Robust Capacitated Facility Location Problem: Optimization Model and Solution Algorithms

Size: px
Start display at page:

Download "Robust Capacitated Facility Location Problem: Optimization Model and Solution Algorithms"

Transcription

1 Journal of Uncertan Sytem Vol.7, No.1, pp.22-35, 2013 Onlne at: Robut Capactated Faclty Locaton Problem: Optmzaton Model and Soluton Algorthm Ragheb Rahmanan *, Mohammad Sad-Mehrabad, Hoat Ahour Department of Indutral Engneerng, Iran Unverty of Scence and Technology, P.C , Narmak, Tehran, Iran Receved 6 July 2011; Reved 13 November 2012 Abtract In th artcle, we propoe an extenon of the capactated faclty locaton problem under uncertanty, where uncertanty may appear n the model key parameter uch a demand and cot. In th model, t aumed that faclte have hard contrant on the amount of demand they can erve and, a a reult, ome cutomer may not be fully atfed. Unfortunately, tradtonal model gnore th tuaton and f faclte do not erve all demand, the model become nfeable. Accordngly, we develop the mathematcal formulaton n order to allow partal atfacton by ntroducng penalty cot for unatfed demand. In general, th model optmze locaton for predefned number of capactated faclte n uch a way that mnmze total expected cot of tranportaton, contructon, and penalty cot of uncovered demand, whle relatve regret n each cenaro mut be no greater than a potve number ( p 0 ). The developed model NP-hard and very challengng to olve. Therefore, an effcent heurtc oluton algorthm baed on the varable neghborhood earch developed to olve the problem. The algorthm effcency compared wth the mulated annealng algorthm and CPLEX olver by olvng varety of tet problem.computatonal experment how that the propoed algorthm more effectve and effcent n term of CPU tme and oluton qualty World Academc Pre, UK. All rght reerved. Keyword: capactated faclty locaton problem, uncertanty, robut optmzaton, partal atfacton, varable neghborhood earch, mulated annealng 1 Introducton Locaton problem are crtcal manageral decon wth a large body of lterature and numerou applcaton n the real-world applcaton [5, 9, 10, 24, 26]. In fact, faclty locaton decon are long-term trategcal decon and almot mpoble to revere. Thee trategc decon have a great mpact on the network flow and cutomer atfacton. Eentally, thee problem deal wth locatng a number of faclte to upply a et of cutomer at the mnmum cot wth repect to everal contrant and nvolve varou obectve whch naturally are n conflct [27]. For example, n ome locaton problem there mght be hard contrant on amount of avalable budget, number of faclte, faclte capacty, coverage dtance, and o forth, whle the obectve functon maxmze total covered demand and mnmze total contructon and tranportaton cot. In lne wth th ubect, numerou paper have been publhed n whch the Capactated Faclty Locaton Problem (CFLP) one of the mot mportant and bac model. Th problem ha been ntenvely condered n the lterature, nce t nclude more realtc aumpton n compare wth, for example, uncapactated locaton problem. In the CFLP, a number of capactated faclte, P, are to be located among J poble te n order to atfy demand of I cutomer by mnmzng total cot of tranportaton and fxed charge of etablhng faclte. The CFLP NP-hard problem whch generalzed from the mple plant locaton problem. The CFLP ha been effectvely mplemented to olve real-world applcaton uch a plant locaton, power taton locaton, warehoue locaton, to ut name a few. Though th problem and t varaton have been ntenvely tuded n the lterature, th problem ha been manly nvetgated wth determntc data and wth th aumpton that faclte capacty can meet all cutomer demand. * Correpondng author. Emal: ragheb.rahmanan@gmal.com (R. Rahmanan).

2 Journal of Uncertan Sytem, Vol.7, No.1, pp.22-35, However, n the realty, thee aumpton are rarely atfed. Utlzed data n manageral decon are encountered wthn complete factor uch a noy and erroneou defect. Meanwhle, n many cae due to lack of htorcal nformaton, t almot mpoble to obtan a relable and certan etmaton for the model key parameter. On the other hand, locaton decon are long-term trategc decon and very entve to the change n demand and other parameter. Therefore, gnorng uncertanty n parameter yeld n neffectve decon. A a reult, optmzaton under uncertanty ha receved ncreang attenton n the locaton theory durng the lat few decade [2, 13, 32, 30, 35] and, varou approache are propoed to face and reolve the uncertanty. Thee approache can be generally categorzed nto robut and tochatc optmzaton approache. In the robut approach, the uncertan parameter are etmated by dcrete or nterval data and the obectve functon typcally mnmze mnmax cot or mnmax regret. The regret n each cenaro defned a dtance between the current cot of that cenaro and t optmal cot. Dcrete cenaro are ued when probablty dtrbuton for uncertan parameter not known. Robut oluton are naturally too conervatve decon and mpoe unneceary long-term cot. On the other hand, tochatc programmng model partton decon n twotage n whch obectve functon mnmze the um cot of the frt-tage and the expected cot of the econdtage. In general, the typcal tochatc programmng may yeld nexpenve oluton n the long run but perform poorly under certan realzaton of the random data. Bede, the robut problem due to the mnmax tructure are much harder to olve. Baed on thee weaknee, Snyder and Dakn [32] ntroduced the tochatcp-robut optmzaton approach whch come through combnng man concept of thee approache. One of the man goal of th approach to degn a more robut ytem wth a lttle ncreae n the expected cot. In th tudy, th approach appled to optmze the CFLP under uncertanty. Conderng uncertanty ncreae relablty of oluton. Moreover, tradtonal CFLP model become nfeable f total opened capacty lower than um of cutomer demand. Wherea, n the realty, t lkely that faclte due to ther lmted capacty leave ome of thee demand unatfed. For ntance, power taton have hard contrant on the amount of generated electrcty and they cannot provde unbounded amount of electrcty to the cutomer. Moreover, due to the huge nvetment cot of etablhng power taton, manager cannot contruct more than a lmted number of power taton. Hence, wth a lmted number of capactated power taton, logcally, we wll not be able to upply all cutomer ncreang demand. Th tuaton can be oberved n thrd-world and deprved area by far. Unfortunately, tradtonal model of CFLP cannot fnd even a feable oluton for uch problem. Therefore, we alo develop the mathematcal formulaton of CFLP wth th aumpton that our faclte may not be able to upply all demand and any proporton of cutomer demand may reman unatfed. On the other hand, long any proporton of demand mpoe a great deal of cot to the company, nce they loe the market hare. A a reult, the obectve functon mnmze penalty cot of unatfed demand, a well. Generally peakng, the man goal of th tudy to develop the mathematcal formulaton of CFLP under uncertanty by takng thee fact nto account that each cutomer may not be fully uppled and ther demand are uncertan and characterzed by a gven et of cenaro. Moreover, we aocate a penalty cot to any proporton of unatfed demand. Therefore, the obectve functon mnmze total cot of etablhng new faclte, expected cot of tranportaton, and expected penalty cot of unatfed demand. The developed model wll be a two-tage model n whch the frt-tage optmze faclte locaton and the econd-tage optmze cutomer agnment. Each cutomer may be uppled by everal faclte (.e. multple agnment property) and any proporton of t demand may be atfed (.e. ay 60% of t total demand). Th formulzaton wll be referred to u a the Stochatc p-robut Capactated Faclty Locaton Problem, or p-scflp a an abbrevaton. In addton, nce the developed model formulaton hard and cannot be effectvely olved ung ordnary optmzaton method, we propoe oluton algorthm baed on Varable Neghborhood Search (VNS) and Smulated Annealng (SA) to olve the developed model. We tet the effcency of thee algorthm on a varety tet ntance and compare ther reult wth the CPLEX olver n term of optmalty gap and CPU tme requrement. The remnder of th paper organzed a follow: n the next Secton the related lterature revewed. In Secton 3 the mathematcal formulaton of the p-scflp preented. Subequently, oluton approache baed on the VNS and SA are outlned n Secton 4. Lkewe, our numercal experment are ummarzed n the Secton 5 and fnally, our concluon and future reearch drecton are dcued n the Secton 6. 2 Related Lterature The CFLP a clacal faclty locaton problem wth large body of the lterature. The CFLP ha been extenvely nvetgate don the both de of model formulaton and olvng algorthm. Varou algorthm that have been mplemented to olve the Uncapactated Faclty Locaton Problem (UFLP) are uually generalzed to olve the

3 24 R. Rahmanan et al.: Robut Capactated Faclty Locaton Problem CFLP. For example, Kuehn and Hamburger [22] propoed an algorthm to olve the UFLP and then Jacoben [17] extended th algorthm for the CFLP. However, n th ecton, we are gong to addre the mot relevant paper to the condered problem and the ntereted reader are referred to [1, 26, 33] and reference theren. Lagrangan Relaxaton (LR) ha been wdely condered a an effcent oluton algorthm to olve the CFLP. Context preented by Cornueol et al. [7] provd an excellent theoretcal analy of all poble Lagrangan relaxaton and the lnear programmng relaxaton for the CFLP. A Lagrangan Heurtc (LH) framework preented by the Bealey [4] to olve dfferent faclty locaton problem. In the propoed method for the CFLP, allocaton contrant and capacty contrant are ncorporated nto obectve functon by ung Lagrangan multpler. In addton, LH method for the both UFLP and the CFLP alo propoed by other author [3, 23]. Barahona and Chudak [3] ntally provded the lnear programmng relaxaton of the CFLP and then propoed the Lagrangan relaxaton to olve the lnear problem. Heurtc and meta-heurtc algorthm are alo appled to olve the CFLP n a wde range. Sun [34] appled the Tabu Search (TS) to olve the CFLP and compared t wth the Lagrangan and the urrogate/lagrangan heurtc method. He ued long term memory baed on prmogentary lnked quad tree to tore vted oluton and prohbt them from beng vted agan. Cortnhal and Captvo [8] propoed upper and lower bound for the ngle ource capactated locaton problem. In that artcle, Lagrangan relaxaton ued to obtan lower bound for th problem and upper bound are obtaned by Lagrangan heurtc followed by earch method (e.g. TS). A mple local earch heurtc for the capactated faclty locaton problem preented by Korupolu et al. [20]. Hnd and Penkoz [15] preented a heurtc that combne Lagrangan relaxaton wth retrcted neghborhood earch for the capactated ngle ource locaton problem. In th paper, a heurtc procedure wth three phae ued to fnd feable oluton. An effcent VNS heurtc for the capactated P-medan problem nvetgated by Flezar and Hnd [12]. They alo propoed everal ntermedate heurtc earch algorthm. Thee algorthm were able to fnd very good oluton wthn much horter computaton tme than the full VNS algorthm.smlarly, exact method have been nvetgated to olve the CFLP. Bacally, the propoed exact method are baed on branch and bound and et parttonng technque wth dfference n tratege to ncreae the lower bound and type of relaxaton. For ntance, Sa [28] relaxed the CFLP to obtan tranportaton problem but Nau [25] relaxed capacty contrant and added et of urrogate contrant to obtan tghter bound. Neverthele, ntereted reader are referred to the context by Srdharan [33] to get comprehenve revew on the oluton algorthm for the CFLP. However, the publhed work related to CFLP have motly concentrated on oluton algorthm, and extenon of mathematcal formulaton are uually ntroduced a newer work uch a capactated hub locaton, capactated maxmal coverng, capactated faclty locaton and network degn, and o forth. Faclty locaton problem are ntenvely tuded under uncertanty [2, 13, 16, 32, 30, 35]. For ntance, Ghezavat et al. [13] propoed a robut approach to locaton-allocaton problem under uncertanty. Snyder and Dakn [32] propoed a new approach for optmzng faclty locaton problem under uncertanty. They combned the mnmum-expected-cot and p-robutne meaure together n order to ntroduce the tochatc p-robut optmzaton model. Th approach mplemented on the clacal model of the UFLP and P-medan problem n order to reolve the uncertanty n demand and dtance. They ntended to fnd oluton that ha mnmum-expected-cot whle the obtaned oluton p-robut;.e., cot under each cenaro for each feable oluton mut be no greater than 100(1+p) % of the optmal cot of that cenaro, where p a non-negatve value known a robutne coeffcent. They olved the propoed model by ung Lagrangan decompoton and reducng them to the multple-choce knapack problem. Addtonally, they dcued a mechanm for detectng nfeablty. The nteret reader are referred to the [5, 21, 26, 29, 31] to get more nformaton about locaton and locaton under uncertanty. 3 Problem Decrpton and Formulaton Conder a producton-dtrbuton network that ha I maor center of cutomer and J potental faclty te. A company want to locate P faclte n th network n order to ervce thee cutomer. Thee faclte cannot upply unbounded amount of demand for cutomer, and conequently, ome cutomer may not be fully atfed or even may reman completely or partally unatfed. On the other hand, nablty n atfyng cutomer demand ncur huge cot for the company n conequence of long the market hare. A a reult, we defne penalty cot for unatfed demand a a functon of unmet-demand. Therefore, f X, denote fracton of demand of cutomer whch atfed by faclty at node under cenaro, we wll have 1 X, percentage of t demand unatfed n cenaro. Where, X, J J ndcate total fracton of demand of cutomer whch atfed by faclte through the network. Accordngly, we defne a varable n order to capture the proporton of demand of cutomer under

4 Journal of Uncertan Sytem, Vol.7, No.1, pp.22-35, cenaro whch unuppled,.e., Z. By keepng th n mnd that um of total fracton of atfed and unatfed demand for each cutomer n each cenaro hould be equal to one or, one hundred percent of t demand, we have equaton (1). X, Z 1 I, S. (1) J Logcally, each cutomer may be agned to everal faclte and may be partally uppled (ay 80 percentage of t total demand). Th more cloe to the realty becaue n the clacal model f faclte could not upply all demand, the model become nfeable, whle there tll a uboptmal oluton to the problem. To mplfy the preentaton of the mathematcal formulaton n th paper, the followng notaton are defned.the ndex et and model parameter are decrbed n Table 1. Table 1: Parameter and ndex et Symbol Indexed by Decrpton 1,2,..., J Set of potental faclty te; J I 1,2,..., I S 1,2,..., S Parameter C W f d * Q a Set of cotumer; Set of cenaro; Decrpton Faclte capacty Probablty of that cenaro occur Dered robutne coeffcent Demand at node under cenaro The fxed charge of openng a faclty on node Number of faclte to be open Travel dtance between node and under cenaro Optmal cot of CFLP under data from cenaro A contant number * where Q the optmal obectve value of the determntc CFLP problem wth partal atfacton (ee equaton (15-21)) under data from cenaro that can be computed by CPLEX or any exact method. The decon of the tochatc p-robut capactated faclty locaton nclude decon about locatng faclte and agnng cotumer to thee faclte n uch a way that total expected cot beng mnmzed. Thee decon are made n two tage and are defned a follow. 1 f one faclty located at node y 0 otherwe X fracton of demand of node under cenaro that atfed by faclty located at node, 0 Z 0 fracton of demand of node under cenaro that notatfed. Note that, here the locaton decon varable (y ), unlke the agnment varable ( X, and Z ), are ndependent on the ndex n order to reflect the two-tage nature of the problem. Fnally, we aume that uncertanty aocated wth demand and dtance. Ung th notaton and aumpton, the propoed two-tage mxed-nteger model for the problem n hand formulated a follow.

5 26 R. Rahmanan et al.: Robut Capactated Faclty Locaton Problem Mn SCFLP : f y W d X, a W Z J S I J I Subect to: (2) X, Z 1 I, S (3) J I W X C y J S,, f y W d X, a W Z (1 ) Q S (5) * J I J I J y P Z 0 I, S (7) X, 0 I, J, S (8) y 0,1 J. (9) In the propoed model, equaton (2) repreent the obectve functon. Frt term n th equaton mnmze the fxed cot of openng faclte. Second term mnmze total expected tranportaton cot. Fnally, the thrd term mnmze total expected penalty cot of unatfed demand. Penalty cot for each cutomer calculated a a functon of lot demand ( W * Z ) multple to a contant number. Here, we have aumed that there lnear relaton between lot demand and cot mpoed to the company, whle, n the realty, t lkely that penalty cot be a power functon of lot demand a equaton (10), b a* W * Z. (10) However, for the ake of mplcty n olvng the model, we have et b=1. Otherwe, the model would become a nonlnear mxed-nteger formulaton whch much more challengng to olve. Moreover, n the obectve functon, agn weght to each cenaro whch ndcate mportance of that cenaro n decon makng, where 1. S Contrant (3) guarantee that um of total atfed and unatfed demand for cutomer n each cenaro do not exceed than t total demand. Equaton (4) repreent the capacty contrant n whch we enure that the total allocated demand to the faclty at node n cenaro (.e. W * X, ) doe not exceed t capacty. Addtonally, th contran ndcate that I cutomer hould be only agned to the faclte,.e., X, y. Equaton (5) enforce the p-robutne crteron. Accordngly, the obectve functon of each cenaro hould not be greater than 1 % of optmal cot of that cenaro. Robutne coeffcent can be et dfferently n each cenaro and denote by, n order to account for the fact that dfferent cenaro may have dfferent mportance level n our decon. However, for the ake of mplcty, we conder the ame value for each cenaro. Addtonally, th contrant ndcate that we are wllng to make addtonal nvetment n the nfratructure n order protect agant future poble drupton. Note that, th model would be equal to the clacal capactated P-medan problem f (robutne contrant wll be nactve), S 1, a=0, and contrant (3) replaced by the followng equaton X, 1 I, S. (11) J The mnmax regret formulaton whch a typcal obectve functon of robut model generally ued where the cenaro probablty not known. Th obectve functon mnmze maxmum regret over all poble cenaro. We can ealy change the propoed model n order to obtan th formulaton. To do o, we mply replace (4) (6)

6 Journal of Uncertan Sytem, Vol.7, No.1, pp.22-35, the obectve functon wth the robutne coeffcent whch yeld n the followng formulaton. We call th formulaton Robut Capactated Faclty Locaton Problem or, RCFLP. Mn RCFLP : (12) Subect to: 0 (13) (3), (4), (5), (6), (7), (8), (9). (14) Snce locatng large number of faclte not derable, equaton (6) lmt number of faclte that can be opened. Fnally, equaton (7 and 9) declare type of agnng varable and locaton varable repectvely. The determntc form of the propoed model formulaton whch would be ued to determne optmal cot of each cenaro a follow. Mn DCFLP : f y W d X, a W Z (15) J I J I Subect to: J I J X, Z 1 I, W X C y J y,, P (16) (17) (18) 4 Soluton Approach Z 0 I, (19) X, 0 I, J, (20) y 0,1 J. (21) For realtcally zed ntance t very challengng to olve mot of the p-scflp problem to optmalty wthn a reaonable computer CPU tme and memory by the well-known optmzaton olver uch a GAMS/CPLEX or Lngo. Therefore, n th part, two heurtc algorthm baed on the VNS and SA are developed to olve the p-scflp. Varable neghborhood earch a mple and powerful oluton framework to fnd near to optmal oluton to large cale and complex problem n a reaonable computer tme. It wa frtly propoed by Brmberg and Mladenovć [6] to olve contnue locaton allocaton problem. Comprehenve revew on the method and applcaton for th metaheurtc prepared by Hanen and Mladenovć [14]. Lkewe, SA that mmc coolng proce of materal another effcent oluton algorthm to olve hghly complex problem [36]. Broadly peakng, the p-scflp nvolve two man decon: (Ι) decon regardng faclte locaton, (.e., y ) and, (ΙΙ) decon about agnng cutomer to the located faclte (.e. X, and Z ). In the propoed model, f faclte locaton wa gven, an optmal oluton for the allocaton ub-problem eay to obtan. In other word, the problem can be ealy olved under any gven vector for the y. Therefore, the propoed algorthm n each teraton determne faclte' locaton, y, and then ue a local earch to determne the local optma for the allocaton ubproblem. Th local earch baed on an exact optmzaton algorthm that ue Branch & Bound and cuttng plane method n t framework. Addtonally, nce tetng a move everal tme not computatonally effcent, n thee algorthm, a long-term memory appled to tore vted move and prevent them from beng vted agan. In order to how a move vector, a J -dgt bnary vector wth equal length to the total potental faclte te ued. The th cell on the chromoome vector ndcate that at node faclty located (1) or not (0) n the current oluton. For example, Fg. 1 repreent the move vector of a network that nclude 10 potental faclte te and node 1 and 7 are elected n current oluton to be open and other node to be cloe.

7 28 R. Rahmanan et al.: Robut Capactated Faclty Locaton Problem Fgure 1: A chromoome for oluton repreentaton Moreover, the GAP value n percentage (relatve dtance between oluton and bet poble oluton) calculated by mean of equaton (22). In th equaton, Ffnal ndcate fnal oluton obtaned wth a heurtc method, and Fbet denote the lower bound obtaned by CPLEX. Ffnal Fbet % GAP 100. (22) Ffnal At the ubequent ubecton the propoed algorthm are dcued n detal and fnally, the flowchart of propoed VNS and outlne of SA are preented. 4.1 Contructon of an Intal Soluton Wth no doubt, ntal oluton ha a great mpact on every oluton algorthm. More mportantly, the VNS algorthm ue the bet recorded oluton a the ncumbent oluton to generate next move. Therefore, a cloe ntal oluton to the optmal oluton ha an enormou mpact on a ucceful VNS mplementaton. Our computatonal tet ndcated that optmal oluton of each cenaro ha a cloe gap wth the optmal oluton of the propoed model. A a reult, the ntal oluton obtaned from electng the optmal oluton of that cenaro that ha greater probablty of occurrence. 4.2 Local Search Local earch a method to return the local optma for the generated oluton. In th tudy oftware that make an nterface between MATLAB and GAMS ued to obtan the local optmal from CPLEX. Th oftware wa developed by Ferr [11]. In better word, algorthm are coded n MATLAB oftware and t generate the move (ee Fg. 2 and 3) and pa them to GAMS oftware and then the CPLEX olver wll olve the problem whle the locaton varable (.e. y ) are fxed. Th method mple and rather effcent. 4.3 Man Procedure of VNS Algorthm The man procedure of the algorthm llutrated n Fg. 2. The algorthm tart wth an ntal oluton for the locaton of faclte and then determne t local optma by CPLEX olver.in the mprovement phae, everal parameter uch a, pre-elected neghborhood tructure (.e. K max ) and et of oluton n the generated k th neghborhood (.e. N k (x)) have extreme mpact on the algorthm performance and need to be carefully tenured. A a reult, we examned everal tratege and the bet combnaton dcued here. For a gven neghborhood tructure, k, VNS randomly elect k bac node from the ncumbent oluton and replace them wth k non-bac node n order to generate a new oluton. Our computatonal tet ndcated that mot of mprovement move are acheved when k fxed at one. Conequently, we change the neghborhood tructure after a predefned number of teraton wthout mprovement ( U ). In the clacal VNS, U et to one; that mean, k k after any non-mprovement teraton neghborhood tructure ncreae by one. Note that, th algorthm et k to 1 f the obectve functon mproved or k k. max Algorthm take an ntal oluton a the ncumbent oluton, denoted by q bet, and fnd the local optma for th oluton (.e. F ) and whenever a move could mprove the obectve functon they wll be updated. Accordngly, t bet generate k th neghborhood from the bet recorded move, denoted by q, and f th oluton ha not been teted prevouly, t wll be checked to ee f t mprove the obectve functon. After applyng the local earch for th oluton, denoted by F current, followng tep are appled n order to decde whether we can change the neghborhood tructure or not, and the algorthm wll be repeated untl one of the toppng crtera perform.

8 Journal of Uncertan Sytem, Vol.7, No.1, pp.22-35, If F current < F bet then let { q bet = q, k=1, and UI=0} Ele f UI<=U k, then let {UI=UI+1, k =k, and q bet = q bet } Otherwe let {k =k+1, UI=0, and q bet = q bet } where, UI a counter for teraton and, whenever k k t wll be et to zero. Fnally, the algorthm top after max elapng the condered CPU tme, havng 3* J teraton wthout mprovement, or achevng the optmal oluton f t wa known beforehand. 4.4 Man Procedure of SA Algorthm Smulated Annealng (SA) a probabltc meta-heurtc that ha wdely been ued to fnd reaonable oluton n a lmted amount of tme (ee, [18, 19]). SA, unlke VNS, accept wore move wth a mall probablty n order to ecape from trappng n local optma. Th probablty calculated by the Boltzmann functon whch ue equaton (23). T Pr e r (23) where, C change n the evaluaton functon (.e. Fcurrent Fbet ), a contant, and T current temperature. If Pr wa greater than a random number, r, n nterval [0, 1], t accept wore move. Fgure 2: Man flow chart of the VNS algorthm The peudo-code of propoed SA algorthm to olve the model outlned at Fg. 3. To begn wth, an ntal oluton obtaned, then, the algorthm fnd t local optma, denoted by F bet. At the tep (2), the algorthm generate

9 30 R. Rahmanan et al.: Robut Capactated Faclty Locaton Problem a new oluton from X by randomly electng one bac node and replacng t wth one non-bac node. Accordngly, f the new oluton mproved the obectve functon both F bet and X wll be updated. Moreover, f the obectve functon dd not mprove, we can tll update X and accept wore move wth a mall probablty. Fnally, at each teraton temperature (.e. T) updated by ung the coolng rate parameter. Th parameter uually between 0.8 and In th way, a tme elape the probablty of acceptng wore move reduce. Moreover, we record the bet acheved move and after a predefned number of teraton (ay 100), the algorthm retart from that oluton f obectve functon dd not mproved. Fnally, the algorthm termnated n four cae: (1) the temperature to be le than TF, (2) the obectve functon not mproved durng 3* J teraton, (3) the relatve gap wth the bet poble oluton become approxmately zero and (4) maxmum condered CPU run tme elape. Intalze nput parameter {model,, T0, TF, toppng crtera} Step (1): Calculate an ntal oluton: X Fnd a local optma for th oluton X: F bet Repeat followng tep untl one of the toppng crtera perform Step (2): Generate a new oluton Y from X baed on k=1 Step (3): Apply the local earch to fnd the local optma of Y: F current Step (4): Calculate change n the obectve functon: Fcurrent Fbet If 0then let { F bet F Current and X Y } Otherwe {r = random(0, 1)} If exp( ) r then let { X Y } T Step (5): T T Numercal Experence 0 Fgure 3: Outlne of the propoed SA algorthm In th part, expermental reult are preented. The propoed problem wa mplemented on a range of tet problem and were olved wth tandard mathematcal programmng oftware GAMS , namely wth the branch-andbound algorthm of CPLEX 12.1, and algorthm were coded n MATLAB 7.6 and run on a Core 2 DELL Notebook wth 2GB RAM. In all data et, each node erve a both a cutomer and a potental faclty te (.e., I = J ). 5.1 Data Generaton To tet the algorthm performance, we appled numercal experment on both randomly generated and tandard tet problem. The tandard tet problem were obtaned from that publhed by Snyder and Dakn [32]. Thee tet problem have 49, 55, 88, and 150 node wth repectvely 9, 5, 9, and 9cenaro. We have ummarzed computatonal reult for thee ntance n Table 3 and 5. In the generated tet problem, the locaton of the cotumer are generated randomly and unformly dtrbuted over an area. In each data et for the generated tet problem, demand and fxed cot for cenaro-1 were drawn unformly from [0,10000] and [4000, 8000] repectvely and then rounded to the nearet nteger. Addtonal cenaro are dentfed by multplyng cenaro-1 to a random number drawn unformly from nterval [0.5, 1.5]. Travel dtance or travel tme between faclte and cutomer for cenaro-1 were equal to Eucldean dtance, and other cenaro are obtaned by multplyng random number drawn unformly from [0.8, 1.3] to the cenaro-1 data. Moreover, probablty of that cenaro occur calculated from equaton (24). I W S I W S. Condered capacty for each tet problem computed by mean of equaton (25). In other word, maxmum demand n all poble cenaro determned and then dvded to number of faclte. (24)

10 Journal of Uncertan Sytem, Vol.7, No.1, pp.22-35, max C S P I W. (25) 5.2 Computatonal Reult The propoed p-scflp ha not been condered formerly n the lterature. Conequently, we cannot provde comparon wth other computatonal tet. Accordngly, we tet the effcency of algorthm by olvng eghteen tet problem wth varou dmenon and comparng obtaned reult wth CPLEX. Followng parameter n th experment were ued. Maxmum condered CPU tme for all tet problem wa fxed at 4000 econd. Ten and ffteen percent of network node are condered a the number of faclte, or P. In addton, parameter of the propoed VNS algorthm were et to K max 3, B1 15, B2 10, B3 5 and the SA parameter were 0.97, T0 100, TF Regardng the propoed model, the requred parameter a and b were repectvely et to10 and 1whch yeld n a lnear mxed-nteger formulaton. To carryng the expermental tet, n th ecton, we frt olve the mnmax regret formulaton, RCFLP, and then preent computatonal reult for the propoed mn-expected-cot formulaton, p- SCFLP. In Table 2 through 5, thoe column under Tet Problem" whch are labeled by "TP", "N, "P", and "S" ndcate number of tet problem, number of network node, number of allowable faclte to be open, and number of cenaro, repectvely. Lkewe, thoe column whch are labeled by "L.B","GAP%", and"cpu Tme()" repreent lower bound, gap or relatve error (ee equaton (22)), and elaped CPU tme n econd, repectvely. In addton, the average CPU tme requrement n econd and average gap are lted n the lat row of Table 2-5. Moreover, n thee table bet performance n olvng each tet ntance boldfaced for the better algorthm. Table 2: Computatonal reult for the generated tet ntance: RCFLP 5.3 Numercal Experment for the RCFLP Note from Table 2 that, for TP1-3, on the average, VNS and SA found the optmal oluton n le than 1.66 and 7.48 econd, repectvely, whle CPLEX requred econd n order to acheve to the optmal oluton. Lkewe, for larger ntance (TP4 and 5), the heurtc repectvely by conumng and econd alo outperformed CPLEX whch elaped econd. All n all, VNS, SA, and CPLEX repectvely by elapng , , and econd obtaned oluton that have 2.47, 3.73, and 3.81 percentage gap wth the lower bound. A a reult, the propoed VNS algorthm ha been able to outperform both SA and CPLEX n term of

11 32 R. Rahmanan et al.: Robut Capactated Faclty Locaton Problem optmalty gap and CPU tme requrement. More nteretngly that n one cae the VNS wa able to mprove CPLEX oluton by more than 10 percentage. The propoed algorthmhave been alo appled to olve well-known tet problem taken from lterature. Our heurtc olved mot of thee ntance to optmalty wthn a mall fracton of CPLEX CPU tme. On the average, VNS wa able to mprove oluton optmalty gap by 40.04% and reduce CPU tme requrement by 30.47% n compare wth CPLEX. On the other hand, thee value for SA algorthm are repectvely and percentage. Accordngly, the propoed VNS outperform agan both SA and CPLEX. Table 3: Computatonal reult for the tandard tet ntance: RCFLP 5.4 Numercal Experment for the p-scflp In th part, we preent computatonal reult for the propoed model wth mn-expected-cot obectve functon (equaton (2 to 9)). To carryng th experment, the p value for each ntance wa et to t lower cae taken from Table 2 and 3. In general, nce the mn-expected-cot formulaton much le complex than mnmax formulaton, CPLEX could fnd optmal oluton almot for all ntance wthn a mall fracton of the condered CPU tme. A a reult, the algorthm were not able to mprove CPLEX reult too much. The next two conecutve table ummarze the computatonal reult of th experment. Table 4: Computatonal reult for the generated tet ntance: p-scflp

12 Journal of Uncertan Sytem, Vol.7, No.1, pp.22-35, Table 5: Computatonal reult for the tandard tet ntance: p-scflp Table 4 ummarze performance comparon between algorthm and CPLEX for the generated tet ntance. A hown n Table 4, VNS and SA, unlke CPLEX, faled to obtan the optmal oluton for 2 and 1 ntance, repectvely. Moreover, on the average, CPLEX, VNS, and SA for the ntance whch are olved optmally requred 63.61, 21.20, and econd, repectvely. However, takng all ntance together, the propoed VNS outperform both SA and CPLEX n term of CPU tme requrement and oluton qualty. Table 5 reveal computatonal reult on the tandard tet problem. For the maorty of ntance, all approache found the optmal oluton takng a few mnute. However, the bet oluton qualte and fatet performance wa acheved by VNS. Note that, the propoed VNS n two cae, TP9, n addton to reducng the CPLEX CPU tme, mproved optmalty gap a well. Thee expermental tude clearly demontrate the advantage of ung the propoed VNS n olvng the p-scflp. In order to nvetgate the convergence peed of propoed oluton algorthm, we took TP5, the tet problem wth 80 node and 8 faclte, a a repreentatve and oberved algorthm and CPLEX behavor n reducng optmalty gap over tme. A a reult, whenever the obectve functon mproved, t gap wth the lower bound reported n Table 2 and t tme were recorded and the reult are depcted n Fg. 4. A th plot how, the propoed ntal oluton ha le optmalty gap n compare wth the frt feable oluton of the CPLEX, roughly 34%. Clearly, the propoed VNS much fater than SA or CPLEX converge to the bet oluton. Though after approxmately 600 econd the VNS could not mprove the oluton any further, t ha acheved to the cloet oluton to the lower bound. Fgure 4: Convergence comparon between CPLEX, VNS and SA to olve TP5 over tme

13 34 R. Rahmanan et al.: Robut Capactated Faclty Locaton Problem 6 Concluon and Future Reearch Drecton In th paper an extenon of the capactated faclty locaton problem under uncertanty wa nvetgated. The mathematcal formulaton wa developed wth thee aumpton that demand and tranportaton cot are uncertan. In addton, nce faclte have hard contrant on the amount of demand they can upply, t lkely that they cannot completely erve all demand. A a reult, we developed the model formulaton to allow partal atfacton by ntroducng a penalty cot a a functon of unuppled demand. In th model, the obectve functon mnmzed expected cot whle relatve regret n each cenaro wa retrcted.in addton, we dcued the mnmax formulaton n whch the wort cae mnmzed. The propoed model formulaton a NP-Hard problem and very challengng to olve. A a reult, two fx-andoptmze heurtc baed on varable neghborhood earch and mulated annealng were developed to olve the model. In lne wth th ubect, the problem wa plt nto two ub-problem, locaton ub-problem and agnment ub-problem. The propoed algorthm teratvely determne the faclte locaton and then the agnment ubproblem olved by CPLEX. Algorthm performance n term of optmalty gap and CPU tme requrement wa teted by ung a varety of tet problem up to 150 node. We preented numercal experment for both mnmax regret formulaton and mn-expected-cot obectve functon. The reult demontrated that the algorthm bede of mplcty outperform CPLEX n term of oluton qualty and computer tme requrement. In general, the bet performance wa acheved by the VNS. Further attenton alo requred n the future reearche to nclude addtonal real aumpton uch a drupton n whch faclte may randomly fal to ervce cutomer or any practcal aumpton whch can be helpfully condered. Another nteretng reearch avenue to extend the model to take nto account acceblty to dfferent clae of faclte wth varou level of herarchy. Another lne of reearch would be developng other alternatve oluton algorthm to olve larger ntance. Moreover, developng an effcent local earch algorthm mght be an approprate future reearche drecton. Acknowledgement We would lke to thank two anonymou referee for ther contructve comment and uggeton whch mproved both the nght and preentaton of the paper. Reference [1] Aardal, K., Capactated faclty locaton: eparaton algorthm and computatonal experence, Mathematcal Programmng, vol.81, pp , [2] Averback, I., The mnmax relatve regret medan problem on network, INFORMS Journal on Computng, vol.17, no.4, pp , [3] Barahona, F., and F.A. Chudak, Near-optmal oluton to large-cale faclty locaton problem, Dcrete Optmzaton, vol.2, pp.35 50, [4] Bealey, J.E., Lagrangan heurtc for locaton problem, European Journal of Operatonal Reearch, vol.65, pp , [5] Brandeau, M.L., and S.S. Chu, An overvew of repreentatve problem n locaton reearch, Management Scence, vol.35, no.6, pp , [6] Brmberg, J., and N. Mladenovć, A varable neghborhood algorthm for olvng the contnuou locaton-allocaton problem, Stude n Locatonal Analy, vol.10, pp.1 12, [7] Cornueol, G., R. Srdharan, and J.M. Thzy, A comparon of heurtc and relaxaton for the capactated plant locaton problem, European Journal of Operatonal Reearch, vol.50, pp , [8] Cortnhal, M.J., and M.E. Captvo, Upper and lower bound for the ngle ource capactated locaton problem, European Journal of Operatonal Reearch, vol.151, pp , [9] Dakn, M.S., Network and Dcrete Locaton: Model, Algorthm, and Applcaton, Wley, New York, [10] Drezner, Z., and H.W. Hamacher, Faclty Locaton: Theory and Algorthm, Sprnger, [11] Ferr, M.C., MATLAB and GAMS: Interfacng Optmzaton and Vualzaton Software, Unverty of Wconn, 2005.

14 Journal of Uncertan Sytem, Vol.7, No.1, pp.22-35, [12] Flezar, K., and K.S. Hnd, An effectve VNS for the capactated p-medan problem, European Journal of Operatonal Reearch, vol.191, pp , [13] Ghezavat, V.R., M. Sad-Mehrabad, and S.J. Sadad, A robut approach to locaton-allocaton problem under uncertanty, Journal of Uncertan Sytem, vol.3, no.2, pp , [14] Hanen, P., N. Mladenovć, and J.A.M. Pérez, Varable neghbourhood earch: method and applcaton, Annal of Operaton Reearch, vol.175, pp , [15] Hnd, K.S., and K. Penkoz, Effcent oluton of large cale, ngle ource, capactated plant locaton problem, Journal of the Operatonal Reearch Socety, vol.50, pp , [16] Jabal-Amel, M.S., N. Azad, and A. Ratpour, Degnng a upply chan network model wth uncertan demand and lead tme, Journal of Uncertan Sytem, vol.3, no.2, pp , [17] Jacoben, S.K., Heurtc for the capactated plant locaton model, European Journal of Operatonal Reearch, vol.12, pp , [18] Jayaraman, V., and A. Ro, A mulated annealng methodology to dtrbuton network degn and management, European Journal of Operatonal Reearch, vol.144, pp , [19] Krkpatrck, S., C.D. Gelatt, and M.P. Vecch, Optmzaton by mulated annealng, Scence, vol.220, pp , [20] Korupolu, M., C. Plaxton, and R. Raaraman, Analy of a local earch heurtc for faclty locaton problem, Journal of Algorthm, vol.37, pp , [21] Kouvel, P., and G. Yu, Robut Dcrete Optmzaton and It Applcaton. Kluwer Academc Publher, [22] Kuehn, A.A., and M.J. Hamburger, A heurtc program for locatng warehoue, Management Scence, vol.9, pp , [23] Lorena, L.A.N., and E.L.F. Senne, Improvng tradtonal ubgradent cheme for Lagrangean relaxaton: an applcaton to locaton problem, Internatonal Journal of Mathematcal Algorthm, vol.1, pp , [24] Maranov, V., and D. Serra, Locaton problem n the publc ector, Faclty Locaton: Applcaton and Theory, edted by Z. Drezner and H.W. Hamacher, Sprnger, Berln, [25] Nau, R.M., An mproved algorthm for the capactated faclty locaton problem, Journal of the Operatonal Reearch Socety, vol.29, pp , [26] Owen, S.H., and M.S. Dakn, Strategc faclty locaton: a revew, European Journal of Operatonal Reearch, vol.111, pp , [27] Pahlavan, A., and M. Sad-Mehrabad, A fuzzy mult-obectve programmng for optmzaton of herarchcal ervce centre locaton, Journal of Uncertan Sytem, vol.5, no.3, pp , [28] Sa, G., Branch and bound and approxmate oluton to the capactated plant locaton problem, Operaton Reearch, vol.17, pp , [29] Sahnd, N.V., Optmzaton under uncertanty: tate-of-the-art and opportunte, Computer & Chemcal Engneerng, vol.28, pp , [30] Snyder, L.V., M.S. Dakn, and C.P. Teo, The tochatc locaton model wth rk poolng, European Journal of Operatonal Reearch, vol.179, pp , [31] Snyder, L.V., Faclty locaton under uncertanty: a revew, IIE Tranacton, vol.38, pp , [32] Snyder, L.V., and M.S. Dakn, Stochatc p-robut locaton problem, IIE Tranacton, vol.38, pp , [33] Srdharan, R., Invted revew the capactated plant locaton problem, European Journal of Operatonal Reearch, vol.87, pp , [34] Sun, M., A tabu earch heurtc procedure for the capactated faclty locaton problem, Journal of Heurtc, vol.18, pp , [35] Weaver, J.R., and R.L. Church, Computatonal procedure for locaton problem on tochatc network, Tranportaton Scence, vol.17, pp , [36] Yu, V.F., S.-W. Ln, W. Lee, and C.-J. Tng, A mulated annealng heurtc for the capactated locaton routng problem, Computer and Indutral Engneerng, vol.58, no.2, pp , 2010.

Additional File 1 - Detailed explanation of the expression level CPD

Additional File 1 - Detailed explanation of the expression level CPD Addtonal Fle - Detaled explanaton of the expreon level CPD A mentoned n the man text, the man CPD for the uterng model cont of two ndvdual factor: P( level gen P( level gen P ( level gen 2 (.).. CPD factor

More information

Solution Methods for Time-indexed MIP Models for Chemical Production Scheduling

Solution Methods for Time-indexed MIP Models for Chemical Production Scheduling Ian Davd Lockhart Bogle and Mchael Farweather (Edtor), Proceedng of the 22nd European Sympoum on Computer Aded Proce Engneerng, 17-2 June 212, London. 212 Elever B.V. All rght reerved. Soluton Method for

More information

Improvements on Waring s Problem

Improvements on Waring s Problem Improvement on Warng Problem L An-Png Bejng, PR Chna apl@nacom Abtract By a new recurve algorthm for the auxlary equaton, n th paper, we wll gve ome mprovement for Warng problem Keyword: Warng Problem,

More information

Adaptive Memory Programming for the Robust Capacitated International Sourcing Problem

Adaptive Memory Programming for the Robust Capacitated International Sourcing Problem Adaptve Memory Programmng for the Robut Capactated Internatonal Sourcng Problem Joé Lu González Velarde Centro de Stema de Manufactura, ITESM Monterrey, Méxco Lugonzal@campu.mty.tem.mx Rafael Martí Departamento

More information

Chapter 6 The Effect of the GPS Systematic Errors on Deformation Parameters

Chapter 6 The Effect of the GPS Systematic Errors on Deformation Parameters Chapter 6 The Effect of the GPS Sytematc Error on Deformaton Parameter 6.. General Beutler et al., (988) dd the frt comprehenve tudy on the GPS ytematc error. Baed on a geometrc approach and aumng a unform

More information

Estimation of Finite Population Total under PPS Sampling in Presence of Extra Auxiliary Information

Estimation of Finite Population Total under PPS Sampling in Presence of Extra Auxiliary Information Internatonal Journal of Stattc and Analy. ISSN 2248-9959 Volume 6, Number 1 (2016), pp. 9-16 Reearch Inda Publcaton http://www.rpublcaton.com Etmaton of Fnte Populaton Total under PPS Samplng n Preence

More information

Chapter 11. Supplemental Text Material. The method of steepest ascent can be derived as follows. Suppose that we have fit a firstorder

Chapter 11. Supplemental Text Material. The method of steepest ascent can be derived as follows. Suppose that we have fit a firstorder S-. The Method of Steepet cent Chapter. Supplemental Text Materal The method of teepet acent can be derved a follow. Suppoe that we have ft a frtorder model y = β + β x and we wh to ue th model to determne

More information

m = 4 n = 9 W 1 N 1 x 1 R D 4 s x i

m = 4 n = 9 W 1 N 1 x 1 R D 4 s x i GREEDY WIRE-SIZING IS LINEAR TIME Chr C. N. Chu D. F. Wong cnchu@c.utexa.edu wong@c.utexa.edu Department of Computer Scence, Unverty of Texa at Autn, Autn, T 787. ABSTRACT In nterconnect optmzaton by wre-zng,

More information

Root Locus Techniques

Root Locus Techniques Root Locu Technque ELEC 32 Cloed-Loop Control The control nput u t ynthezed baed on the a pror knowledge of the ytem plant, the reference nput r t, and the error gnal, e t The control ytem meaure the output,

More information

Specification -- Assumptions of the Simple Classical Linear Regression Model (CLRM) 1. Introduction

Specification -- Assumptions of the Simple Classical Linear Regression Model (CLRM) 1. Introduction ECONOMICS 35* -- NOTE ECON 35* -- NOTE Specfcaton -- Aumpton of the Smple Clacal Lnear Regreon Model (CLRM). Introducton CLRM tand for the Clacal Lnear Regreon Model. The CLRM alo known a the tandard lnear

More information

Start Point and Trajectory Analysis for the Minimal Time System Design Algorithm

Start Point and Trajectory Analysis for the Minimal Time System Design Algorithm Start Pont and Trajectory Analy for the Mnmal Tme Sytem Degn Algorthm ALEXANDER ZEMLIAK, PEDRO MIRANDA Department of Phyc and Mathematc Puebla Autonomou Unverty Av San Claudo /n, Puebla, 757 MEXICO Abtract:

More information

Small signal analysis

Small signal analysis Small gnal analy. ntroducton Let u conder the crcut hown n Fg., where the nonlnear retor decrbed by the equaton g v havng graphcal repreentaton hown n Fg.. ( G (t G v(t v Fg. Fg. a D current ource wherea

More information

Introduction to Interfacial Segregation. Xiaozhe Zhang 10/02/2015

Introduction to Interfacial Segregation. Xiaozhe Zhang 10/02/2015 Introducton to Interfacal Segregaton Xaozhe Zhang 10/02/2015 Interfacal egregaton Segregaton n materal refer to the enrchment of a materal conttuent at a free urface or an nternal nterface of a materal.

More information

Method Of Fundamental Solutions For Modeling Electromagnetic Wave Scattering Problems

Method Of Fundamental Solutions For Modeling Electromagnetic Wave Scattering Problems Internatonal Workhop on MehFree Method 003 1 Method Of Fundamental Soluton For Modelng lectromagnetc Wave Scatterng Problem Der-Lang Young (1) and Jhh-We Ruan (1) Abtract: In th paper we attempt to contruct

More information

Team. Outline. Statistics and Art: Sampling, Response Error, Mixed Models, Missing Data, and Inference

Team. Outline. Statistics and Art: Sampling, Response Error, Mixed Models, Missing Data, and Inference Team Stattc and Art: Samplng, Repone Error, Mxed Model, Mng Data, and nference Ed Stanek Unverty of Maachuett- Amhert, USA 9/5/8 9/5/8 Outlne. Example: Doe-repone Model n Toxcology. ow to Predct Realzed

More information

MULTIPLE REGRESSION ANALYSIS For the Case of Two Regressors

MULTIPLE REGRESSION ANALYSIS For the Case of Two Regressors MULTIPLE REGRESSION ANALYSIS For the Cae of Two Regreor In the followng note, leat-quare etmaton developed for multple regreon problem wth two eplanator varable, here called regreor (uch a n the Fat Food

More information

Amiri s Supply Chain Model. System Engineering b Department of Mathematics and Statistics c Odette School of Business

Amiri s Supply Chain Model. System Engineering b Department of Mathematics and Statistics c Odette School of Business Amr s Supply Chan Model by S. Ashtab a,, R.J. Caron b E. Selvarajah c a Department of Industral Manufacturng System Engneerng b Department of Mathematcs Statstcs c Odette School of Busness Unversty of

More information

Harmonic oscillator approximation

Harmonic oscillator approximation armonc ocllator approxmaton armonc ocllator approxmaton Euaton to be olved We are fndng a mnmum of the functon under the retrcton where W P, P,..., P, Q, Q,..., Q P, P,..., P, Q, Q,..., Q lnwgner functon

More information

Resource Allocation with a Budget Constraint for Computing Independent Tasks in the Cloud

Resource Allocation with a Budget Constraint for Computing Independent Tasks in the Cloud Resource Allocaton wth a Budget Constrant for Computng Independent Tasks n the Cloud Wemng Sh and Bo Hong School of Electrcal and Computer Engneerng Georga Insttute of Technology, USA 2nd IEEE Internatonal

More information

Valid Inequalities Based on Demand Propagation for Chemical Production Scheduling MIP Models

Valid Inequalities Based on Demand Propagation for Chemical Production Scheduling MIP Models Vald Inequalte Baed on Demand ropagaton for Chemcal roducton Schedulng MI Model Sara Velez, Arul Sundaramoorthy, And Chrto Maravela 1 Department of Chemcal and Bologcal Engneerng Unverty of Wconn Madon

More information

Improvements on Waring s Problem

Improvements on Waring s Problem Imrovement on Warng Problem L An-Png Bejng 85, PR Chna al@nacom Abtract By a new recurve algorthm for the auxlary equaton, n th aer, we wll gve ome mrovement for Warng roblem Keyword: Warng Problem, Hardy-Lttlewood

More information

On the SO 2 Problem in Thermal Power Plants. 2.Two-steps chemical absorption modeling

On the SO 2 Problem in Thermal Power Plants. 2.Two-steps chemical absorption modeling Internatonal Journal of Engneerng Reearch ISSN:39-689)(onlne),347-53(prnt) Volume No4, Iue No, pp : 557-56 Oct 5 On the SO Problem n Thermal Power Plant Two-tep chemcal aborpton modelng hr Boyadjev, P

More information

Confidence intervals for the difference and the ratio of Lognormal means with bounded parameters

Confidence intervals for the difference and the ratio of Lognormal means with bounded parameters Songklanakarn J. Sc. Technol. 37 () 3-40 Mar.-Apr. 05 http://www.jt.pu.ac.th Orgnal Artcle Confdence nterval for the dfference and the rato of Lognormal mean wth bounded parameter Sa-aat Nwtpong* Department

More information

A Weighted UTASTAR Method for the Multiple Criteria Decision Making with Interval Numbers

A Weighted UTASTAR Method for the Multiple Criteria Decision Making with Interval Numbers 3rd Internatonal Conference on Management Scence and Management Innovaton MSMI 2016) A Weghted UTASTAR Method for the Multple Crtera Decon Makng wth Interval Number Wen-Tao Xong Jng Cheng School of Mathematc

More information

Kernel Methods and SVMs Extension

Kernel Methods and SVMs Extension Kernel Methods and SVMs Extenson The purpose of ths document s to revew materal covered n Machne Learnng 1 Supervsed Learnng regardng support vector machnes (SVMs). Ths document also provdes a general

More information

Annexes. EC.1. Cycle-base move illustration. EC.2. Problem Instances

Annexes. EC.1. Cycle-base move illustration. EC.2. Problem Instances ec Annexes Ths Annex frst llustrates a cycle-based move n the dynamc-block generaton tabu search. It then dsplays the characterstcs of the nstance sets, followed by detaled results of the parametercalbraton

More information

Integrated approach in solving parallel machine scheduling and location (ScheLoc) problem

Integrated approach in solving parallel machine scheduling and location (ScheLoc) problem Internatonal Journal of Industral Engneerng Computatons 7 (2016) 573 584 Contents lsts avalable at GrowngScence Internatonal Journal of Industral Engneerng Computatons homepage: www.growngscence.com/ec

More information

Pythagorean triples. Leen Noordzij.

Pythagorean triples. Leen Noordzij. Pythagorean trple. Leen Noordz Dr.l.noordz@leennoordz.nl www.leennoordz.me Content A Roadmap for generatng Pythagorean Trple.... Pythagorean Trple.... 3 Dcuon Concluon.... 5 A Roadmap for generatng Pythagorean

More information

Seismic Reliability Analysis and Topology Optimization of Lifeline Networks

Seismic Reliability Analysis and Topology Optimization of Lifeline Networks The 4 th World Conference on Earthquake Engneerng October 2-7, 2008, Beng, Chna Semc Relablty Analy and Topology Optmzaton of Lfelne Network ABSTRACT: Je L and We Lu 2 Profeor, Dept. of Buldng Engneerng,

More information

Generalized Linear Methods

Generalized Linear Methods Generalzed Lnear Methods 1 Introducton In the Ensemble Methods the general dea s that usng a combnaton of several weak learner one could make a better learner. More formally, assume that we have a set

More information

Single-Facility Scheduling over Long Time Horizons by Logic-based Benders Decomposition

Single-Facility Scheduling over Long Time Horizons by Logic-based Benders Decomposition Sngle-Faclty Schedulng over Long Tme Horzons by Logc-based Benders Decomposton Elvn Coban and J. N. Hooker Tepper School of Busness, Carnege Mellon Unversty ecoban@andrew.cmu.edu, john@hooker.tepper.cmu.edu

More information

Energy Saving for Automatic Train Control in. Moving Block Signaling System

Energy Saving for Automatic Train Control in. Moving Block Signaling System Energy Savng for Automatc Tran Control n Movng Block Sgnalng Sytem Qng Gu, Xao-Yun Lu and Tao Tang Abtract Wth rapd development of the ralway traffc, the movng block gnalng ytem (MBS) method ha become

More information

Module 3 LOSSY IMAGE COMPRESSION SYSTEMS. Version 2 ECE IIT, Kharagpur

Module 3 LOSSY IMAGE COMPRESSION SYSTEMS. Version 2 ECE IIT, Kharagpur Module 3 LOSSY IMAGE COMPRESSION SYSTEMS Verson ECE IIT, Kharagpur Lesson 6 Theory of Quantzaton Verson ECE IIT, Kharagpur Instructonal Objectves At the end of ths lesson, the students should be able to:

More information

Discrete Simultaneous Perturbation Stochastic Approximation on Loss Function with Noisy Measurements

Discrete Simultaneous Perturbation Stochastic Approximation on Loss Function with Noisy Measurements 0 Amercan Control Conference on O'Farrell Street San Francco CA USA June 9 - July 0 0 Dcrete Smultaneou Perturbaton Stochatc Approxmaton on Lo Functon wth Noy Meaurement Q Wang and Jame C Spall Abtract

More information

OPTIMAL COMPUTING BUDGET ALLOCATION FOR MULTI-OBJECTIVE SIMULATION MODELS. David Goldsman

OPTIMAL COMPUTING BUDGET ALLOCATION FOR MULTI-OBJECTIVE SIMULATION MODELS. David Goldsman Proceedng of the 004 Wnter Smulaton Conference R.G. Ingall, M. D. Roett, J. S. Smth, and B. A. Peter, ed. OPTIMAL COMPUTING BUDGET ALLOCATION FOR MULTI-OBJECTIVE SIMULATION MODELS Loo Hay Lee Ek Peng Chew

More information

COMPARISON OF SOME RELIABILITY CHARACTERISTICS BETWEEN REDUNDANT SYSTEMS REQUIRING SUPPORTING UNITS FOR THEIR OPERATIONS

COMPARISON OF SOME RELIABILITY CHARACTERISTICS BETWEEN REDUNDANT SYSTEMS REQUIRING SUPPORTING UNITS FOR THEIR OPERATIONS Avalable onlne at http://sck.org J. Math. Comput. Sc. 3 (3), No., 6-3 ISSN: 97-537 COMPARISON OF SOME RELIABILITY CHARACTERISTICS BETWEEN REDUNDANT SYSTEMS REQUIRING SUPPORTING UNITS FOR THEIR OPERATIONS

More information

Lecture Notes on Linear Regression

Lecture Notes on Linear Regression Lecture Notes on Lnear Regresson Feng L fl@sdueducn Shandong Unversty, Chna Lnear Regresson Problem In regresson problem, we am at predct a contnuous target value gven an nput feature vector We assume

More information

EEL 6266 Power System Operation and Control. Chapter 3 Economic Dispatch Using Dynamic Programming

EEL 6266 Power System Operation and Control. Chapter 3 Economic Dispatch Using Dynamic Programming EEL 6266 Power System Operaton and Control Chapter 3 Economc Dspatch Usng Dynamc Programmng Pecewse Lnear Cost Functons Common practce many utltes prefer to represent ther generator cost functons as sngle-

More information

Alpha Risk of Taguchi Method with L 18 Array for NTB Type QCH by Simulation

Alpha Risk of Taguchi Method with L 18 Array for NTB Type QCH by Simulation Proceedng of the World Congre on Engneerng 00 Vol II WCE 00, July -, 00, London, U.K. Alpha Rk of Taguch Method wth L Array for NTB Type QCH by Smulaton A. Al-Refae and M.H. L Abtract Taguch method a wdely

More information

The Study of Teaching-learning-based Optimization Algorithm

The Study of Teaching-learning-based Optimization Algorithm Advanced Scence and Technology Letters Vol. (AST 06), pp.05- http://dx.do.org/0.57/astl.06. The Study of Teachng-learnng-based Optmzaton Algorthm u Sun, Yan fu, Lele Kong, Haolang Q,, Helongang Insttute

More information

MULTISTART OPTIMIZATION WITH A TRAINABLE DECISION MAKER FOR AVOIDING HIGH-VALUED LOCAL MINIMA

MULTISTART OPTIMIZATION WITH A TRAINABLE DECISION MAKER FOR AVOIDING HIGH-VALUED LOCAL MINIMA 3 rd Internatonal Conference on Experment/Proce/Sytem Modelng/Smulaton & Optmzaton 3 rd IC-EpMO Athen, 8- July, 2009 IC-EpMO MULTISTART OPTIMIZATION WITH A TRAINABLE DECISION MAKER FOR AVOIDING HIGH-VALUED

More information

arxiv: v1 [cs.gt] 15 Jan 2019

arxiv: v1 [cs.gt] 15 Jan 2019 Model and algorthm for tme-content rk-aware Markov game Wenje Huang, Pham Vet Ha and Wllam B. Hakell January 16, 2019 arxv:1901.04882v1 [c.gt] 15 Jan 2019 Abtract In th paper, we propoe a model for non-cooperatve

More information

MMA and GCMMA two methods for nonlinear optimization

MMA and GCMMA two methods for nonlinear optimization MMA and GCMMA two methods for nonlnear optmzaton Krster Svanberg Optmzaton and Systems Theory, KTH, Stockholm, Sweden. krlle@math.kth.se Ths note descrbes the algorthms used n the author s 2007 mplementatons

More information

Extension of VIKOR Method for MCDM Problem with Hesitant Linguistic Fuzzy Set and Possibility Degree

Extension of VIKOR Method for MCDM Problem with Hesitant Linguistic Fuzzy Set and Possibility Degree Extenon of VIKOR Method for MCDM Problem wth Hetant Lngutc Fuzzy et Poblty Degree Xnrong Yang a, Gang Qan b, Xangqan Feng c chool of Computer cence & Technology, Nanng Normal Unverty, Nanng 00, Chna a

More information

Chapter 5. Solution of System of Linear Equations. Module No. 6. Solution of Inconsistent and Ill Conditioned Systems

Chapter 5. Solution of System of Linear Equations. Module No. 6. Solution of Inconsistent and Ill Conditioned Systems Numercal Analyss by Dr. Anta Pal Assstant Professor Department of Mathematcs Natonal Insttute of Technology Durgapur Durgapur-713209 emal: anta.bue@gmal.com 1 . Chapter 5 Soluton of System of Lnear Equatons

More information

MODELLING OF TRANSIENT HEAT TRANSPORT IN TWO-LAYERED CRYSTALLINE SOLID FILMS USING THE INTERVAL LATTICE BOLTZMANN METHOD

MODELLING OF TRANSIENT HEAT TRANSPORT IN TWO-LAYERED CRYSTALLINE SOLID FILMS USING THE INTERVAL LATTICE BOLTZMANN METHOD Journal o Appled Mathematc and Computatonal Mechanc 7, 6(4), 57-65 www.amcm.pcz.pl p-issn 99-9965 DOI:.75/jamcm.7.4.6 e-issn 353-588 MODELLING OF TRANSIENT HEAT TRANSPORT IN TWO-LAYERED CRYSTALLINE SOLID

More information

Design and Optimization of Fuzzy Controller for Inverse Pendulum System Using Genetic Algorithm

Design and Optimization of Fuzzy Controller for Inverse Pendulum System Using Genetic Algorithm Desgn and Optmzaton of Fuzzy Controller for Inverse Pendulum System Usng Genetc Algorthm H. Mehraban A. Ashoor Unversty of Tehran Unversty of Tehran h.mehraban@ece.ut.ac.r a.ashoor@ece.ut.ac.r Abstract:

More information

Two-Layered Model of Blood Flow through Composite Stenosed Artery

Two-Layered Model of Blood Flow through Composite Stenosed Artery Avalable at http://pvamu.edu/aam Appl. Appl. Math. ISSN: 93-9466 Vol. 4, Iue (December 9), pp. 343 354 (Prevouly, Vol. 4, No.) Applcaton Appled Mathematc: An Internatonal Journal (AAM) Two-ayered Model

More information

Scattering of two identical particles in the center-of. of-mass frame. (b)

Scattering of two identical particles in the center-of. of-mass frame. (b) Lecture # November 5 Scatterng of two dentcal partcle Relatvtc Quantum Mechanc: The Klen-Gordon equaton Interpretaton of the Klen-Gordon equaton The Drac equaton Drac repreentaton for the matrce α and

More information

College of Computer & Information Science Fall 2009 Northeastern University 20 October 2009

College of Computer & Information Science Fall 2009 Northeastern University 20 October 2009 College of Computer & Informaton Scence Fall 2009 Northeastern Unversty 20 October 2009 CS7880: Algorthmc Power Tools Scrbe: Jan Wen and Laura Poplawsk Lecture Outlne: Prmal-dual schema Network Desgn:

More information

Outline and Reading. Dynamic Programming. Dynamic Programming revealed. Computing Fibonacci. The General Dynamic Programming Technique

Outline and Reading. Dynamic Programming. Dynamic Programming revealed. Computing Fibonacci. The General Dynamic Programming Technique Outlne and Readng Dynamc Programmng The General Technque ( 5.3.2) -1 Knapsac Problem ( 5.3.3) Matrx Chan-Product ( 5.3.1) Dynamc Programmng verson 1.4 1 Dynamc Programmng verson 1.4 2 Dynamc Programmng

More information

Winter 2008 CS567 Stochastic Linear/Integer Programming Guest Lecturer: Xu, Huan

Winter 2008 CS567 Stochastic Linear/Integer Programming Guest Lecturer: Xu, Huan Wnter 2008 CS567 Stochastc Lnear/Integer Programmng Guest Lecturer: Xu, Huan Class 2: More Modelng Examples 1 Capacty Expanson Capacty expanson models optmal choces of the tmng and levels of nvestments

More information

728. Mechanical and electrical elements in reduction of vibrations

728. Mechanical and electrical elements in reduction of vibrations 78. Mechancal and electrcal element n reducton of vbraton Katarzyna BIAŁAS The Slean Unverty of Technology, Faculty of Mechancal Engneerng Inttute of Engneerng Procee Automaton and Integrated Manufacturng

More information

Two Approaches to Proving. Goldbach s Conjecture

Two Approaches to Proving. Goldbach s Conjecture Two Approache to Provng Goldbach Conecture By Bernard Farley Adved By Charle Parry May 3 rd 5 A Bref Introducton to Goldbach Conecture In 74 Goldbach made h mot famou contrbuton n mathematc wth the conecture

More information

STOCHASTIC BEHAVIOUR OF COMMUNICATION SUBSYSTEM OF COMMUNICATION SATELLITE

STOCHASTIC BEHAVIOUR OF COMMUNICATION SUBSYSTEM OF COMMUNICATION SATELLITE IJS 4 () July Sharma & al ehavour of Subytem of ommuncaton Satellte SOHSI HVIOU O OMMUNIION SUSYSM O OMMUNIION SLLI SK Mttal eepankar Sharma & Neelam Sharma 3 S he author n th paper have dcued the tochatc

More information

The multivariate Gaussian probability density function for random vector X (X 1,,X ) T. diagonal term of, denoted

The multivariate Gaussian probability density function for random vector X (X 1,,X ) T. diagonal term of, denoted Appendx Proof of heorem he multvarate Gauan probablty denty functon for random vector X (X,,X ) px exp / / x x mean and varance equal to the th dagonal term of, denoted he margnal dtrbuton of X Gauan wth

More information

Structure and Drive Paul A. Jensen Copyright July 20, 2003

Structure and Drive Paul A. Jensen Copyright July 20, 2003 Structure and Drve Paul A. Jensen Copyrght July 20, 2003 A system s made up of several operatons wth flow passng between them. The structure of the system descrbes the flow paths from nputs to outputs.

More information

Hongyi Miao, College of Science, Nanjing Forestry University, Nanjing ,China. (Received 20 June 2013, accepted 11 March 2014) I)ϕ (k)

Hongyi Miao, College of Science, Nanjing Forestry University, Nanjing ,China. (Received 20 June 2013, accepted 11 March 2014) I)ϕ (k) ISSN 1749-3889 (prnt), 1749-3897 (onlne) Internatonal Journal of Nonlnear Scence Vol.17(2014) No.2,pp.188-192 Modfed Block Jacob-Davdson Method for Solvng Large Sparse Egenproblems Hongy Mao, College of

More information

Supporting Information. Hydroxyl Radical Production by H 2 O 2 -Mediated. Conditions

Supporting Information. Hydroxyl Radical Production by H 2 O 2 -Mediated. Conditions Supportng Informaton Hydroxyl Radcal Producton by H 2 O 2 -Medated Oxdaton of Fe(II) Complexed by Suwannee Rver Fulvc Acd Under Crcumneutral Frehwater Condton Chrtopher J. Mller, Andrew L. Roe, T. Davd

More information

A New Inverse Reliability Analysis Method Using MPP-Based Dimension Reduction Method (DRM)

A New Inverse Reliability Analysis Method Using MPP-Based Dimension Reduction Method (DRM) roceedng of the ASME 007 Internatonal Degn Engneerng Techncal Conference & Computer and Informaton n Engneerng Conference IDETC/CIE 007 September 4-7, 007, La Vega, eada, USA DETC007-35098 A ew Inere Relablty

More information

Resonant FCS Predictive Control of Power Converter in Stationary Reference Frame

Resonant FCS Predictive Control of Power Converter in Stationary Reference Frame Preprnt of the 9th World Congre The Internatonal Federaton of Automatc Control Cape Town, South Afrca. Augut -9, Reonant FCS Predctve Control of Power Converter n Statonary Reference Frame Lupng Wang K

More information

Chapter Newton s Method

Chapter Newton s Method Chapter 9. Newton s Method After readng ths chapter, you should be able to:. Understand how Newton s method s dfferent from the Golden Secton Search method. Understand how Newton s method works 3. Solve

More information

NUMERICAL DIFFERENTIATION

NUMERICAL DIFFERENTIATION NUMERICAL DIFFERENTIATION 1 Introducton Dfferentaton s a method to compute the rate at whch a dependent output y changes wth respect to the change n the ndependent nput x. Ths rate of change s called the

More information

Module 9. Lecture 6. Duality in Assignment Problems

Module 9. Lecture 6. Duality in Assignment Problems Module 9 1 Lecture 6 Dualty n Assgnment Problems In ths lecture we attempt to answer few other mportant questons posed n earler lecture for (AP) and see how some of them can be explaned through the concept

More information

On the Multicriteria Integer Network Flow Problem

On the Multicriteria Integer Network Flow Problem BULGARIAN ACADEMY OF SCIENCES CYBERNETICS AND INFORMATION TECHNOLOGIES Volume 5, No 2 Sofa 2005 On the Multcrtera Integer Network Flow Problem Vassl Vasslev, Marana Nkolova, Maryana Vassleva Insttute of

More information

Extended Prigogine Theorem: Method for Universal Characterization of Complex System Evolution

Extended Prigogine Theorem: Method for Universal Characterization of Complex System Evolution Extended Prgogne Theorem: Method for Unveral Characterzaton of Complex Sytem Evoluton Sergey amenhchkov* Mocow State Unverty of M.V. Lomonoov, Phycal department, Rua, Mocow, Lennke Gory, 1/, 119991 Publhed

More information

Information Acquisition in Global Games of Regime Change (Online Appendix)

Information Acquisition in Global Games of Regime Change (Online Appendix) Informaton Acquton n Global Game of Regme Change (Onlne Appendx) Mchal Szkup and Iabel Trevno Augut 4, 05 Introducton Th appendx contan the proof of all the ntermedate reult that have been omtted from

More information

The Minimum Universal Cost Flow in an Infeasible Flow Network

The Minimum Universal Cost Flow in an Infeasible Flow Network Journal of Scences, Islamc Republc of Iran 17(2): 175-180 (2006) Unversty of Tehran, ISSN 1016-1104 http://jscencesutacr The Mnmum Unversal Cost Flow n an Infeasble Flow Network H Saleh Fathabad * M Bagheran

More information

LINEAR REGRESSION ANALYSIS. MODULE IX Lecture Multicollinearity

LINEAR REGRESSION ANALYSIS. MODULE IX Lecture Multicollinearity LINEAR REGRESSION ANALYSIS MODULE IX Lecture - 30 Multcollnearty Dr. Shalabh Department of Mathematcs and Statstcs Indan Insttute of Technology Kanpur 2 Remedes for multcollnearty Varous technques have

More information

Week 5: Neural Networks

Week 5: Neural Networks Week 5: Neural Networks Instructor: Sergey Levne Neural Networks Summary In the prevous lecture, we saw how we can construct neural networks by extendng logstc regresson. Neural networks consst of multple

More information

APPROXIMATE FUZZY REASONING BASED ON INTERPOLATION IN THE VAGUE ENVIRONMENT OF THE FUZZY RULEBASE AS A PRACTICAL ALTERNATIVE OF THE CLASSICAL CRI

APPROXIMATE FUZZY REASONING BASED ON INTERPOLATION IN THE VAGUE ENVIRONMENT OF THE FUZZY RULEBASE AS A PRACTICAL ALTERNATIVE OF THE CLASSICAL CRI Kovác, Sz., Kóczy, L.T.: Approxmate Fuzzy Reaonng Baed on Interpolaton n the Vague Envronment of the Fuzzy Rulebae a a Practcal Alternatve of the Clacal CRI, Proceedng of the 7 th Internatonal Fuzzy Sytem

More information

A Novel Approach for Testing Stability of 1-D Recursive Digital Filters Based on Lagrange Multipliers

A Novel Approach for Testing Stability of 1-D Recursive Digital Filters Based on Lagrange Multipliers Amercan Journal of Appled Scence 5 (5: 49-495, 8 ISSN 546-939 8 Scence Publcaton A Novel Approach for Tetng Stablty of -D Recurve Dgtal Flter Baed on Lagrange ultpler KRSanth, NGangatharan and Ponnavakko

More information

Calculating Jacobian coefficients of primitive constraints with respect to Euler parameters

Calculating Jacobian coefficients of primitive constraints with respect to Euler parameters Calculatng Jacoban coeffcent of prmtve contrant wth repect to Euler parameter Yong Lu, Ha-Chuan Song, Jun-Ha Yong To cte th veron: Yong Lu, Ha-Chuan Song, Jun-Ha Yong. Calculatng Jacoban coeffcent of prmtve

More information

ECE559VV Project Report

ECE559VV Project Report ECE559VV Project Report (Supplementary Notes Loc Xuan Bu I. MAX SUM-RATE SCHEDULING: THE UPLINK CASE We have seen (n the presentaton that, for downlnk (broadcast channels, the strategy maxmzng the sum-rate

More information

FUZZY GOAL PROGRAMMING VS ORDINARY FUZZY PROGRAMMING APPROACH FOR MULTI OBJECTIVE PROGRAMMING PROBLEM

FUZZY GOAL PROGRAMMING VS ORDINARY FUZZY PROGRAMMING APPROACH FOR MULTI OBJECTIVE PROGRAMMING PROBLEM Internatonal Conference on Ceramcs, Bkaner, Inda Internatonal Journal of Modern Physcs: Conference Seres Vol. 22 (2013) 757 761 World Scentfc Publshng Company DOI: 10.1142/S2010194513010982 FUZZY GOAL

More information

A new Approach for Solving Linear Ordinary Differential Equations

A new Approach for Solving Linear Ordinary Differential Equations , ISSN 974-57X (Onlne), ISSN 974-5718 (Prnt), Vol. ; Issue No. 1; Year 14, Copyrght 13-14 by CESER PUBLICATIONS A new Approach for Solvng Lnear Ordnary Dfferental Equatons Fawz Abdelwahd Department of

More information

Computing Correlated Equilibria in Multi-Player Games

Computing Correlated Equilibria in Multi-Player Games Computng Correlated Equlbra n Mult-Player Games Chrstos H. Papadmtrou Presented by Zhanxang Huang December 7th, 2005 1 The Author Dr. Chrstos H. Papadmtrou CS professor at UC Berkley (taught at Harvard,

More information

No! Yes! Only if reactions occur! Yes! Ideal Gas, change in temperature or pressure. Survey Results. Class 15. Is the following possible?

No! Yes! Only if reactions occur! Yes! Ideal Gas, change in temperature or pressure. Survey Results. Class 15. Is the following possible? Survey Reult Chapter 5-6 (where we are gong) % of Student 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% Hour Spent on ChE 273 1-2 3-4 5-6 7-8 9-10 11+ Hour/Week 2008 2009 2010 2011 2012 2013 2014 2015 2017 F17

More information

MODELLING OF STOCHASTIC PARAMETERS FOR CONTROL OF CITY ELECTRIC TRANSPORT SYSTEMS USING EVOLUTIONARY ALGORITHM

MODELLING OF STOCHASTIC PARAMETERS FOR CONTROL OF CITY ELECTRIC TRANSPORT SYSTEMS USING EVOLUTIONARY ALGORITHM MODELLING OF STOCHASTIC PARAMETERS FOR CONTROL OF CITY ELECTRIC TRANSPORT SYSTEMS USING EVOLUTIONARY ALGORITHM Mkhal Gorobetz, Anatoly Levchenkov Inttute of Indutral Electronc and Electrotechnc, Rga Techncal

More information

A special case of the generalized pooling problem arising in the mining industry

A special case of the generalized pooling problem arising in the mining industry A pecal cae of the generalzed poolng problem arng n the mnng ndutry Nataha Boland Thoma Kalnowk Faban Rgternk Martn Savelbergh Aprl 26, 216 Abtract Iron ore and coal are ubtantal contrbutor to Autrala

More information

This appendix presents the derivations and proofs omitted from the main text.

This appendix presents the derivations and proofs omitted from the main text. Onlne Appendx A Appendx: Omtted Dervaton and Proof Th appendx preent the dervaton and proof omtted from the man text A Omtted dervaton n Secton Mot of the analy provded n the man text Here, we formally

More information

Supporting Information

Supporting Information Supportng Informaton The neural network f n Eq. 1 s gven by: f x l = ReLU W atom x l + b atom, 2 where ReLU s the element-wse rectfed lnear unt, 21.e., ReLUx = max0, x, W atom R d d s the weght matrx to

More information

Designing Service Competitions among Heterogeneous Suppliers

Designing Service Competitions among Heterogeneous Suppliers Degnng Servce Competton among Heterogeneou Suppler Ehan Elah and Saf Benaafar Graduate Program n Indutral & Sytem Engneerng Department of Mechancal Engneerng Unverty of Mnneota, Mnneapol, M 55455 elah@me.umn.edu

More information

Foresighted Resource Reciprocation Strategies in P2P Networks

Foresighted Resource Reciprocation Strategies in P2P Networks Foreghted Reource Recprocaton Stratege n PP Networ Hyunggon Par and Mhaela van der Schaar Electrcal Engneerng Department Unverty of Calforna Lo Angele (UCLA) Emal: {hgpar mhaela@ee.ucla.edu Abtract We

More information

Lecture 10 Support Vector Machines II

Lecture 10 Support Vector Machines II Lecture 10 Support Vector Machnes II 22 February 2016 Taylor B. Arnold Yale Statstcs STAT 365/665 1/28 Notes: Problem 3 s posted and due ths upcomng Frday There was an early bug n the fake-test data; fxed

More information

bounds compared to SB and SBB bounds as the former two have an index parameter, while the latter two

bounds compared to SB and SBB bounds as the former two have an index parameter, while the latter two 1 Queung Procee n GPS and PGPS wth LRD Traffc Input Xang Yu, Ian L-Jn Thng, Yumng Jang and Chunmng Qao Department of Computer Scence and Engneerng State Unverty of New York at Buffalo Department of Electrcal

More information

Speeding up Computation of Scalar Multiplication in Elliptic Curve Cryptosystem

Speeding up Computation of Scalar Multiplication in Elliptic Curve Cryptosystem H.K. Pathak et. al. / (IJCSE) Internatonal Journal on Computer Scence and Engneerng Speedng up Computaton of Scalar Multplcaton n Ellptc Curve Cryptosystem H. K. Pathak Manju Sangh S.o.S n Computer scence

More information

A Hybrid Evolution Algorithm with Application Based on Chaos Genetic Algorithm and Particle Swarm Optimization

A Hybrid Evolution Algorithm with Application Based on Chaos Genetic Algorithm and Particle Swarm Optimization Natonal Conference on Informaton Technology and Computer Scence (CITCS ) A Hybrd Evoluton Algorthm wth Applcaton Baed on Chao Genetc Algorthm and Partcle Swarm Optmzaton Fu Yu School of Computer & Informaton

More information

CHAPTER IV RESEARCH FINDING AND ANALYSIS

CHAPTER IV RESEARCH FINDING AND ANALYSIS CHAPTER IV REEARCH FINDING AND ANALYI A. Descrpton of Research Fndngs To fnd out the dfference between the students who were taught by usng Mme Game and the students who were not taught by usng Mme Game

More information

For now, let us focus on a specific model of neurons. These are simplified from reality but can achieve remarkable results.

For now, let us focus on a specific model of neurons. These are simplified from reality but can achieve remarkable results. Neural Networks : Dervaton compled by Alvn Wan from Professor Jtendra Malk s lecture Ths type of computaton s called deep learnng and s the most popular method for many problems, such as computer vson

More information

Not at Steady State! Yes! Only if reactions occur! Yes! Ideal Gas, change in temperature or pressure. Yes! Class 15. Is the following possible?

Not at Steady State! Yes! Only if reactions occur! Yes! Ideal Gas, change in temperature or pressure. Yes! Class 15. Is the following possible? Chapter 5-6 (where we are gong) Ideal gae and lqud (today) Dente Partal preure Non-deal gae (next tme) Eqn. of tate Reduced preure and temperature Compreblty chart (z) Vapor-lqud ytem (Ch. 6) Vapor preure

More information

A Hybrid Variational Iteration Method for Blasius Equation

A Hybrid Variational Iteration Method for Blasius Equation Avalable at http://pvamu.edu/aam Appl. Appl. Math. ISSN: 1932-9466 Vol. 10, Issue 1 (June 2015), pp. 223-229 Applcatons and Appled Mathematcs: An Internatonal Journal (AAM) A Hybrd Varatonal Iteraton Method

More information

Batch RL Via Least Squares Policy Iteration

Batch RL Via Least Squares Policy Iteration Batch RL Va Leat Square Polcy Iteraton Alan Fern * Baed n part on lde by Ronald Parr Overvew Motvaton LSPI Dervaton from LSTD Expermental reult Onlne veru Batch RL Onlne RL: ntegrate data collecton and

More information

An Interactive Optimisation Tool for Allocation Problems

An Interactive Optimisation Tool for Allocation Problems An Interactve Optmsaton ool for Allocaton Problems Fredr Bonäs, Joam Westerlund and apo Westerlund Process Desgn Laboratory, Faculty of echnology, Åbo Aadem Unversty, uru 20500, Fnland hs paper presents

More information

Strong-Stability-Preserving, K-Step, 5- to 10-Stage, Hermite-Birkhoff Time-Discretizations of Order 12

Strong-Stability-Preserving, K-Step, 5- to 10-Stage, Hermite-Birkhoff Time-Discretizations of Order 12 Amercan Journal of Computatonal Mathematc 20 72-82 do:04236/acm202008 Publhed Onlne June 20 (http://wwwcrporg/ournal/acm) Strong-Stablty-Preervng K-Step 5- to 0-Stage Hermte-Brkhoff Tme-Dcretzaton of Order

More information

Curve Fitting with the Least Square Method

Curve Fitting with the Least Square Method WIKI Document Number 5 Interpolaton wth Least Squares Curve Fttng wth the Least Square Method Mattheu Bultelle Department of Bo-Engneerng Imperal College, London Context We wsh to model the postve feedback

More information

This is a repository copy of An iterative orthogonal forward regression algorithm.

This is a repository copy of An iterative orthogonal forward regression algorithm. Th a repotory copy of An teratve orthogonal forward regreon algorthm. Whte Roe Reearch Onlne URL for th paper: http://eprnt.whteroe.ac.uk/0735/ Veron: Accepted Veron Artcle: Guo, Y., Guo, L. Z., Bllng,

More information

A NOVEL FAMILY OF WEIGHTED AVERAGE VOTERS FOR FAULT-TOLERANT COMPUTER CONTROL SYSTEMS

A NOVEL FAMILY OF WEIGHTED AVERAGE VOTERS FOR FAULT-TOLERANT COMPUTER CONTROL SYSTEMS A OVEL FAMILY OF WEIGHTED AVERAGE VOTERS FOR FAULT-TOLERAT COMPUTER COTROL SYSTEMS G. Latf-Shabgah *, A. J. Hrt *, and S. Bennett + * Department of Telematc, Open Unverty Walton Hall, Mlton Keyne, MK7

More information

A NUMERICAL MODELING OF MAGNETIC FIELD PERTURBATED BY THE PRESENCE OF SCHIP S HULL

A NUMERICAL MODELING OF MAGNETIC FIELD PERTURBATED BY THE PRESENCE OF SCHIP S HULL A NUMERCAL MODELNG OF MAGNETC FELD PERTURBATED BY THE PRESENCE OF SCHP S HULL M. Dennah* Z. Abd** * Laboratory Electromagnetc Sytem EMP BP b Ben-Aknoun 606 Alger Algera ** Electronc nttute USTHB Alger

More information

Synchronization Protocols. Task Allocation Bin-Packing Heuristics: First-Fit Subtasks assigned in arbitrary order To allocate a new subtask T i,j

Synchronization Protocols. Task Allocation Bin-Packing Heuristics: First-Fit Subtasks assigned in arbitrary order To allocate a new subtask T i,j End-to-End Schedulng Framework 1. Tak allocaton: bnd tak to proceor 2. Synchronzaton protocol: enforce precedence contrant 3. Subdeadlne agnment 4. Schedulablty analy Tak Allocaton Bn-Packng eurtc: Frt-Ft

More information