Durng the subseuent crossover and mutaton operatons, the hardware constrants among all servers reman the same [22]. In the present study, the problem of the locaton selecton of vrtual machnes can be regarded as a bn packng problem. Three chromosome encodng mechansms are avalable: codng based on cases, codng based on tems, and codng based on groups [3]. Codng based on cases and codng based on groups focus on ndvdual tems. The goal functon of bn packng depends on tem groups. In ths study, the encodng mechansm s manly based on tem groups. Accordng to the combned redundances of group encodng [23], a doubly lnked lst codng method s proposed. M vrtual machnes are assgned to N physcal nodes. M s generally larger than N. A random vrtual seuence that ncludes M vrtual machnes s generated. For chromosome encodng, a prorty heurstc algorthm s employed to place a random vrtual machne seuence on physcal nodes. Ths prorty heurstc algorthm selects a physcal node from used physcal nodes. If the selected node can satsfy the fve resource reurements (CPU, memory, network bandwdth, dsk, and I/O) of the frst vrtual machne, then the vrtual machne can be placed on the physcal node. Otherwse, the subseuent physcal nodes are selected untl the physcal node that satsfes such reurements s found (let the resources of N physcal machnes satsfy the resource reurements of M vrtual machnes). If no proper physcal node from the used physcal nodes can satsfy the resource reurement, then the frst new physcal node that has not been used s selected for the placement of the vrtual machne (when the algorthm s mplemented ntally, all physcal machne nodes are not used). Accordng to the descrpton above, all vrtual machnes can be placed on physcal nodes. The concrete algorthm s descrbed as follows: Algorthm 1: Prorty heurstc algorthm For (=0; <=n; ++) { VMFlag=0; For (=0;<=m;++) { If (VMp ==1) { If Rp(c,m,s,b,/o)>Rv(c,m,s,b,/o) { VM stored n P; VMFlag=1; Break; } Else ++; } } If (VMFlag==0) VM stored n the frst unused P }

10 KSII TRANSACTIONS ON INTERNET AND INFORMATION SYSTEMS VOL. 9, NO. 12, December The encodng process of assgnng vrtual machnes to physcal nodes, whch s based on group nformaton and a doubly lnked lst, s shown n the example. Let us assume the avalablty of vrtual machnes, whch need to be allocated at a certan perod, as well as a seuence number of. The seuence number s random. Vrtual machnes are assgned to physcal nodes randomly accordng to the pror algorthm, and the ntal code can then be generated. The ntal code s optmzed accordng to the doubly lnked lst packng method. Fg. 1. Encodng process 4.2 Evaluaton Functon The optmal goals are a mnmal number of used physcal machnes, low unt prce costs of resources for users, a hgh performance of an applcaton system acheved wth a balanced load, and a low useful rato of physcal nodes. An evaluaton functon can be used to evaluate ndvduals. The frst parameter s a key factor to evaluate the chromosome load performance of used physcal nodes. A small load varance denotes the good load performance of used physcal nodes. The second parameter s the lowest possble resource cost for users. The thrd parameter s used to evaluate the energy consumpton level n a chromosome usng the number of physcal

11 4786 Lu et al.: An Adaptve Vrtual Machne Locaton Selecton Mechansm n Dstrbuted Cloud nodes. A mnmal number of used physcal nodes denote nsgnfcant energy consumpton. The fourth parameter s used to evaluate the physcal machne utlzaton of a specfc resource type. In ths study, the ftness functon s defned as follows: 4.3 Evoluton Operators Ftfuncton( m, M, P, U, ) = {mn( m), mn M ( Y ), mn( P), max U, }. (4) N Q = 1 = 1 For the problem of assgnng a vrtual machne to physcal machne, a new codng method based on group nformaton and doubly lnked lst nodes s used Crossover operator The vrtual machne code contans the precursor and subseuent nodes. A chromosome gene s expressed n dfferent groups and n an nternal bdrectonal chan structure. Consderng that the arrval of the vrtual machne seuence s random, the seuence length s not fxed. The number of vrtual machnes that each physcal machne can accommodate also dffers. The varable length of the chromosomes referred by the crossover operator s not fxed as well. The genetc algorthm nvolves physcal machne codng and vrtual machne group codng. Based on the proposed codng method, a sngle-pont crossover method wth the longest doubly lnked lst s put forward n ths study. The crossover process n the genetc algorthm manly allows the offsprng to nhert excellent genes from parent generaton. The crossover process s dvded nto two parts. One part s the process based on cross-group codng to mnmze the number of used physcal machnes. The other part s the process based on cross-resource codng to maxmze the utlty of physcal machne resources. In the frst part (as shown n the Fg. 2 below), the man steps of the sngle-pont crossover optmzaton method for sub-ndvdual codng optmzaton based on the largest chan length (for the optmzaton of ndvdual speces) are as follows: Step a) Two parent nodes are selected randomly. The longest group n parent ndvdual A s selected and replaced wth the correspondng group of parent ndvdual B. Step b) After the crossng operaton for parent ndvdual B, the correspondng physcal nodes that comprse repeated vrtual machnes are removed from the collecton of used physcal nodes and are then added to the unused physcal node set; n ths case, the vrtual machnes are deleted [16]. Step c) The unallocated vrtual machne seres s generated on the bass of the retaned vrtual machne chan structure of parent ndvdual B. Step d) The prorty heurstc algorthm s used to generate new chld enttes from the unallocated vrtual machne seuence. Step e) A group s randomly selected from parent ndvdual B and s then used to replace the correspondng group n parent ndvdual A. Indvduals B are generated accordng to Steps a) d). For example, 12 vrtual machnes and 4 physcal nodes are smulated. The process of the sngle-pont crossover method [16] s shown n Fg Mutaton operator Two scenaros emerge n chromosome varaton. One s based on the group encodng varatons of a doubly lnked lst,.e., deletng a vrtual machne n the parent node lst randomly. The other scenaro refers to resource codng varants, n whch we can delete a doubly lnked lst node of a vrtual machne. After such deleton, the precursor and subseuent nodes change

12 KSII TRANSACTIONS ON INTERNET AND INFORMATION SYSTEMS VOL. 9, NO. 12, December smultaneously. The occurrence of these two types of mutaton does not follow a specfc dscplne; they may occur n the form of a sngle ndvdual or n a specfc moment. When mutaton occurs, X physcal nodes are randomly selected from the chromosome (where X < M / 2). X physcal nodes are removed from the collecton of used physcal nodes and are then added to the collecton of unused physcal nodes. The vrtual machnes deployed on the physcal nodes are also deleted. The unallocated vrtual machne seres s then generated on the bass of the chan structure of ndvdual vrtual machnes, whch preserves ndvdual varaton. Fnally, the unallocated seuence of vrtual machnes s redstrbuted n accordance wth the dstrbuton of the prorty heurstc algorthm. The value of parameter X, whch can affect the performance of the mult-obectve evolutonary algorthm, depends on mutaton rate. The value s determned wth an experment [16]. We take note of the followng detals. 1. When vrtual machnes have been ntegrated, we wll ensure that the amount of resources of the physcal machne s less than or eual to the threshold of each type of resource value durng the process of crossover or mutaton. 2. In the process of crossover or mutaton based on resource codng, the remanng resources of the cross secton selected by the physcal machne must be greater than those of the selected vrtual machne. The process of the mutaton operator method s shown n Fg. 3. Fg. 2. Sngle-pont crossover process

13 4788 Lu et al.: An Adaptve Vrtual Machne Locaton Selecton Mechansm n Dstrbuted Cloud Populaton update method Fg. 3. Mutaton process The domnant relaton s a weakened form of the Pareto domnance relaton [24]. In ths study, the form s added for the gven. The concrete defnton of ths form s as follows: Defnton 8: domnant relaton Let whch s called, f and only f and, whch s marked as. Defnton 9: -Pareto optmal approxmate soluton set Set s called a -Pareto optmal approxmate soluton set of f and only f for any, the attrbute s always true and s always true [24]. Defnton 10: -Pareto soluton set Set s called a -Pareto soluton set of f and only f s a -Pareto optmal approxmate soluton set of X and [24].

14 KSII TRANSACTIONS ON INTERNET AND INFORMATION SYSTEMS VOL. 9, NO. 12, December Defnton 11: Immune memory Immune memory s an mportant characterstc of the mmune system. When the body s exposed to an antgen for the second tme, the ncubaton perod of the antbodes relatve to the frst response tme s obvously short. In ths case, the antbody levels rse rapdly. When the same antgen nvades the body agan, a strong prmary mmune system develops; the phenomenon n whch antbodes gan a hgher degree of affnty s referred to as mmune memory [25]. The current non-domnant ndvduals are stored on the bass of the memory set. ε domnant s an effectve form of the relaxaton mechansm of the Pareto domnance, whch s used to mantan the unform dstrbuton of a soluton. Ths form s wdely appled to mantan dversty doman. ε domnant s adopted to update the memory populaton. For a 1 f K, = 1,2,..., m, mult-goal optmzaton problem that ncludes m goal functons, f ( ) then the goal space can be dvded nto (( K 1 ) / ε ) m sub-spaces accordng to the rules of ε domnant. In every sub-space, only one ndvdual exsts, and other ndvduals are deleted from the populaton space. The ε domnant mechansm s senstve to a varety of problems. Gven such varety, dfferent numbers of antbodes are kept. From a practcal vewpont, decders cannot know the geometrc dstrbuton form of the Pareto fronter n advance. If a target value of multple antbodes s less than a certan value, decders do not vew two antbodes as dfferent. In ths case, the use of the ε domnant mechansm s approprate. We adopt a control mechansm to provde decson rghts that can meet such a stuaton and to mantan populaton dversty. In practce, decson makers can dynamcally control the values of a vector. Allocatng an dentfcaton vector for ndvduals s a smple method that can be used to dvde the antbodes of a lvng space [25]. The B X = B X, B X,..., B X and dentfcaton vector can be defned as ( a) ( 1( a) 2( a) m( a) ) B ( Xa) f ( Xa) / ε, ( 1,2,..., m ) as MS ( t) ( MS MS MS ) = =. The current memory set can be denoted = 1, 2,..., n. A new generaton populaton s generated after the crossover and mutaton operatons and s then crossed wth sub-populatons. The result s sorted wth the rule of ε non-domnated sortng. Accordng to the domnaton relatonshp among ndvduals, the memory nformaton of antbodes, the Eucldean dstance, and super volume [26], the subseuent generaton of speces evolves further. The update mechansm of concrete super body populatons s descrbed as follows: a) If the current generaton s the frst generaton, then the non-domnant ndvdual should be selected n the ntal populaton. Otherwse, the non-domnant ndvdual of the current populaton and the parent populaton s assgned to the ndvdual populaton set. The large ndvdual populaton s sorted after beng overlad accordng to the domnant relaton. Let the grade number be n, and let the non-domnant set be denoted as L1, L2,..., Ln. The new dentfcaton vector for the non-domnant populaton s assgned. If some of the dentfcaton vectors of some ndvduals are the same, then they belong to the same super ndvdual. All dentfcaton vectors can be used to extract and dentfy super ndvduals. b) Let the populaton scale be M, and let the ndvdual number of L 1 be M. The set of L 1 can be regarded as the subseuent generaton. c) If the ndvdual number of L 1 s more than M, then all the memory nformaton of the antbodes and the contrbuton value of the super ndvduals are calculated n L 1, except for

15 4790 Lu et al.: An Adaptve Vrtual Machne Locaton Selecton Mechansm n Dstrbuted Cloud some ndvduals wth goal lmt values. The n 2 contrbuton values that are relatvely large are selected, and the ndvduals whose values satsfy the goal value are selected as the subseuent generaton populaton and marked as Z( g+ 1). d) If the ndvdual number of L 1 s less than M, then all the ndvduals of M are added to the subseuent generaton marked as Z( g+ 1). The ndvduals of L2, L3,..., Ln are added to Z( g+ 1) untl the scale of the populaton reaches M. After such addton, f the number of ndvduals of L s more than M, then the Eucldean dstance of the vector ndvduals wth the same super ndvduals s calculated. The ndvduals wth small values of Eucldean dstance on the subseuent generaton untl the populaton scale reaches M. 5. Immune Memory-based Algorthm for the Locaton Selecton of Vrtual Machnes The optmzaton process wth the mult-obectve optmzaton algorthm has obvous dsadvantages despte the wde use of the non-domnated sortng genetc algorthm 2 and the double F-shaped resonator (DFR). For the problem of mult-target optmzaton, mproved performances and optmal results can be acheved for dfferent frontal shapes. In ths study, the basc evoluton operators of cloud resource allocaton are thus embedded n SMS-EMOA [16]. We propose a new mult-obectve vrtual machne locaton selecton algorthm (MOVMLSA) accordng to the mult-obectve selecton process based on domnated hyper-volume [27]. The man process of the algorthm s descrbed as follows: a) Accordng to the group nformaton based on doubly lnked lst encodng, parent ndvduals are ntalzed to form the parent populaton wth a scale of M. b) The ftness evaluaton functon s calculated n vew of the populaton ndvduals. The parent populaton that has been evaluated s sorted wth the ε non-domnated rule. The control level of the ndvduals n the populaton s dvded. c) Two parent ndvduals are selected from populatons usng the tournament selecton method, and a hybrd ndvdual s generated usng the sngle-pont crossover method for the two parent ndvduals. d) Intellgent hybrd ndvdual varaton occurs, and ndvduals are generated. The chld ndvduals are frst evaluated wth the ftness evaluaton functon. The chld ndvduals are then added to the populaton pool. Ths step s repeated untl the populaton sze reaches M. e) The best M ndvduals are selected as the subseuent generaton of the parent populaton on the bass of the update mechansm of the mmune memory. f) For the new generaton of the parent populaton, steps c) f) are repeated untl the evoluton condton s met (maxmum number of teratons). The algorthm s then termnated.

16 KSII TRANSACTIONS ON INTERNET AND INFORMATION SYSTEMS VOL. 9, NO. 12, December Fg. 4. Algorthm flow 6. Expermental Results and Analyss The expermental results show that the proposed MOVMLSA based on mmune memory can determne the reasonable placement of vrtual machnes, mprove the utlzaton of physcal resources, and reduce the cost of resources for users. The algorthm features a certan degree of feasblty and hgh accuracy. A prorty heurstc algorthm (.e., the proportonal hazards model (PHM)), the classc NSG-2 algorthm, the mult-dmensonal domnant resource farness (DRF) schedulng algorthm [28], and the mult-obectve evolutonary algorthm based on mmune memory are all smulated n ths study to verfy the performance of the mmune system-based mult-dmensonal strategy for the locaton selecton of vrtual machnes. The number of used physcal machnes, the load balance of the physcal machnes, and the resource cost are all recorded; these values can be used to compare the performance of the dfferent algorthms. The smulaton expermental platform s the CloudSm, whch was proposed by the teams from the Grd Laboratory of the Unversty of Melbourne n Australa and the Grdbus Proect [6]. Some correspondng classes based on the base classes of CloudSm are modfed to mplement all the algorthms. Several nherted classes and methods are also desgned. The exstng useful resource vector of every physcal machne can be obtaned through the host class. The reured resource vector of a vrtual machne can be generated from the Datacenter class. The VMProvsoner class s used to acheve the mappng from a physcal machne to a vrtual machne. The VMAllocaton polcy s an abstract class that s used to acheve the locaton

17 4792 Lu et al.: An Adaptve Vrtual Machne Locaton Selecton Mechansm n Dstrbuted Cloud selecton of vrtual machnes. The HostForVm allocaton method can realze the placement of a specal vrtual machne to a fxed physcal machne. The nherted VMAllocatonPolcy class can provde the assgnment strategy of vrtual machnes. The allocaton algorthm of vrtual machnes based on the mult-obectve genetc mmune memory can be mplemented by edtng the user-defned nhertance class named VMAllocaton Polcy. Wth ths class, vrtual machnes can select proper physcal machnes to locate. Wth the method of extendng classes n CloudSm, all classes can be rebult. Forty physcal nodes are smulated n the CloudSm smulaton platform. Every node s eupped wth two processors (Intel(R) Core(TM) U 1.7 GHz), a 4 GB memory storage, 8 MB L2 cache, and two dsks of 7,200 turns wth 500 GB. The vrtual numbers are 30, 45, 55, 65, and 80 n turn. Table 3. Parameter desgn of the algorthm for the locaton selecton of vrtual machnes n the CloudSm platform Number of vrtual Number of physcal nodes machnes Algorthm PHM, NSG-2, DRF, MOVMLSA PHM, NSG-2, DRF, MOVMLSA PHM, NSG-2, DRF, MOVMLSA PHM, NSG-2, DRF, MOVMLSA PHM, NSG-2, DRF, MOVMLSA The populaton sze of the evoluton algorthm and the evoluton number are set to 50 and 1,000, respectvely. The crossover and mutaton ratos are set to 0.6 and 0.01, respectvely. Crosser and mutaton rates have very mportant nfluence on the expermental results. For general generc algorthm, the crosser and mutaton rates are always constant. However, determnng the concrete value of both rates s dffcult. If the value s extremely small, an optmal soluton s always more approxmately acheved; however, ensurng that the soluton s the global optmal soluton s dffcult. On the other hand, f the value of crosser and mutaton rates are too large, the number of teratons also become large. Moreover, the searchng ablty becomes extremely low. At the same tme, the algorthm s dffcult to converge. To determne the expermental parameters, we conducted experments, whch can be used to determne the correlaton between mutaton rate and convergence speed, 1000 tmes. Table 4. Class crossover and mutaton rates Crossover rate Mutaton rate Maxmum evoluton number Populaton sze , An evoluton experment s conducted accordng to the above parameters n Table 3 and Table 4. The expermental results are shown n Table 5.

18 KSII TRANSACTIONS ON INTERNET AND INFORMATION SYSTEMS VOL. 9, NO. 12, December Algorthm PHM NSG-2 DRF MOVMLSA PHM NSG-2 DRF MOVMLSA PHM NSG-2 DRF MOVMLSA PHM NSG-2 DRF MOVMLSA PHM NSG-2 DRF MOVMLSA Table 5. Performance comparson of four algorthms Number of Number of Load balance vrtual physcal varance of physcal machnes machnes set Resource utlzaton The ntutve data graph shown n Fgs. 5 7 s used to further analyze the algorthm performance data shown n Table 4. The numbers of the enabled physcal nodes n the locaton selecton algorthm based on mmune memory and the DFR algorthm are approxmately smlar, but they are lower than those n the PHM and NSG-2 algorthms. A low number of enabled physcal nodes ndcates consderable savngs n energy and resources. For the load performance of a cluster, the load performance varance n the locaton selecton algorthm based on the mult-obectve genetc algorthm s less than those n the other three algorthms. A small load performance varance denotes the good effect of the load-balancng server cluster. The precedng analyss mples that the locaton selecton algorthm based on the mult-obectve genetc algorthm can greatly reduce the number of used physcal servers and can acheve a good load-balancng server cluster effect PHM MSG-2 DFR MOVMSLA Resource ultly (%) Number of Vrtual Machnes Fg. 5. Comparson of resource utlty

19 4794 Lu et al.: An Adaptve Vrtual Machne Locaton Selecton Mechansm n Dstrbuted Cloud Number of Physcal machnes PHM NSG-2 DRF MOVMLSA Number of Vrtual Machnes Fg. 6. Comparson of the number of enabled physcal machnes Degree of Loadbalance varance PHM NSG-2 DFR MOVMLSA Number of Vrtual Machnes Fg. 7. Comparson of load balance degrees To obtan the comparson results of PHM, NSG-2, DFR, and MOVMLSA, we performed fve ndependent group experments (the reured numbers of vrtual machnes are separately set to 30, 45, 55, 65, and 80). Thus, the physcal machne number, resource load balance, and resource utlty can be analyzed adeuately. (1) Analyss comparson of comprehensve resource utlty: The comprehensve resource denotes CPU, memory, network, and I/O resource. From the comparatve experments of the frst group data to the ffth group data, the changes were recorded accordng to the changes n the number of vrtual machnes. The resource utlty stuaton of PHM, NSG-2, DRF, and MOVMLSA are shown n Fg. 5. As shown n the fgure, under the same condtons, the

20 KSII TRANSACTIONS ON INTERNET AND INFORMATION SYSTEMS VOL. 9, NO. 12, December physcal resource utlty s the hghest when MOVMLSA s used. When the number of vrtual machnes s adusted from 30 to 80, the comprehensve resource utlty usng MOVMLSA s 10% 14% hgher than that of PHM, NSG-2, and DRF. Thus, MOVMLSA can obvously mprove resource utlty and save energy. (2)Analyss comparson of the number of enabled physcal machne: The number of vrtual machnes s adusted, and the reured physcal machne number s shown n Fg. 6. The fgure reflects the dfference between NSG-2, DRF, and MOVMLSA. When dfferent algorthms are used to assgn the vrtual machne to the physcal machne, MOVMLSA can assgn the vrtual machne effcently, whch s reflected by usng physcal machnes at least under the same condtons. Of course, the number of reured physcal machnes ncreases wth the ncrease n the number of vrtual machnes. The expermental results show that compared wth PHM, whch consumes plenty of resources, MOVMLSA can degrade the comprehensve resource utlty from 35% to 62%. Compared wth NSG-2 and DRF, MOVMLSA can degrade the comprehensve resource utlty as well. Thus, MOVMLSA can mprove resource utlzaton. (3) Analyss comparson of resource load balance: In the experments, resource threshold value, attrbutes of physcal machnes, and vrtual machne tasks are consstent wth one another, except for the deployment method. The cross and mutaton rates are set at 0.6 and 0.01, respectvely, and the maxmum genetc algebra s set at The vrtual machne deployment process was recorded from 30 vrtual machnes to 80 vrtual machnes n the experment. The man goal of ths experment s to obtan the comparson of the result of resource load balance n a fxed number of vrtual machnes wth dfferent locaton selecton algorthms. The comparson results are presented n Fg. 7. The experments show that wth the ncrease n the number of vrtual machnes, the algorthm can decrease the unbalanced degree of the resources. From the longtudnal contrast, under the same condtons relatve to PHM, NSG-2, and DFR, MOVMSLA can reduce the resource mbalance degree from 0.03% to 0.23%. Compared wth three other algorthms, MOVMLSA contrbutes more effectvely to the performance of resource load balance. However, ths advantage s not partcularly obvous because many uncertan factors n determnng genetc algorthm parameters exst. In our future work, the aspects of optmzaton parameters based on the characterstcs of the algorthm tself would be mproved. In theory, the mult-obect vrtual machne locaton selecton problem, whch s proposed n ths paper, s a combnatoral optmzaton problem. In cloud computng envronment, however, combnatoral optmzaton problems are lkely to lead to combnaton exploson. Prortzaton process between targets can be avoded because the genetc algorthm can be processed parallel to each target. Thus, the genetc algorthm can effectvely reduce the nvald combnaton. Therefore, the algorthm s sutable for solvng a mult-obect optmzaton problem [10]. MOVMLSA s a mult-obect locaton selecton algorthm based on mmune memory, whch s also classfed as a knd of genetc algorthm. MOVMLSA s effectve for solvng mult-obect combnatoral optmzaton problems. However, at present, the genetc algorthm faces the problem of slow convergence speed. To mprove the convergence speed of the genetc algorthm, MOVMLSA can choose excellent ndvduals from each generaton populaton; then, the mmune nformaton s extracted from the excellent ndvduals, and a vaccnaton for mmunzaton s prescrbed to the dagnosed offsprng. Immune memory can speed up the breedng of good modes and repar the damaged modes by crossover and mutaton model of excellence. Vrtual machne populaton and mmune memory nformaton can nteract and cooperate wth one another, whch can greatly mprove the convergence speed of MOVMLSA. MOVMLSA can automatcally obtan the mmunzaton nformaton of an ndvdual, and ths nformaton s dynamcally updated wth the evoluton of the populaton.

21 4796 Lu et al.: An Adaptve Vrtual Machne Locaton Selecton Mechansm n Dstrbuted Cloud In ths way, we can fnd the subspace that may be ncluded n the optmum soluton and mprove the searchng effcency of the algorthm. The scale of mmune memory lbrary s an mportant factor that can nfluence the effcency of the algorthm. However, n ths study, we dd not examne the relatonshp between the populaton sze and the mmunzaton type database scale, whch wll be nvestgated n our future work. 7. Concluson The MOVMLSA based on the mult-obectve evolutonary algorthm and that based on mmune memory are dscussed n ths work. To solve the problem of mappng physcal machnes to vrtual machnes, we transform the problem of the locaton selecton of vrtual machnes nto a mult-obectve packng problem accordng to the mult-dmensonal resource characterstcs of vrtual machnes and the characterstcs of optmzaton goals. The soluton of the mult-obectve genetc algorthm s obtaned on the bass of the genetc memory nformaton. A chromosome evaluaton functon and a group chan code are desgned accordng to the doubly lnked lst and group nformaton. The maxmum lengths of the cross operator, sngle-pont crossover operator, and mutaton operator of the X-pont mutaton are desgned accordng to the code nformaton. Compared wth other evolutonary algorthms, the mult-obectve genetc algorthm based on mmune memory shows a better performance for dfferent frontal shapes. An algorthm for the locaton selecton of vrtual machnes based on mmune factors s desgned n a cloud computng platform. The algorthm update polcy s based on the volume of the update mechansm, whch can ensure the populaton ualty and dversty. PHM, NSG-2, DFR, and MOVMLSA are smulated to verfy the performance of the new algorthm. The expermental results show that the MOVMLSA based on mmune operators performs better than the other algorthms n terms of resource utlzaton, cluster load balance, and resource cost. References [1] D. Bretgand and A. Epsten, Improvng consoldaton of vrtual machnes wth rsk-aware bandwdth oversubscrpton n compute clouds, IEEE INFOCOM 2012-IEEE Conference on Computer Communcatons, pp , Artcle (CrossRef Lnk). [2] M. Alcherry and T. Lakshman, Optmzng data access latences n cloud systems by ntellgent vrtual machne placement, IEEE INFOCOM 2013-IEEE Conference on Computer Communcatons, pp , Artcle (CrossRef Lnk). [3] Xaoao Chen, Shhpng Chen and Fang Fang, Vrtual machne resource allocaton algorthm n cloud computng, Computer Applcaton Research, vol. 31, no. 9, pp , Artcle (CrossRef Lnk). [4] M. Stllwell, D. Schanzenbach, F. Vven, and H. Casanova, Resource allocaton algorthms for vrtualzed servce hostng platforms, Journal of Parallel and Dstrbuted Computng, vol. 70,no. 9, pp , Artcle (CrossRef Lnk). [5] M. Gahlawat and P. Sharma, Survey of vrtual machne placement n federated Clouds, n Proc. of Advance Computng Conference (IACC), 2014 IEEE Internatonal, pp , Artcle (CrossRef Lnk). [6] Bo Xu, Chao Zhao, Yanun Zhu and Zhpng Peng, Vrtual machne resource schedulng mult-obectve optmzaton n cloud computng, Journal of System Smulaton, vol. 26,no. 3, pp , Artcle (CrossRef Lnk).

22 KSII TRANSACTIONS ON INTERNET AND INFORMATION SYSTEMS VOL. 9, NO. 12, December [7] X. Kong, C. Ln, Y. Jang, W. Yan, and X. Chu, Effcent dynamc task schedulng n vrtualzed data centers wth fuzzy predcton, Journal of Network and Computer Applcatons, vol. 34, no. 4 pp , Artcle (CrossRef Lnk). [8] F. Hao, M. Kodalam, T. Lakshman, and S. Mukheree, Onlne allocaton of vrtual machnes n a dstrbuted cloud, n Proc. of IEEE INFOCOM IEEE Conference on Computer Communcatons, pp , Artcle (CrossRef Lnk). [9] D. Warneke and O. Kao, Explotng dynamc resource allocaton for effcent parallel data processng n the cloud, IEEE Transactons on Parallel and Dstrbuted Systems, vol. 22, no. 6, pp , Artcle (CrossRef Lnk). [10] S. T. Magulur, R. Srkant, and L. Yng, Stochastc models of load balancng and schedulng n cloud computng clusters, n Proc. of IEEE INFOCOM 2012-IEEE Conference on Computer Communcatons, pp , Artcle (CrossRef Lnk). [11] J. W. Jang, T. Lan, S. Ha, M. Chen, and M. Chang, Jont VM placement and routng for data center traffc engneerng, n Proc. of IEEE INFOCOM 2012-IEEE Conference on Computer Communcatons, pp , Artcle (CrossRef Lnk). [12] K. Sato, M. Samema, and N. Komoda, Dynamc optmzaton of vrtual machne placement by resource usage predcton, n Proc. of th IEEE Internatonal Conference on Industral Informatcs, pp.86-91, Artcle (CrossRef Lnk). [13] J. Xu and J. A. Fortes, Mult-obectve vrtual machne placement n vrtualzed data center envronments, n Proc. of Green Computng and Communcatons (GreenCom), 2010 IEEE/ACM Int'l Conference on & Int'l Conference on Cyber, Physcal and Socal Computng (CPSCom), pp , Artcle (CrossRef Lnk). [14] M. Chen, H. Zhang, Y.-Y. Su, X. Wang, G. Jang, and K. Yoshhra, Effectve VM szng n vrtualzed data centers, n Proc. of Integrated Network Management (IM), 2011 IFIP/IEEE Internatonal Symposum on, pp ,2011. Artcle (CrossRef Lnk). [15] Q. L, Q.-F. Hao, L.-M. Xao, and Z.-J. L, Adaptve management and mult-obectve optmzaton for vrtual machne placement n cloud computng, Chnese Journal of Computers, vol. 34, no.12, pp , Artcle (CrossRef Lnk). [16] Haoun A, Suwen Gong and Yuanmng Yuan, Research of cloud computng vrtual machne allocated strategy on Mult-obectve evolutonary algorthm, Computer Scence, vol. 41, no.6, pp , Artcle (CrossRef Lnk). [17] S. Sawant, A genetc algorthm schedulng approach for vrtual machne resources n a cloud computng envronment, Master's Proects. Paper 198, Artcle (CrossRef Lnk). [18] J. Gu, J. Hu, T. Zhao, and G. Sun, A new resource schedulng strategy based on genetc algorthm n cloud computng envronment, Journal of Computers, vol. 7, no.1 pp.42-52, Artcle (CrossRef Lnk). [19] L MF, B JP, L ZC, Resource Schedulng Watng Aware Vrtual Machne Consoldaton, Journal of Software, vol.25, no. 7, pp , Artcle (CrossRef Lnk). [20] G. Hamlton, Dstrbuted vrtual machne mgraton for cloud data center envronments, Unversty of Glasgow, Artcle (CrossRef Lnk). [21] Da-we Sun, Guran Chang, Fengyun L, Chuan Wang, and Xngwe Wang, Optmzng mult-dmensonal QoS cloud resource schedulng by mmune clonal wth preference, Acta Electronca Snca, vol. 39, no.8, pp , Artcle (CrossRef Lnk). [22] Apng Yuan and Canun Wan, Vrtual machne deployment strategy based on mproved genetc algorthm n cloud computng envronment, Journal of Computer Applcatons, vol. 34, no. 2, pp.357-9,364, Artcle (CrossRef Lnk). [23] Ö. Ülker, E. E. Korkmaz, and E. Özcan, A groupng genetc algorthm usng lnear lnkage encodng for bn packng, n Proc. of Parallel Problem Solvng from Nature PPSN X, ed: Sprnger, pp ,2008. Artcle (CrossRef Lnk). [24] Dongdong Yang, Llcheng Jao,Maoguo Gong and Hang Yu, Clonal selecton algorthm for solvng preference mult-obectve optmzaton, Journal of Software, vol. 21, no.1, pp , Artcle (CrossRef Lnk).

23 4798 Lu et al.: An Adaptve Vrtual Machne Locaton Selecton Mechansm n Dstrbuted Cloud [25] Yong Lu, Xnhua Wang, Changmng Xng and Shuo Wang, Resources schedulng strategy based on ant colony optmzaton algorthms n cloud computng, Artcle (CrossRef Lnk). [26] M. Emmerch, N. Beume, and B. Nauoks, An EMO algorthm usng the hypervolume measure as selecton crteron, Lecture Notes n Computer Scence, vol.3410, pp.62-76, Artcle (CrossRef Lnk). [27] N. Beume, B. Nauoks, and M. Emmerch, SMS-EMOA: Mult-obectve selecton based on domnated hypervolume, European Journal of Operatonal Research, vol. 181, no. 3, pp , Artcle (CrossRef Lnk). [28] Png Guo and Q l, Load balance schedulng algorthm based on the load on the server status classfcaton, Journal of Huazhong Unversty of scence and technology: Natural Scence Edton, vol. 40, no. 1, pp.62-65, Artcle (CrossRef Lnk). Shukun Lu receved the M.S. degree n computer scence and technology from Unversty of South Chna, PR Chna, n Currently, He has been a Ph.D. canddate at Central South Unversty snce September 2013, Chna. Hs maor research nterests nclude cloud computng, vrtualzaton technology, performance analyss, computer networks, database technology, data mnng and software engneerng. He has publshed nearly ffteen papers n related ournals. And he s a member of CCF and a member of ACM. Prof. Wea Ja s a Zhyuan Char Professor n the Department of Computer Scence and Engneerng, Shangha Jaotong Unversty. He oned German Natonal Research Center for Informaton Scence (GMD) n Bonn (St. Augustne) from 1993 to 1995 as a research fellow. Durng , he oned Department of Computer Scence, Cty Unversty of HK as a professor. Wea Ja receved BSc and MSc from Center South Unversty, Chna n 1982 and 1984 and Master of Appled Sc. and PhD from Polytechnc Faculty of Mons, Belgum n 1992 and 1993 respectvely, all n Computer Scence. He s the guest Professor of Beng Unversty of Scence and Technology of Chna, Beng Jao Tong Unversty, Jnan Unversty, Guangzhou, Chengdu Unversty, Chna. He has served as the edtors of IEEE TPDS and ComCom and PC chars and members/keynote speakers for varous prestge nternatonal conferences. He s the Senor Member of IEEE and the Member of ACM. Hs research nterests nclude next generaton wreless communcaton, protocols and heterogeneous networks; dstrbuted systems, multcast and anycast QoS routng protocols. In these felds, he has a number of publcatons n the prestge nternatonal ournals (IEEE Transactons, e.g., ToN, TPDS, TC, TMC etc.), books/chapters and refereed nternatonal conference proceedngs (e.g. ACM CCS, WSec, MobHoc, SenSys, IEEE ICDCS, INFOCOM etc.). He (wth W. Zhou) has publshed a book Dstrbuted Network Systems by Sprnger where the book contans extensve research materals and mplementaton examples. He has receved the best paper award n a prestge (IEEE) conference.

Cluster Analysis of Electrical Behavior

Cluster Analysis of Electrical Behavior Journal of Computer and Communcatons, 205, 3, 88-93 Publshed Onlne May 205 n ScRes. http://www.scrp.org/ournal/cc http://dx.do.org/0.4236/cc.205.350 Cluster Analyss of Electrcal Behavor Ln Lu Ln Lu, School

More information

Application of Improved Fish Swarm Algorithm in Cloud Computing Resource Scheduling

Application of Improved Fish Swarm Algorithm in Cloud Computing Resource Scheduling , pp.40-45 http://dx.do.org/10.14257/astl.2017.143.08 Applcaton of Improved Fsh Swarm Algorthm n Cloud Computng Resource Schedulng Yu Lu, Fangtao Lu School of Informaton Engneerng, Chongqng Vocatonal Insttute

More information

An Optimal Algorithm for Prufer Codes *

An Optimal Algorithm for Prufer Codes * J. Software Engneerng & Applcatons, 2009, 2: 111-115 do:10.4236/jsea.2009.22016 Publshed Onlne July 2009 (www.scrp.org/journal/jsea) An Optmal Algorthm for Prufer Codes * Xaodong Wang 1, 2, Le Wang 3,

More information

Multi-objective Virtual Machine Placement for Load Balancing

Multi-objective Virtual Machine Placement for Load Balancing Mult-obectve Vrtual Machne Placement for Load Balancng Feng FANG and Bn-Bn Qu,a School of Computer Scence & Technology, Huazhong Unversty Of Scence And Technology, Wuhan, Chna Abstract. The vrtual machne

More information

Virtual Machine Migration based on Trust Measurement of Computer Node

Virtual Machine Migration based on Trust Measurement of Computer Node Appled Mechancs and Materals Onlne: 2014-04-04 ISSN: 1662-7482, Vols. 536-537, pp 678-682 do:10.4028/www.scentfc.net/amm.536-537.678 2014 Trans Tech Publcatons, Swtzerland Vrtual Machne Mgraton based on

More information

Resource and Virtual Function Status Monitoring in Network Function Virtualization Environment

Resource and Virtual Function Status Monitoring in Network Function Virtualization Environment Journal of Physcs: Conference Seres PAPER OPEN ACCESS Resource and Vrtual Functon Status Montorng n Network Functon Vrtualzaton Envronment To cte ths artcle: MS Ha et al 2018 J. Phys.: Conf. Ser. 1087

More information

Parallelism for Nested Loops with Non-uniform and Flow Dependences

Parallelism for Nested Loops with Non-uniform and Flow Dependences Parallelsm for Nested Loops wth Non-unform and Flow Dependences Sam-Jn Jeong Dept. of Informaton & Communcaton Engneerng, Cheonan Unversty, 5, Anseo-dong, Cheonan, Chungnam, 330-80, Korea. seong@cheonan.ac.kr

More information

Meta-heuristics for Multidimensional Knapsack Problems

Meta-heuristics for Multidimensional Knapsack Problems 2012 4th Internatonal Conference on Computer Research and Development IPCSIT vol.39 (2012) (2012) IACSIT Press, Sngapore Meta-heurstcs for Multdmensonal Knapsack Problems Zhbao Man + Computer Scence Department,

More information

Load Balancing for Hex-Cell Interconnection Network

Load Balancing for Hex-Cell Interconnection Network Int. J. Communcatons, Network and System Scences,,, - Publshed Onlne Aprl n ScRes. http://www.scrp.org/journal/jcns http://dx.do.org/./jcns.. Load Balancng for Hex-Cell Interconnecton Network Saher Manaseer,

More information

Efficient Distributed File System (EDFS)

Efficient Distributed File System (EDFS) Effcent Dstrbuted Fle System (EDFS) (Sem-Centralzed) Debessay(Debsh) Fesehaye, Rahul Malk & Klara Naherstedt Unversty of Illnos-Urbana Champagn Contents Problem Statement, Related Work, EDFS Desgn Rate

More information

Degree-Constrained Minimum Spanning Tree Problem Using Genetic Algorithm

Degree-Constrained Minimum Spanning Tree Problem Using Genetic Algorithm Degree-Constraned Mnmum Spannng Tree Problem Usng Genetc Algorthm Keke Lu, Zhenxang Chen, Ath Abraham *, Wene Cao and Shan Jng Shandong Provncal Key Laboratory of Network Based Intellgent Computng Unversty

More information

A mathematical programming approach to the analysis, design and scheduling of offshore oilfields

A mathematical programming approach to the analysis, design and scheduling of offshore oilfields 17 th European Symposum on Computer Aded Process Engneerng ESCAPE17 V. Plesu and P.S. Agach (Edtors) 2007 Elsever B.V. All rghts reserved. 1 A mathematcal programmng approach to the analyss, desgn and

More information

A Model Based on Multi-agent for Dynamic Bandwidth Allocation in Networks Guang LU, Jian-Wen QI

A Model Based on Multi-agent for Dynamic Bandwidth Allocation in Networks Guang LU, Jian-Wen QI 216 Jont Internatonal Conference on Artfcal Intellgence and Computer Engneerng (AICE 216) and Internatonal Conference on etwork and Communcaton Securty (CS 216) ISB: 978-1-6595-362-5 A Model Based on Mult-agent

More information

CHAPTER 2 PROPOSED IMPROVED PARTICLE SWARM OPTIMIZATION

CHAPTER 2 PROPOSED IMPROVED PARTICLE SWARM OPTIMIZATION 24 CHAPTER 2 PROPOSED IMPROVED PARTICLE SWARM OPTIMIZATION The present chapter proposes an IPSO approach for multprocessor task schedulng problem wth two classfcatons, namely, statc ndependent tasks and

More information

FAHP and Modified GRA Based Network Selection in Heterogeneous Wireless Networks

FAHP and Modified GRA Based Network Selection in Heterogeneous Wireless Networks 2017 2nd Internatonal Semnar on Appled Physcs, Optoelectroncs and Photoncs (APOP 2017) ISBN: 978-1-60595-522-3 FAHP and Modfed GRA Based Network Selecton n Heterogeneous Wreless Networks Xaohan DU, Zhqng

More information

THE PATH PLANNING ALGORITHM AND SIMULATION FOR MOBILE ROBOT

THE PATH PLANNING ALGORITHM AND SIMULATION FOR MOBILE ROBOT Journal of Theoretcal and Appled Informaton Technology 30 th Aprl 013. Vol. 50 No.3 005-013 JATIT & LLS. All rghts reserved. ISSN: 199-8645 www.jatt.org E-ISSN: 1817-3195 THE PATH PLANNING ALGORITHM AND

More information

NUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS

NUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS ARPN Journal of Engneerng and Appled Scences 006-017 Asan Research Publshng Network (ARPN). All rghts reserved. NUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS Igor Grgoryev, Svetlana

More information

A New Token Allocation Algorithm for TCP Traffic in Diffserv Network

A New Token Allocation Algorithm for TCP Traffic in Diffserv Network A New Token Allocaton Algorthm for TCP Traffc n Dffserv Network A New Token Allocaton Algorthm for TCP Traffc n Dffserv Network S. Sudha and N. Ammasagounden Natonal Insttute of Technology, Truchrappall,

More information

Available online at Available online at Advanced in Control Engineering and Information Science

Available online at   Available online at   Advanced in Control Engineering and Information Science Avalable onlne at wwwscencedrectcom Avalable onlne at wwwscencedrectcom Proceda Proceda Engneerng Engneerng 00 (2011) 15000 000 (2011) 1642 1646 Proceda Engneerng wwwelsevercom/locate/proceda Advanced

More information

BioTechnology. An Indian Journal FULL PAPER. Trade Science Inc.

BioTechnology. An Indian Journal FULL PAPER. Trade Science Inc. [Type text] [Type text] [Type text] ISSN : 0974-74 Volume 0 Issue BoTechnology 04 An Indan Journal FULL PAPER BTAIJ 0() 04 [684-689] Revew on Chna s sports ndustry fnancng market based on market -orented

More information

Available online at ScienceDirect. Procedia Computer Science 46 (2015 )

Available online at  ScienceDirect. Procedia Computer Science 46 (2015 ) Avalable onlne at www.scencedrect.com ScenceDrect Proceda Computer Scence 46 (2015 ) 558 565 Internatonal Conference on Informaton and Communcaton Technologes (ICICT 2014) A Novel Famly Genetc Approach

More information

Virtual Memory. Background. No. 10. Virtual Memory: concept. Logical Memory Space (review) Demand Paging(1) Virtual Memory

Virtual Memory. Background. No. 10. Virtual Memory: concept. Logical Memory Space (review) Demand Paging(1) Virtual Memory Background EECS. Operatng System Fundamentals No. Vrtual Memory Prof. Hu Jang Department of Electrcal Engneerng and Computer Scence, York Unversty Memory-management methods normally requres the entre process

More information

Load-Balanced Anycast Routing

Load-Balanced Anycast Routing Load-Balanced Anycast Routng Chng-Yu Ln, Jung-Hua Lo, and Sy-Yen Kuo Department of Electrcal Engneerng atonal Tawan Unversty, Tape, Tawan sykuo@cc.ee.ntu.edu.tw Abstract For fault-tolerance and load-balance

More information

Cost-efficient deployment of distributed software services

Cost-efficient deployment of distributed software services 1/30 Cost-effcent deployment of dstrbuted software servces csorba@tem.ntnu.no 2/30 Short ntroducton & contents Cost-effcent deployment of dstrbuted software servces Cost functons Bo-nspred decentralzed

More information

Comparison of Heuristics for Scheduling Independent Tasks on Heterogeneous Distributed Environments

Comparison of Heuristics for Scheduling Independent Tasks on Heterogeneous Distributed Environments Comparson of Heurstcs for Schedulng Independent Tasks on Heterogeneous Dstrbuted Envronments Hesam Izakan¹, Ath Abraham², Senor Member, IEEE, Václav Snášel³ ¹ Islamc Azad Unversty, Ramsar Branch, Ramsar,

More information

Video Proxy System for a Large-scale VOD System (DINA)

Video Proxy System for a Large-scale VOD System (DINA) Vdeo Proxy System for a Large-scale VOD System (DINA) KWUN-CHUNG CHAN #, KWOK-WAI CHEUNG *# #Department of Informaton Engneerng *Centre of Innovaton and Technology The Chnese Unversty of Hong Kong SHATIN,

More information

Positive Semi-definite Programming Localization in Wireless Sensor Networks

Positive Semi-definite Programming Localization in Wireless Sensor Networks Postve Sem-defnte Programmng Localzaton n Wreless Sensor etworks Shengdong Xe 1,, Jn Wang, Aqun Hu 1, Yunl Gu, Jang Xu, 1 School of Informaton Scence and Engneerng, Southeast Unversty, 10096, anjng Computer

More information

Scheduling Remote Access to Scientific Instruments in Cyberinfrastructure for Education and Research

Scheduling Remote Access to Scientific Instruments in Cyberinfrastructure for Education and Research Schedulng Remote Access to Scentfc Instruments n Cybernfrastructure for Educaton and Research Je Yn 1, Junwe Cao 2,3,*, Yuexuan Wang 4, Lanchen Lu 1,3 and Cheng Wu 1,3 1 Natonal CIMS Engneerng and Research

More information

A Time-driven Data Placement Strategy for a Scientific Workflow Combining Edge Computing and Cloud Computing

A Time-driven Data Placement Strategy for a Scientific Workflow Combining Edge Computing and Cloud Computing > REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < 1 A Tme-drven Data Placement Strategy for a Scentfc Workflow Combnng Edge Computng and Cloud Computng Bng Ln, Fangnng

More information

Sum of Linear and Fractional Multiobjective Programming Problem under Fuzzy Rules Constraints

Sum of Linear and Fractional Multiobjective Programming Problem under Fuzzy Rules Constraints Australan Journal of Basc and Appled Scences, 2(4): 1204-1208, 2008 ISSN 1991-8178 Sum of Lnear and Fractonal Multobjectve Programmng Problem under Fuzzy Rules Constrants 1 2 Sanjay Jan and Kalash Lachhwan

More information

A GENETIC ALGORITHM FOR PROCESS SCHEDULING IN DISTRIBUTED OPERATING SYSTEMS CONSIDERING LOAD BALANCING

A GENETIC ALGORITHM FOR PROCESS SCHEDULING IN DISTRIBUTED OPERATING SYSTEMS CONSIDERING LOAD BALANCING A GENETIC ALGORITHM FOR PROCESS SCHEDULING IN DISTRIBUTED OPERATING SYSTEMS CONSIDERING LOAD BALANCING M. Nkravan and M. H. Kashan Department of Electrcal Computer Islamc Azad Unversty, Shahrar Shahreqods

More information

AADL : about scheduling analysis

AADL : about scheduling analysis AADL : about schedulng analyss Schedulng analyss, what s t? Embedded real-tme crtcal systems have temporal constrants to meet (e.g. deadlne). Many systems are bult wth operatng systems provdng multtaskng

More information

Real-time Fault-tolerant Scheduling Algorithm for Distributed Computing Systems

Real-time Fault-tolerant Scheduling Algorithm for Distributed Computing Systems Real-tme Fault-tolerant Schedulng Algorthm for Dstrbuted Computng Systems Yun Lng, Y Ouyang College of Computer Scence and Informaton Engneerng Zheang Gongshang Unversty Postal code: 310018 P.R.CHINA {ylng,

More information

Application of VCG in Replica Placement Strategy of Cloud Storage

Application of VCG in Replica Placement Strategy of Cloud Storage Internatonal Journal of Grd and Dstrbuted Computng, pp.27-40 http://dx.do.org/10.14257/jgdc.2016.9.4.03 Applcaton of VCG n Replca Placement Strategy of Cloud Storage Wang Hongxa Computer Department, Bejng

More information

Simulation Based Analysis of FAST TCP using OMNET++

Simulation Based Analysis of FAST TCP using OMNET++ Smulaton Based Analyss of FAST TCP usng OMNET++ Umar ul Hassan 04030038@lums.edu.pk Md Term Report CS678 Topcs n Internet Research Sprng, 2006 Introducton Internet traffc s doublng roughly every 3 months

More information

Problem Definitions and Evaluation Criteria for Computational Expensive Optimization

Problem Definitions and Evaluation Criteria for Computational Expensive Optimization Problem efntons and Evaluaton Crtera for Computatonal Expensve Optmzaton B. Lu 1, Q. Chen and Q. Zhang 3, J. J. Lang 4, P. N. Suganthan, B. Y. Qu 6 1 epartment of Computng, Glyndwr Unversty, UK Faclty

More information

Determining the Optimal Bandwidth Based on Multi-criterion Fusion

Determining the Optimal Bandwidth Based on Multi-criterion Fusion Proceedngs of 01 4th Internatonal Conference on Machne Learnng and Computng IPCSIT vol. 5 (01) (01) IACSIT Press, Sngapore Determnng the Optmal Bandwdth Based on Mult-crteron Fuson Ha-L Lang 1+, Xan-Mn

More information

An Iterative Solution Approach to Process Plant Layout using Mixed Integer Optimisation

An Iterative Solution Approach to Process Plant Layout using Mixed Integer Optimisation 17 th European Symposum on Computer Aded Process Engneerng ESCAPE17 V. Plesu and P.S. Agach (Edtors) 2007 Elsever B.V. All rghts reserved. 1 An Iteratve Soluton Approach to Process Plant Layout usng Mxed

More information

Maximum Variance Combined with Adaptive Genetic Algorithm for Infrared Image Segmentation

Maximum Variance Combined with Adaptive Genetic Algorithm for Infrared Image Segmentation Internatonal Conference on Logstcs Engneerng, Management and Computer Scence (LEMCS 5) Maxmum Varance Combned wth Adaptve Genetc Algorthm for Infrared Image Segmentaton Huxuan Fu College of Automaton Harbn

More information

Learning the Kernel Parameters in Kernel Minimum Distance Classifier

Learning the Kernel Parameters in Kernel Minimum Distance Classifier Learnng the Kernel Parameters n Kernel Mnmum Dstance Classfer Daoqang Zhang 1,, Songcan Chen and Zh-Hua Zhou 1* 1 Natonal Laboratory for Novel Software Technology Nanjng Unversty, Nanjng 193, Chna Department

More information

Helsinki University Of Technology, Systems Analysis Laboratory Mat Independent research projects in applied mathematics (3 cr)

Helsinki University Of Technology, Systems Analysis Laboratory Mat Independent research projects in applied mathematics (3 cr) Helsnk Unversty Of Technology, Systems Analyss Laboratory Mat-2.08 Independent research projects n appled mathematcs (3 cr) "! #$&% Antt Laukkanen 506 R ajlaukka@cc.hut.f 2 Introducton...3 2 Multattrbute

More information

Querying by sketch geographical databases. Yu Han 1, a *

Querying by sketch geographical databases. Yu Han 1, a * 4th Internatonal Conference on Sensors, Measurement and Intellgent Materals (ICSMIM 2015) Queryng by sketch geographcal databases Yu Han 1, a * 1 Department of Basc Courses, Shenyang Insttute of Artllery,

More information

Adaptive Energy and Location Aware Routing in Wireless Sensor Network

Adaptive Energy and Location Aware Routing in Wireless Sensor Network Adaptve Energy and Locaton Aware Routng n Wreless Sensor Network Hong Fu 1,1, Xaomng Wang 1, Yngshu L 1 Department of Computer Scence, Shaanx Normal Unversty, X an, Chna, 71006 fuhong433@gmal.com {wangxmsnnu@hotmal.cn}

More information

Network Coding as a Dynamical System

Network Coding as a Dynamical System Network Codng as a Dynamcal System Narayan B. Mandayam IEEE Dstngushed Lecture (jont work wth Dan Zhang and a Su) Department of Electrcal and Computer Engneerng Rutgers Unversty Outlne. Introducton 2.

More information

For instance, ; the five basic number-sets are increasingly more n A B & B A A = B (1)

For instance, ; the five basic number-sets are increasingly more n A B & B A A = B (1) Secton 1.2 Subsets and the Boolean operatons on sets If every element of the set A s an element of the set B, we say that A s a subset of B, or that A s contaned n B, or that B contans A, and we wrte A

More information

Cloud testing scheduling based on improved ACO

Cloud testing scheduling based on improved ACO Internatonal Symposum on Computers & Informatcs (ISCI 2015) Cloud testng schedulng based on mproved ACO Yang Zheng 1,2 a, Lzh Ca *2,3 b, Shdong Huang 4,c, Jawen Lu 1,d and Pan Lu 5,e 1 College of Informaton

More information

An Indian Journal FULL PAPER ABSTRACT KEYWORDS. Trade Science Inc.

An Indian Journal FULL PAPER ABSTRACT KEYWORDS. Trade Science Inc. [Type text] [Type text] [Type text] ISSN : 97-735 Volume Issue 9 BoTechnology An Indan Journal FULL PAPER BTAIJ, (9), [333-3] Matlab mult-dmensonal model-based - 3 Chnese football assocaton super league

More information

A New Approach For the Ranking of Fuzzy Sets With Different Heights

A New Approach For the Ranking of Fuzzy Sets With Different Heights New pproach For the ankng of Fuzzy Sets Wth Dfferent Heghts Pushpnder Sngh School of Mathematcs Computer pplcatons Thapar Unversty, Patala-7 00 Inda pushpndersnl@gmalcom STCT ankng of fuzzy sets plays

More information

An Application of the Dulmage-Mendelsohn Decomposition to Sparse Null Space Bases of Full Row Rank Matrices

An Application of the Dulmage-Mendelsohn Decomposition to Sparse Null Space Bases of Full Row Rank Matrices Internatonal Mathematcal Forum, Vol 7, 2012, no 52, 2549-2554 An Applcaton of the Dulmage-Mendelsohn Decomposton to Sparse Null Space Bases of Full Row Rank Matrces Mostafa Khorramzadeh Department of Mathematcal

More information

Compiler Design. Spring Register Allocation. Sample Exercises and Solutions. Prof. Pedro C. Diniz

Compiler Design. Spring Register Allocation. Sample Exercises and Solutions. Prof. Pedro C. Diniz Compler Desgn Sprng 2014 Regster Allocaton Sample Exercses and Solutons Prof. Pedro C. Dnz USC / Informaton Scences Insttute 4676 Admralty Way, Sute 1001 Marna del Rey, Calforna 90292 pedro@s.edu Regster

More information

A MOVING MESH APPROACH FOR SIMULATION BUDGET ALLOCATION ON CONTINUOUS DOMAINS

A MOVING MESH APPROACH FOR SIMULATION BUDGET ALLOCATION ON CONTINUOUS DOMAINS Proceedngs of the Wnter Smulaton Conference M E Kuhl, N M Steger, F B Armstrong, and J A Jones, eds A MOVING MESH APPROACH FOR SIMULATION BUDGET ALLOCATION ON CONTINUOUS DOMAINS Mark W Brantley Chun-Hung

More information

Module Management Tool in Software Development Organizations

Module Management Tool in Software Development Organizations Journal of Computer Scence (5): 8-, 7 ISSN 59-66 7 Scence Publcatons Management Tool n Software Development Organzatons Ahmad A. Al-Rababah and Mohammad A. Al-Rababah Faculty of IT, Al-Ahlyyah Amman Unversty,

More information

Constructing Minimum Connected Dominating Set: Algorithmic approach

Constructing Minimum Connected Dominating Set: Algorithmic approach Constructng Mnmum Connected Domnatng Set: Algorthmc approach G.N. Puroht and Usha Sharma Centre for Mathematcal Scences, Banasthal Unversty, Rajasthan 304022 usha.sharma94@yahoo.com Abstract: Connected

More information

Subspace clustering. Clustering. Fundamental to all clustering techniques is the choice of distance measure between data points;

Subspace clustering. Clustering. Fundamental to all clustering techniques is the choice of distance measure between data points; Subspace clusterng Clusterng Fundamental to all clusterng technques s the choce of dstance measure between data ponts; D q ( ) ( ) 2 x x = x x, j k = 1 k jk Squared Eucldean dstance Assumpton: All features

More information

Classifier Selection Based on Data Complexity Measures *

Classifier Selection Based on Data Complexity Measures * Classfer Selecton Based on Data Complexty Measures * Edth Hernández-Reyes, J.A. Carrasco-Ochoa, and J.Fco. Martínez-Trndad Natonal Insttute for Astrophyscs, Optcs and Electroncs, Lus Enrque Erro No.1 Sta.

More information

MODULE DESIGN BASED ON INTERFACE INTEGRATION TO MAXIMIZE PRODUCT VARIETY AND MINIMIZE FAMILY COST

MODULE DESIGN BASED ON INTERFACE INTEGRATION TO MAXIMIZE PRODUCT VARIETY AND MINIMIZE FAMILY COST INTERNATIONAL CONFERENCE ON ENGINEERING DESIGN, ICED 07 28-31 AUGUST 2007, CITE DES SCIENCES ET DE L'INDUSTRIE, PARIS, FRANCE MODULE DESIGN BASED ON INTERFACE INTEGRATION TO MAIMIZE PRODUCT VARIETY AND

More information

Type-2 Fuzzy Non-uniform Rational B-spline Model with Type-2 Fuzzy Data

Type-2 Fuzzy Non-uniform Rational B-spline Model with Type-2 Fuzzy Data Malaysan Journal of Mathematcal Scences 11(S) Aprl : 35 46 (2017) Specal Issue: The 2nd Internatonal Conference and Workshop on Mathematcal Analyss (ICWOMA 2016) MALAYSIAN JOURNAL OF MATHEMATICAL SCIENCES

More information

A Clustering Algorithm for Chinese Adjectives and Nouns 1

A Clustering Algorithm for Chinese Adjectives and Nouns 1 Clusterng lgorthm for Chnese dectves and ouns Yang Wen, Chunfa Yuan, Changnng Huang 2 State Key aboratory of Intellgent Technology and System Deptartment of Computer Scence & Technology, Tsnghua Unversty,

More information

Wishing you all a Total Quality New Year!

Wishing you all a Total Quality New Year! Total Qualty Management and Sx Sgma Post Graduate Program 214-15 Sesson 4 Vnay Kumar Kalakband Assstant Professor Operatons & Systems Area 1 Wshng you all a Total Qualty New Year! Hope you acheve Sx sgma

More information

Reliable and Efficient Routing Using Adaptive Genetic Algorithm in Packet Switched Networks

Reliable and Efficient Routing Using Adaptive Genetic Algorithm in Packet Switched Networks IJCSI Internatonal Journal of Computer Scence Issues, Vol. 9, Issue 1, No 3, January 2012 ISSN (Onlne): 1694-0814 www.ijcsi.org 168 Relable and Effcent Routng Usng Adaptve Genetc Algorthm n Packet Swtched

More information

Parallel matrix-vector multiplication

Parallel matrix-vector multiplication Appendx A Parallel matrx-vector multplcaton The reduced transton matrx of the three-dmensonal cage model for gel electrophoress, descrbed n secton 3.2, becomes excessvely large for polymer lengths more

More information

Distributed Resource Scheduling in Grid Computing Using Fuzzy Approach

Distributed Resource Scheduling in Grid Computing Using Fuzzy Approach Dstrbuted Resource Schedulng n Grd Computng Usng Fuzzy Approach Shahram Amn, Mohammad Ahmad Computer Engneerng Department Islamc Azad Unversty branch Mahallat, Iran Islamc Azad Unversty branch khomen,

More information

A Genetic Algorithm Based Dynamic Load Balancing Scheme for Heterogeneous Distributed Systems

A Genetic Algorithm Based Dynamic Load Balancing Scheme for Heterogeneous Distributed Systems Proceedngs of the Internatonal Conference on Parallel and Dstrbuted Processng Technques and Applcatons, PDPTA 2008, Las Vegas, Nevada, USA, July 14-17, 2008, 2 Volumes. CSREA Press 2008, ISBN 1-60132-084-1

More information

Tsinghua University at TAC 2009: Summarizing Multi-documents by Information Distance

Tsinghua University at TAC 2009: Summarizing Multi-documents by Information Distance Tsnghua Unversty at TAC 2009: Summarzng Mult-documents by Informaton Dstance Chong Long, Mnle Huang, Xaoyan Zhu State Key Laboratory of Intellgent Technology and Systems, Tsnghua Natonal Laboratory for

More information

Research of Dynamic Access to Cloud Database Based on Improved Pheromone Algorithm

Research of Dynamic Access to Cloud Database Based on Improved Pheromone Algorithm , pp.197-202 http://dx.do.org/10.14257/dta.2016.9.5.20 Research of Dynamc Access to Cloud Database Based on Improved Pheromone Algorthm Yongqang L 1 and Jn Pan 2 1 (Software Technology Vocatonal College,

More information

Solving two-person zero-sum game by Matlab

Solving two-person zero-sum game by Matlab Appled Mechancs and Materals Onlne: 2011-02-02 ISSN: 1662-7482, Vols. 50-51, pp 262-265 do:10.4028/www.scentfc.net/amm.50-51.262 2011 Trans Tech Publcatons, Swtzerland Solvng two-person zero-sum game by

More information

Research Article Adaptive Cost-Based Task Scheduling in Cloud Environment

Research Article Adaptive Cost-Based Task Scheduling in Cloud Environment Scentfc Programmng Volume 2016, Artcle ID 8239239, 9 pages http://dx.do.org/10.1155/2016/8239239 Research Artcle Adaptve Cost-Based Task Schedulng n Cloud Envronment Mohammed A. S. Mosleh, 1 G. Radhaman,

More information

An Efficient Genetic Algorithm with Fuzzy c-means Clustering for Traveling Salesman Problem

An Efficient Genetic Algorithm with Fuzzy c-means Clustering for Traveling Salesman Problem An Effcent Genetc Algorthm wth Fuzzy c-means Clusterng for Travelng Salesman Problem Jong-Won Yoon and Sung-Bae Cho Dept. of Computer Scence Yonse Unversty Seoul, Korea jwyoon@sclab.yonse.ac.r, sbcho@cs.yonse.ac.r

More information

Content Based Image Retrieval Using 2-D Discrete Wavelet with Texture Feature with Different Classifiers

Content Based Image Retrieval Using 2-D Discrete Wavelet with Texture Feature with Different Classifiers IOSR Journal of Electroncs and Communcaton Engneerng (IOSR-JECE) e-issn: 78-834,p- ISSN: 78-8735.Volume 9, Issue, Ver. IV (Mar - Apr. 04), PP 0-07 Content Based Image Retreval Usng -D Dscrete Wavelet wth

More information

Support Vector Machines

Support Vector Machines /9/207 MIST.6060 Busness Intellgence and Data Mnng What are Support Vector Machnes? Support Vector Machnes Support Vector Machnes (SVMs) are supervsed learnng technques that analyze data and recognze patterns.

More information

Study on Multi-objective Flexible Job-shop Scheduling Problem considering Energy Consumption

Study on Multi-objective Flexible Job-shop Scheduling Problem considering Energy Consumption Journal of Industral Engneerng and Management JIEM, 2014 7(3): 589-604 nlne ISSN: 2014-0953 Prnt ISSN: 2014-8423 http://dx.do.org/10.3926/jem.1075 Study on Mult-objectve Flexble Job-shop Schedulng Problem

More information

Improved Resource Allocation Algorithms for Practical Image Encoding in a Ubiquitous Computing Environment

Improved Resource Allocation Algorithms for Practical Image Encoding in a Ubiquitous Computing Environment JOURNAL OF COMPUTERS, VOL. 4, NO. 9, SEPTEMBER 2009 873 Improved Resource Allocaton Algorthms for Practcal Image Encodng n a Ubqutous Computng Envronment Manxong Dong, Long Zheng, Kaoru Ota, Song Guo School

More information

RAP. Speed/RAP/CODA. Real-time Systems. Modeling the sensor networks. Real-time Systems. Modeling the sensor networks. Real-time systems:

RAP. Speed/RAP/CODA. Real-time Systems. Modeling the sensor networks. Real-time Systems. Modeling the sensor networks. Real-time systems: Speed/RAP/CODA Presented by Octav Chpara Real-tme Systems Many wreless sensor network applcatons requre real-tme support Survellance and trackng Border patrol Fre fghtng Real-tme systems: Hard real-tme:

More information

DESIGNING TRANSMISSION SCHEDULES FOR WIRELESS AD HOC NETWORKS TO MAXIMIZE NETWORK THROUGHPUT

DESIGNING TRANSMISSION SCHEDULES FOR WIRELESS AD HOC NETWORKS TO MAXIMIZE NETWORK THROUGHPUT DESIGNING TRANSMISSION SCHEDULES FOR WIRELESS AD HOC NETWORKS TO MAXIMIZE NETWORK THROUGHPUT Bran J. Wolf, Joseph L. Hammond, and Harlan B. Russell Dept. of Electrcal and Computer Engneerng, Clemson Unversty,

More information

EVALUATION OF THE PERFORMANCES OF ARTIFICIAL BEE COLONY AND INVASIVE WEED OPTIMIZATION ALGORITHMS ON THE MODIFIED BENCHMARK FUNCTIONS

EVALUATION OF THE PERFORMANCES OF ARTIFICIAL BEE COLONY AND INVASIVE WEED OPTIMIZATION ALGORITHMS ON THE MODIFIED BENCHMARK FUNCTIONS Academc Research Internatonal ISS-L: 3-9553, ISS: 3-9944 Vol., o. 3, May 0 EVALUATIO OF THE PERFORMACES OF ARTIFICIAL BEE COLOY AD IVASIVE WEED OPTIMIZATIO ALGORITHMS O THE MODIFIED BECHMARK FUCTIOS Dlay

More information

Machine Learning: Algorithms and Applications

Machine Learning: Algorithms and Applications 14/05/1 Machne Learnng: Algorthms and Applcatons Florano Zn Free Unversty of Bozen-Bolzano Faculty of Computer Scence Academc Year 011-01 Lecture 10: 14 May 01 Unsupervsed Learnng cont Sldes courtesy of

More information

Classifying Acoustic Transient Signals Using Artificial Intelligence

Classifying Acoustic Transient Signals Using Artificial Intelligence Classfyng Acoustc Transent Sgnals Usng Artfcal Intellgence Steve Sutton, Unversty of North Carolna At Wlmngton (suttons@charter.net) Greg Huff, Unversty of North Carolna At Wlmngton (jgh7476@uncwl.edu)

More information

An Improved Image Segmentation Algorithm Based on the Otsu Method

An Improved Image Segmentation Algorithm Based on the Otsu Method 3th ACIS Internatonal Conference on Software Engneerng, Artfcal Intellgence, Networkng arallel/dstrbuted Computng An Improved Image Segmentaton Algorthm Based on the Otsu Method Mengxng Huang, enjao Yu,

More information

Priority-Based Scheduling Algorithm for Downlink Traffics in IEEE Networks

Priority-Based Scheduling Algorithm for Downlink Traffics in IEEE Networks Prorty-Based Schedulng Algorthm for Downlnk Traffcs n IEEE 80.6 Networks Ja-Mng Lang, Jen-Jee Chen, You-Chun Wang, Yu-Chee Tseng, and Bao-Shuh P. Ln Department of Computer Scence Natonal Chao-Tung Unversty,

More information

Performance Evaluation of Information Retrieval Systems

Performance Evaluation of Information Retrieval Systems Why System Evaluaton? Performance Evaluaton of Informaton Retreval Systems Many sldes n ths secton are adapted from Prof. Joydeep Ghosh (UT ECE) who n turn adapted them from Prof. Dk Lee (Unv. of Scence

More information

ELEC 377 Operating Systems. Week 6 Class 3

ELEC 377 Operating Systems. Week 6 Class 3 ELEC 377 Operatng Systems Week 6 Class 3 Last Class Memory Management Memory Pagng Pagng Structure ELEC 377 Operatng Systems Today Pagng Szes Vrtual Memory Concept Demand Pagng ELEC 377 Operatng Systems

More information

The Research of Support Vector Machine in Agricultural Data Classification

The Research of Support Vector Machine in Agricultural Data Classification The Research of Support Vector Machne n Agrcultural Data Classfcaton Le Sh, Qguo Duan, Xnmng Ma, Me Weng College of Informaton and Management Scence, HeNan Agrcultural Unversty, Zhengzhou 45000 Chna Zhengzhou

More information

Delay Variation Optimized Traffic Allocation Based on Network Calculus for Multi-path Routing in Wireless Mesh Networks

Delay Variation Optimized Traffic Allocation Based on Network Calculus for Multi-path Routing in Wireless Mesh Networks Appl. Math. Inf. Sc. 7, No. 2L, 467-474 2013) 467 Appled Mathematcs & Informaton Scences An Internatonal Journal http://dx.do.org/10.12785/ams/072l13 Delay Varaton Optmzed Traffc Allocaton Based on Network

More information

An Efficient Genetic Algorithm Based Approach for the Minimum Graph Bisection Problem

An Efficient Genetic Algorithm Based Approach for the Minimum Graph Bisection Problem 118 An Effcent Genetc Algorthm Based Approach for the Mnmum Graph Bsecton Problem Zh-Qang Chen, Rong-Long WAG and Kozo OKAZAKI Faculty of Engneerng, Unversty of Fuku, Bunkyo 3-9-1,Fuku-sh, Japan 910-8507

More information

An Entropy-Based Approach to Integrated Information Needs Assessment

An Entropy-Based Approach to Integrated Information Needs Assessment Dstrbuton Statement A: Approved for publc release; dstrbuton s unlmted. An Entropy-Based Approach to ntegrated nformaton Needs Assessment June 8, 2004 Wllam J. Farrell Lockheed Martn Advanced Technology

More information

Recommended Items Rating Prediction based on RBF Neural Network Optimized by PSO Algorithm

Recommended Items Rating Prediction based on RBF Neural Network Optimized by PSO Algorithm Recommended Items Ratng Predcton based on RBF Neural Network Optmzed by PSO Algorthm Chengfang Tan, Cayn Wang, Yuln L and Xx Q Abstract In order to mtgate the data sparsty and cold-start problems of recommendaton

More information

The Comparison of Calibration Method of Binocular Stereo Vision System Ke Zhang a *, Zhao Gao b

The Comparison of Calibration Method of Binocular Stereo Vision System Ke Zhang a *, Zhao Gao b 3rd Internatonal Conference on Materal, Mechancal and Manufacturng Engneerng (IC3ME 2015) The Comparson of Calbraton Method of Bnocular Stereo Vson System Ke Zhang a *, Zhao Gao b College of Engneerng,

More information

Outline. Type of Machine Learning. Examples of Application. Unsupervised Learning

Outline. Type of Machine Learning. Examples of Application. Unsupervised Learning Outlne Artfcal Intellgence and ts applcatons Lecture 8 Unsupervsed Learnng Professor Danel Yeung danyeung@eee.org Dr. Patrck Chan patrckchan@eee.org South Chna Unversty of Technology, Chna Introducton

More information

Study of Data Stream Clustering Based on Bio-inspired Model

Study of Data Stream Clustering Based on Bio-inspired Model , pp.412-418 http://dx.do.org/10.14257/astl.2014.53.86 Study of Data Stream lusterng Based on Bo-nspred Model Yngme L, Mn L, Jngbo Shao, Gaoyang Wang ollege of omputer Scence and Informaton Engneerng,

More information

Clustering Algorithm Combining CPSO with K-Means Chunqin Gu 1, a, Qian Tao 2, b

Clustering Algorithm Combining CPSO with K-Means Chunqin Gu 1, a, Qian Tao 2, b Internatonal Conference on Advances n Mechancal Engneerng and Industral Informatcs (AMEII 05) Clusterng Algorthm Combnng CPSO wth K-Means Chunqn Gu, a, Qan Tao, b Department of Informaton Scence, Zhongka

More information

The Greedy Method. Outline and Reading. Change Money Problem. Greedy Algorithms. Applications of the Greedy Strategy. The Greedy Method Technique

The Greedy Method. Outline and Reading. Change Money Problem. Greedy Algorithms. Applications of the Greedy Strategy. The Greedy Method Technique //00 :0 AM Outlne and Readng The Greedy Method The Greedy Method Technque (secton.) Fractonal Knapsack Problem (secton..) Task Schedulng (secton..) Mnmum Spannng Trees (secton.) Change Money Problem Greedy

More information

GA-Based Learning Algorithms to Identify Fuzzy Rules for Fuzzy Neural Networks

GA-Based Learning Algorithms to Identify Fuzzy Rules for Fuzzy Neural Networks Seventh Internatonal Conference on Intellgent Systems Desgn and Applcatons GA-Based Learnng Algorthms to Identfy Fuzzy Rules for Fuzzy Neural Networks K Almejall, K Dahal, Member IEEE, and A Hossan, Member

More information

A Semi-Distributed Load Balancing Architecture and Algorithm for Heterogeneous Wireless Networks

A Semi-Distributed Load Balancing Architecture and Algorithm for Heterogeneous Wireless Networks A Sem-Dstrbuted oad Balancng Archtecture and Algorthm for Heterogeneous reless Networks Md. Golam Rabul Ala Choong Seon Hong * Kyung Hee Unversty, Korea rob@networkng.khu.ac.kr, cshong@khu.ac.kr Abstract

More information

NGPM -- A NSGA-II Program in Matlab

NGPM -- A NSGA-II Program in Matlab Verson 1.4 LIN Song Aerospace Structural Dynamcs Research Laboratory College of Astronautcs, Northwestern Polytechncal Unversty, Chna Emal: lsssswc@163.com 2011-07-26 Contents Contents... 1. Introducton...

More information

Network Intrusion Detection Based on PSO-SVM

Network Intrusion Detection Based on PSO-SVM TELKOMNIKA Indonesan Journal of Electrcal Engneerng Vol.1, No., February 014, pp. 150 ~ 1508 DOI: http://dx.do.org/10.11591/telkomnka.v1.386 150 Network Intruson Detecton Based on PSO-SVM Changsheng Xang*

More information

AN IMPROVED GENETIC ALGORITHM FOR RECTANGLES CUTTING & PACKING PROBLEM. Wang Shoukun, Wang Jingchun, Jin Yihui

AN IMPROVED GENETIC ALGORITHM FOR RECTANGLES CUTTING & PACKING PROBLEM. Wang Shoukun, Wang Jingchun, Jin Yihui Copyrght 2002 IFAC 5th Trennal World Congress, Barcelona, Span A IPROVED GEETIC ALGORITH FOR RECTAGLES CUTTIG & PACKIG PROBLE Wang Shouun, Wang Jngchun, Jn Yhu Tsnghua Unversty, Beng 00084, P. R. Chna

More information

A Resources Virtualization Approach Supporting Uniform Access to Heterogeneous Grid Resources 1

A Resources Virtualization Approach Supporting Uniform Access to Heterogeneous Grid Resources 1 A Resources Vrtualzaton Approach Supportng Unform Access to Heterogeneous Grd Resources 1 Cunhao Fang 1, Yaoxue Zhang 2, Song Cao 3 1 Tsnghua Natonal Labatory of Inforamaton Scence and Technology 2 Department

More information

A Binarization Algorithm specialized on Document Images and Photos

A Binarization Algorithm specialized on Document Images and Photos A Bnarzaton Algorthm specalzed on Document mages and Photos Ergna Kavalleratou Dept. of nformaton and Communcaton Systems Engneerng Unversty of the Aegean kavalleratou@aegean.gr Abstract n ths paper, a

More information

RESEARCH ON JOB-SHOP SCHEDULING PROBLEM BASED ON IMPROVED PARTICLE SWARM OPTIMIZATION

RESEARCH ON JOB-SHOP SCHEDULING PROBLEM BASED ON IMPROVED PARTICLE SWARM OPTIMIZATION Journal of heoretcal and Appled Informaton echnology 005-013 JAI & LLS. All rghts reserved. RESEARCH ON JOB-SHOP SCHEDULING PROBLEM BASED ON IMPROVED PARICLE SWARM OPIMIZAION 1 ZUFENG ZHONG 1 School of

More information

A Hybrid Genetic Algorithm for Routing Optimization in IP Networks Utilizing Bandwidth and Delay Metrics

A Hybrid Genetic Algorithm for Routing Optimization in IP Networks Utilizing Bandwidth and Delay Metrics A Hybrd Genetc Algorthm for Routng Optmzaton n IP Networks Utlzng Bandwdth and Delay Metrcs Anton Redl Insttute of Communcaton Networks, Munch Unversty of Technology, Arcsstr. 21, 80290 Munch, Germany

More information