In spite of the remarkable improvements in the quality of general purpose mixed-integer

Size: px
Start display at page:

Download "In spite of the remarkable improvements in the quality of general purpose mixed-integer"

Transcription

1 Modellng Praccal Lo-Szng Problems as Mxed-Ineger Programs Gaean Belvaux Laurence A. Wolsey Cockerll-Sambre, Flémalle Work carred ou a CORE, Unversé Caholque de Louvan CORE and INMA, Unversé Caholque de Louvan, Voe du Roman Pays 34, 1348 Louvan-la-Neuve, Belgum gaean.belvaux@cockerll-sambre.com wolsey@core.ucl.ac.be In spe of he remarkable mprovemens n he qualy of general purpose mxed-neger programmng sofware, he effecve soluon of a varey of lo-szng problems depends crucally on he developmen of gh formulaons for he specal problem feaures occurrng n pracce. Afer revewng some of he basc preprocessng echnques for handlng safey socks and mullevel problems, we dscuss a varey of aspecs arsng parcularly n small and large bucke (me perod) models such as sar-ups, changeovers, mnmum bach szes, choce of one or wo se-ups per perod, ec. A se of applcaons s descrbed ha conans one or more of hese specal feaures, and some ndcave compuaonal resuls are presened. Fnally, o show anoher echnque ha s useful, a slghly dfferen (supply chan) applcaon s presened, for whch he a pror addon of some smple mxed-neger nequales, based on aggregaon, leads o mporan mprovemens n he resuls. (Lo-Szng; Producon Plannng; Mxed-Ineger Programmng; Vald Inequales; Reformulaon) 1. Inroducon Much work has been done n acklng academc lo-szng problems by usng Lagrangan relaxaon combned wh branch-and-bound (Afenaks and Gavsh 1986, Daby e al. 1992, Tempelmeer and Dersroff 1996, Thzy and van Wassenhove 1985) or by heurscs usng boh one-pass greedy-ype approaches (Dxon and Slver 1981), meaheurscs such as abu search (Smpson and Erenguc 1998a, b), and column generaon (Carysse e al. (1990, 1993) and Vanderbeck (1998)). Relavely lle compuaon s repored based on mxed-neger programmng reformulaons and/or cung planes, wh he excepon of (Barany e al. 1984, Consanno 1996, Eppen and Marn 1987, Poche and Wolsey 1991). The hess here, and n he companon paper (Belvaux and Wolsey 2000), s ha many praccal lo-szng problems can be effecvely solved va mxed-neger programmng. Two mporan elemens of such an approach are () gh a pror formulaons of he specal problem feaures ncurred n pracce, and () general or specal purpose mxed-neger programmng (MIP) sofware ncorporang cung planes. The second elemen can be found n recen commercal MIP sysems such as XPRESS and CPLEX, or research codes such as MINTO (Savelsbergh and Nemhauser 1993) and BC-OPT (Corder e al. 1999), or n BC- PROD (a modellng and opmzaon sysem specfcally desgned for lo-szng problems, descrbed n a recen paper (Belvaux and Wolsey 2000)). Here we concenrae on he frs elemen, descrbng and formulang varous aspecs of lo-szng encounered n pracce. Specfcally n 2, we remnd he reader of some of he basc nequales ha can be added a pror or as cung planes n lo-szng models, as well as /01/4707/0993$ elecronc ISSN Managemen Scence 2001 INFORMS Vol. 47, No. 7, July 2001 pp

2 Modellng Praccal Lo-Szng Problems as Mxed-Ineger Programs wo basc modellng choces, ways of reang safey socks, and he possble use of echelon socks n mullevel problems. These choces ofen nfluence he compuaonal performance of commercal and specalzed mxed-neger programmng sysems. In 3 we consder how o bes model sar-ups, changeovers, and swch-offs n boh small bucke models wh one or wo se-ups per perod, and n bg bucke models n whch numerous ems can be produced per perod. In 4 we consder how o model aspecs such as mnmum bach szes, fullcapacy producon, and mnmum down-mes n small bucke models. We hen descrbe brefly a se of praccal applcaons conanng one or more of he above feaures, and ndcae how he formulaons can be mproved/modfed usng he observaons from he earler secons. Some lmed compuaonal resuls are presened. In 6 we consder a dfferen, bu sgnfcan (supply chan) applcaon nvolvng a large number of producs, ses, and sales areas, where he dffculy les n he requremen o produce n mulples of a fxed bach sze or o produce mnmum amouns. Here, based on aggregaon, some smple vald nequales are added o he formulaon a pror. We close wh some bref remarks. 2. Some Basc Modellng Opons 2.1. A Basc Model The noaon used resembles as closely as possble ha n Belvaux and Wolsey (2000). We consder frs lo-szng problems wh NI ems/producs ha can be produced n parallel on NK machnes wh a horzon of NT me perods. The basc daa (apar from coss) are s he prespecfed demand for em n perod. d k s he rae of producon of em on machne k n perod. We wll assume ha k = 1 unless oherwse saed. C k s he maxmum amoun of em ha can be produced on machne k n perod. L k s he oal capacy of machne k n perod. a k s he capacy requred per un of em on machne k n perod. b k s he fxed amoun of capacy requred f machne k s se up o produce em n perod. The basc varables are defned as follows: s s he sock of a he end of. r s he backlog of a he end of. x k s he producon of on k n. y k = 1 f machne k s made avalable o produce em n, and y k = 0 oherwse. y k s he se-up varable. The mnmal consrans of a mul-em, mulmachne sngle-level problem n whch one or several ems are produced on each machne durng a perod are s 1 r 1 + k a k xk k xk = d + s r for all (1) x k C k y k for all k (2) + b k yk L k for all k (3) x s r 0 y 0 1 (4) When only one or wo ems can be produced per perod, s ofen mporan o ake no accoun he cos and/or me of preparng he machne for he producon of a new em or for cleanng he machne aferwards. The addonal varables needed follow. z k = 1 f machne k s se up o produce em n perod, bu was no se up for em mmedaely beforehand; z k s called he sar-up varable. w k s he analogous swch-off varable. w k = 1f machne k s se up o produce em n perod, bu s no se up mmedaely aferwards. The basc model exended o nclude sar-ups has he addonal consrans z k +1 wk = y k +1 yk for all k (5) z k y k for all k (6) z y 0 1 (7) Consrans (5) are ofen wren n he alernave form or z k +1 yk +1 yk for all k (8) w k y k y k +1 for all k (9) 994 Managemen Scence/Vol. 47, No. 7, July 2001

3 Modellng Praccal Lo-Szng Problems as Mxed-Ineger Programs when here s no need for he swch-off or he sarup varables, respecvely. Also, wh sar-ups presen he machne capacy consran (3) ofen akes he smple form y k 1 for all k (10) wh L k = 1 for all k, and ak = 0 b k = 1 for all k. In he nex subsecon we presen some of he nequales ha can be added a pror o he basc model; n he followng subsecon we consder how hey can also be used as cung planes Vald Inequales: A Pror Reformulaon Consder frs a sngle em wh a sngle machne and no backloggng. We drop ndces when possble so as o smplfy he noaon. For each em he flow balance consran (1), he capacy consran (2), and he bound consrans (4) gve a se of he form s 1 + x = d + s for all (11) x C y for all (12) x s 0 y 0 1 (13) Ignorng he values of he C, and reang he problem as uncapacaed, he nequales s k 1 d j 1 y k y j (14) are vald, because he sock a he end of perod k 1 mus be used o sasfy he demand n perod j f no producon akes place n he nerval k k + 1 j y k + +y j = 0. As here are only O n 2 such nequales, some or all of hem can be added a pror o he formulaon. Smple expermenaon shows ha even jus addng he n nequales s k 1 d k 1 y k for k = 1 n (15) obaned by akng he nequales (14) wh = k ypcally produces an mporan mprovemen n he lower bound value of he LP relaxaon. When C = max C s resrcve, gher nequales can be obaned. Specfcally by aggregang he balance consrans over he nerval k and hen usng he varable upper bound consrans, one obans he surrogae nequaly s k 1 + C y j d k wh s k 1 0 and y j neger. For hs se, he basc mxed-neger roundng nequaly (Nemhauser and Wolsey 1988) s s k 1 k ( k y j ) (16) where k = d k /C, and k = d k C k 1. Ths agan gves O n 2 nequales ha can possbly be added a pror. We now ndcae how he above nequales can be easly adaped n he presence of sar-up or backloggng varables. Wh sar-ups he nequales (14) can be replaced by s k 1 d j 1 y k z k+1 z j (17) because y k +z k+1 + +z j = 0 also mples ha here s no producon n he nerval k j. Wh backloggng, s no dffcul o see ha (14) can be replaced by s k 1 + r j d j 1 y k y j (18) and (17) can be exended n he same way, whereas (16) can be replaced by s k 1 + r k ( k y j ) (19) 2.3. Vald Inequales: Cus When usng one of he branch-and-cu sysems CPLEX, XPRESS, MINTO, and BC-OPT s possble o add some or all of he above vald nequales a pror. However, hese sysems are also able o generae some of hem auomacally as cung planes. These sysems ypcally generae cus by recognzng and combnng consrans much as above. For nsance, akng (11) wh = k, combnng wh he varable upper bound consran (12), and usng he nonnegavy of s k leads o he surrogae s k 1 + C k y k d k s k 1 0 y k 0 1 Managemen Scence/Vol. 47, No. 7, July

4 Modellng Praccal Lo-Szng Problems as Mxed-Ineger Programs from whch each of he sysems can generae he nequaly (15). By combnng several consecuve flow balance consrans wh he varable upper bound consrans hey can hen generae nequales such as (16), and possbly (14). Boh nequales (14) and (16) are members of larger exponenal famles of nequales. For example, he nequales (14) form par of he exponenal famly s k 1 + ( x j d j 1 y ) j k \T j T for all k T wh k and T k. See Barany e al. (1984), and Poche and Wolsey (1993) for a generalzaon of he nequales (16) for he capacaed case. However, should be poned ou ha he way he nal lo-szng model s formulaed may be crucal o wheher he above general mxed-neger programmng sysems succeed n generang cus or no. For nsance, f he flow balance consrans (11) were wren as x j d 1 for all j=1 none of hese sysems s lkely o recognze he exsence of mplc flow conservaon consrans, and so he nequaly (15) wll no be generaed Safey Socks Calculang Ne Demands Suppose now ha here s an nal sock S 0, and ha here are nonzero lower bounds S (safey socks) on he socks a he end of each perod. S 0 + x 1 = d 1 + s 1 s 1 + x = d + s for = 2 NT s S for = 1 NT Now he nequaly (15) s sll vald, bu s unlkely o be volaed because s 1 S 1. Here, he alernave s o rewre he sysem of equaons based on ne demands. Specfcally, frs se S max { } S S 1 d for = 1 NT Then, defne new varables s = s S 0, and ne demands d = d + S S 1. The resulng sysem s s 1 + x = d + s s 0 Wh hs change, he nequaly (15) ha now akes he form s 1 d 1 y may be volaed, even hough he orgnal nequaly s no. A second modfcaon may be approprae f d >C for some. Workng backwards from = NT 2, we see ha s 1 ŝ 1 = d + ŝ C + wh ŝ NT = 0. We hen reapply he same procedure as above o elmnae he lower bounds ŝ on he modfed sock varables s Mullevel Problems Echelon Socks We now consder mullevel problems. For smplcy of noaon we consder an assembly sysem where each em requres exacly one un of each of s predecessors, bu he ransformaon below apples jus as well o problems wh arbrary produc srucure. For an assembly sysem n whch s he successor em of em, and em 1 s he sngle-end em, he flow conservaon equaons ake he form s 1 + x = x + s for = 2 NI all (20) s x1 = d 1 + s1 for all (21) Here, by combnng hese equaons wh he varable upper bound consran x C y, nequaly (14) can be generaed for produc = 1, bu here s no apparen sngle-em lo-szng problem for he oher producs on whch o generae vald nequales. To overcome hs dffculy, remember ha he echelon sock of an em a an nsan n me s he oal sock of ha em whn he sysem, counng s own sock plus he number of uns of he em conaned n he socks of all s successors (Clark and Scarf 1960). So, we nroduce echelon sock varables defned by e 1 = s1 and e = s + e for 2 for each. Now he sysem (20), (21) can be rewren as e 1 + x = d1 + e for all (22) e e for all (23) 996 Managemen Scence/Vol. 47, No. 7, July 2001

5 Modellng Praccal Lo-Szng Problems as Mxed-Ineger Programs where he addonal consrans (23) come from he nonnegavy of s. Now vald nequaly (15), akng he form e 1 d1 1 y can be generaed for each em and no jus for he end Produc 1. In erms of he orgnal varables, hs nequaly s of he form s j 1 d1 1 y j P where P s he se of all ems conanng Iem. 3. Sar-ups and Changeovers In he nex wo secons we explcly examne ways o formulae or reformulae pars of a parcular loszng problem a pror. Frs, we consder sar-ups, swch-offs, and changeovers. We say ha here s a changeover from o j f here s a swch-off of em followed mmedaely be a sar-up of em j. Here, an mporan decson o be made n buldng a model s he sze of he me nerval. If he me nerval s large, leadng o a bg bucke model, a relavely small number of perods may be needed, bu many evens can hen occur whn a me nerval, whch may lead o dffcules n modellng. Alernavely, f he me nervals are shor, leadng o a small bucke model, reang n deal wha happens n a me nerval or beween consecuve me nervals may be easer, bu he number of nervals may become large Small Bucke Models Frs, we consder small bucke models n whch we are lmed o producng one or wo ems per perod. The rade-off beween hese wo modellng opons s obvously problem-dependen. I s possble ha by allowng wo ems per perod, one can somewha ncrease he sze of he perods, and hus decrease he number of perods and he sze of he LP formulaon. On he oher hand, fndng a gh formulaon may be more complcaed. The specfc goal n hs subsecon s o fnd gh formulaons for sar-ups, swch-offs, and changeovers n he one and wo se-up models. The smples modellng soluon s jus o allow one em o be produced per perod on a sngle machne. Here, a producon sequence s very smple. We jus ndcae whch em, f any, s se up n each perod. Thus, f NT = 5, he sequence ndcaes ha he machne s se up o produce Iem 1 n Perod 1, Iem 2 n Perods 2 and 3, Iem 4 n Perod 4, ec. Also, a sar-up of Iem 2 occurs n Perod 2, of Iem 3 n 5, ec. Ths suaon s easly modelled usng consrans (5), (6), and (10) resrced o a sngle machne and he addonal consrans z 1 y 1 for all. When we allow wo ems o be produced per perod, we allow only one produc change per perod. Now a feasble sequence such as ndcaes ha Iems 1 and 2 are se up n Perod 1 n ha order, Iem 2 s se-up n Perod 2, Iems 2 and 4 are se up n Perod 3 n ha order, and so on. Now Iem 2 s sared up n Perod 1 because he machne s se up for Iem 2 a he end of Perod 1, and was se up for anoher em earler n he perod. Smlarly, Iem 4 s sared up n Perod 3. If we rewre he sequence as , we see ha he second em se up n 1 mus always be he frs em se up n. Noe ha hs makes possble for one em o be produced hroughou perod 1 and for wo oher ems and o be produced n perod by usng he se up sequence, bu s mpossble for four dfferen ems o be produced whn wo me perods. I can be shown by nducon ha he 0 1 pons sasfyng (5), and he consrans y 1 + z for all z 1 for all z + z 1 y for all precsely descrbe such wo se-up sequences. To undersand wheher hese formulaons can be mproved, we sudy a model nvolvng sequence dependen changeovers (Karmarkar and Schrage 1985), whch s mporan n s own rgh. Suppose ha here s a dummy em/produc correspondng o he dle sae. We consder he polyhedron Q represenng he flow of a sngle un from Managemen Scence/Vol. 47, No. 7, July

6 Modellng Praccal Lo-Szng Problems as Mxed-Ineger Programs Fgure 1 Small Buckes: One and Two Se-ups = 0 = 1 = 2 = 3 = 4 = 5 = 1 = 2 = 3 = 4 = 1 = 2 = 3 = 4 = 3 = 5 = 1 = 2 = 3 = 4 = 3 = em o em over me, j = q j for j = 1 NI = 1 NT j = q 1 for = 1 NI = 1 NT j q 0 = 1 j 0 for j = 1 NI = 1 NT shown n Fgure 1, where q s he flow hrough node, and j s he flow from node 1 o node j. Ths formulaon represenng a flow of one un (or a pah) s known o be as gh as possble n he sense ha, f no oher consrans are presen, he lnear programmng relaxaon leads o an neger soluon. The correspondng gh formulaon whou he changeover varables j wh j s gven by he followng resul. Proposon 1 (Consanno 1995). proj q Q s he polyhedron = 1 for = 0 NT (24) q + q 1 + j j q q q 1 jj 1 + for = 1 NI = 1 NT (25) for = 1 NI = 1 NT (26) for = 1 NI = 1 NT (27) 0 for = 1 NI = 1 NT (28) We now nerpre hs proposon n he cases of one and wo se-ups per perod One Se-up per Perod To use he polyhedron Q o represen hs suaon, we need o defne he se-up and sar-up varables appropraely. Specfcally, we nerpre a flow hrough node j hrough he frs nework n Fgure 1 o mean ha here s a se-up of em j n perod. Thus, we se y j = q j. Now j = 1fy 1 = yj = 1, namely, s se up n 1 and j n. Wh hese defnons, z = j j j y y +1 s he sar-up varable, and w = j j j s he swch-off varable. = +1 = Usng hese equaons and proj y Q, we oban he y z w gh formulaon for se-ups, sar-ups, and swch-offs: y = 1 for = 0 NT (29) z y for = 1 NI = 1 NT (30) y z = y 1 w 1 for = 1 NI = 1 NT (31) 998 Managemen Scence/Vol. 47, No. 7, July 2001

7 Modellng Praccal Lo-Szng Problems as Mxed-Ineger Programs y 1 + z + y j z j 1 j j z w 0 for = 1 NI = 1 NT (32) for = 1 NI = 1 NT (33) y + ( j y +1 zj +1 ) zj 1 for =1 NI j j z w 0 =1 NT (37) for =1 NI =1 NT (38) where (29) follows from (24); (33) because z and w 1 are he slack varables n (25) and (26), respecvely; (31) by elmnang n he equaons defnng z and w ; (30) from (28) and he defnon of z ; and (32) from (27) afer subsuon. Noe ha nequaly (32) s a cu srenghenng he formulaon gven a he begnnng of hs secon; see Consanno (1995). I says ha he hree possbles em s se up n perod 1, em s sared up n perod, and some em oher han s produced n boh perods 1 and are muually exclusve. where (34) follows from (24), and he defnons of y and z, (35) from (28) and he equaons defnng z and y, (38) as z and w are he slack varables n (25) and (26), respecvely, (36) from usng he defnons of z and w+1 o elmnae q, and hen he defnon of y o elmnae, and (37) from (27) afer approprae subsuons. Noe ha here (37) s a cu srenghenng he formulaon gven a he begnnng of hs secon. I says ha em s se up durng perod and some em oher han s se up hroughou perod are muually exclusve Two Se-ups per Perod We agan use formulaon Q bu wh he second nework n Fgure 1, so he nerpreaon s dfferen. Iems and j are se up n perod n ha order f and only f q 1 = qj = j = 1. So, q j = 1 means ha he machne s se up o produce em j a he end of perod. Wh hese defnons, z = j j j sar-up varable of em n perod, = q s he w = j j j = q 1 s he swch-off varable of em n perod, and y = q 1 +q = q 1 + z = w +z + s he seup varable of em n perod. Usng hese equaons and proj y Q, we oban he followng y z w formulaon: y z = 1 for = 1 NT (34) z + z 1 y for = 1 NI = 1 NT (35) y z = y 1 w 1 for = 1 NI = 1 NT (36) 3.4. Bg Bucke Models wh Changeovers Consder now a dfferen suaon n whch many ems can be se up n each perod, bu changeover coss and/or mes mus be modelled. Suppose for smplcy ha he problem only nvolves one me perod, and he frs and las ems produced n he perod, denoed by = 0 and = n + 1, respecvely, are known and dsnc. Le y be he number of seups of em n he perod, and j be he number of mes producon s changed from o j, wh q an a pror upper bound on he number of mes a gven em s produced n he perod. Thus, he seup sequence 1, 2, 2, 3, 4, 2, 3 s represened by y 1 = y 4 = 1 y 2 = 3 y 3 = 2, and 01 = 12 = 22 = 34 = 42 = 35 = 1 23 = 2. Ths can be modelled n much he same way as he prze collecng ravellng salesman or he vehcle roung problem (Balas 1989) wh 0j = 1 (39) j 0 j = y for 0 n+ 1 (40) j j = y j for j 0 n+ 1 (41) Managemen Scence/Vol. 47, No. 7, July

8 Modellng Praccal Lo-Szng Problems as Mxed-Ineger Programs n+1 n+1 = 1 (42) y j 0 1 q for all j (43) wh addonal consrans o elmnae subours. The dfference here s ha we mus deal wh neger raher han 0 1 varables. The followng resul s easly verfed. Proposon 2. () A vecor y sasfyng he sysem (39) (43) provdes a feasble producon sequence f and only f he nduced dreced mulgraph does no conan a dsconneced dreced Euleran subgraph. () The vald nequaly j E S j y 1 q yk for k S (44) S elmnaes such a mulgraph on he node se S 1 n, where E S s he se of pars j wh j S. 4. Mnmum Producon Runs and Full-Capacy Producon Here we show how o model varous ypcal consrans arsng n small bucke models. Some of hese reformulaons were frs presened n Poche and Wolsey (1996); see also PAMIPS (1995). Mnmum On and Off Tmes. We consder sngleem consrans hroughou hs subsecon, so he produc superscrp s agan dropped. If a machne mus reman se up for an em for a mnmum of perods, we have z y A gher formulaon s for = + 1 z z y (45) or equvalenly w + +w + 1 y. If afer producng an em, a machne canno produce agan whn perods, we have z 1 y for = and agan a gher formulaon s z z + 1 y (46) or equvalenly w 1 + +w 1 y. Full-Capacy Producon. If an em mus be produced a full capacy n all bu he frs and las perods of a se-up sequence (nonpreemped producon bach), we have ha producon s a full capacy n perod f he machne s se up n perods 1 and, and neher s a swch-off perod. Ths gves he consran x C y 1 + y w 1 w 1 A gher formulaon s x C y 1 w 1 w (47) correspondng o he observaon ha full-capacy producon s enforced n f he machne s se up n 1 and s no swched off n perods 1or. Full-Capacy and Mnmum Producon Runs. Suppose ha he mnmum producon quany for some em s P, full capacy s C, and he maxmum capacy n a changeover perod s C, wh C < C<P. Le a = P C /C and b = P/C. Clearly, he leas number of perods requred o produce P s max a + 1 b, and he mos s b + 1. The basc consran x Pw = b can be ghened. Specfcally f a = b, he nequaly x a + x P a 1 C w (48) s vald for all >a, and f a = b 1, he nequaly x b + x a + x P a 1 C w (49) s vald for all >b. Noe also ha f he mnmum producon quany s fully produced durng some nerval + 1 l and exceeds he oal demand d l l = d for he nerval, hen he excess mus be used o sasfy demands n perods ousde he nerval. Ths shows he valdy of he nequales and s l + r 1 P d l + l =+b w (50) l b s l + r 1 P d l + z (51) = 1000 Managemen Scence/Vol. 47, No. 7, July 2001

9 Modellng Praccal Lo-Szng Problems as Mxed-Ineger Programs 5. A Selecon of Problems Here we descrbe some praccal lo-szng nsances, gvng he orgnal formulaon, and some mprovemens ha can be made based on b4 Ths s a mul-em, mulmachne problem wh backloggng, cleanng mes CLT, and lower and upper bounds on socks. I s a small bucke model wh wo se-ups per perod a mos. The specal feaure of he problem s he requremen of lower bounds MB (n number of days) on producon runs for end ems, denoed EP, and full-capacy producon n all bu he frs and las perods of a producon run. There s a small number of nermedae producs, denoed IP, wh a few machnes dedcaed o hem. The coeffcens RML j represen he number of uns of nermedae produc IP requred o produce one un of fnal produc j EP. Noe ha he producon srucure s no of assembly ype, bu he se of end ems s paroned accordng o he unque nermedae em ha each uses. An nal formulaon of hs problem s mn k s 1 + k s 1 r 1 + k ( p k x k + c w k k x k = k x k j k j ) ( + h s + ) e r RML j j k x j k +s for IP for all (52) = d +s r for EP for all (53) x k +CLT w k C k y k for all k (54) x k + CLT w k C k for all k (55) w k 1 y k 1 y k for all k (56) w k +w k +1 y k for all k (57) y k w k 1 for all k (58) w k 1 for all k (59) w k l k x k l l= k k x k y k for EP all k l + k (60) MB w k C k( y k for EP all k (61) 1 +yk w k w k 1 1) for all k (62) s r x 0 y w 0 1 s 0 = S 0 r 0 =0 s s s where k = MB /C k, and k = MB + CLT C k /C k. Noe ha (57) (59) represen wo se-ups per perod, (60) a mnmum run lengh consran, (61) he mnmum bach consran, and (62) he full-capacy producon consran. A reformulaon of hs problem based on 3 and 4s mn ( ) ( p k x k + c w k + h s + ) e r s 1 + k s 1 r 1 + k k x k = k x k j k j RML j j k x j k +s for IP for all (63) = d +s r for EP for all (64) x k +CLT w k C k y k for all k (65) x k + CLT w k C k for all k (66) y k k + k l= x k l l= k w k 1 y k 1 y k for all k (67) w k 1 for all k (68) w k l +x k x k y k for EP all k (69) ( MB / k ( k 1 ) C k) w k for EP all k (70) C k( ) y k 1 w k w k 1 Managemen Scence/Vol. 47, No. 7, July

10 Modellng Praccal Lo-Szng Problems as Mxed-Ineger Programs for all k (71) r 1 +s l ( ) MB d + l l w k k =+b for k l wh +b l (72) ( l l s 1 + r d 1 1 ) w k u y k = = k u= k s r x 0 y w 0 1 for k l (73) s 0 = S 0 r 0 =0 s s s Noe ha (69) s from he nequaly (45); (70) from ; (71) from (47); (72) from (50); and (73) from a combnaon of (17) and (18). Resuls for nsances wh beween 10 and 20 perods, 22 ems and 7 machnes are presened n Belvaux and Wolsey (2000). Daa and model fles for hs problem are avalable a Belvaux and Wolsey (1999) Chesapeake The CHES problems are a se of fve ndusral problems nvolvng he allocaon and sequencng of producon operaons wh sequence dependen se-up coss on connuous parallel facles. They are based on real problems (CHES 1989). We use he formulaon from he orgnal paper (CHES 1989) wh he addon of he generalzed subour elmnaon consrans proposed n 3. Here, y k s modfed o denoe he number of se-ups of em on machne k n perod wh upper bound q. jk s he number of ransons from producon of o producon of j n on machne k, whereas jk = 1 f he las producon on machne k n s and he frs producon n + 1sj. k = 1 f he las em produced on machne k s. Y jk 0 s daa akng value 1 f machne k s nally se up for em j, and value 0 oherwse. The formulaon proposed s mn h s + k s 1 + k p xk + f y + j k q jk( jk x k = d + s for all k x k C k y k for all k + jk ) jk + jk + j j j x k LB k yk for all a k x k L k for all k s x 0 y 0 1 q jk + Y jk 0 = y jk jk 1 = yjk jk = y k jk + k = y k j E S jk for = 1 for >1 for <NT for = NT y k 1 q yl k for l S S S 1 NI all Compuaonal resuls have been obaned by runnng BC-PROD n branch-and-bound mode wh he formulaon excludng he subour consrans. When an neger soluon s found, he sysem checks f represens a feasble machne sequence, and f no, a generalzed subour elmnaon consran s added, and he ree search connues. Such an approach s easly mplemenable wh sysems such as MINTO, or wh he subroune lbrares of CPLEX or XPRESS. A beer approach would be o also generae consrans o cu off fraconal soluons. Resuls on he unsolved problems of he orgnal daa se are presened n Table 1 (he resuls repored on n Belvaux and Wolsey (2000) are for slghly modfed daa). The frs seven columns descrbe he nsance: s name, he number of Iems NI, he number of Machnes NK, he number of Perods NT, he number of rows r, columns (c), and neger varables n n he MIP formulaon. The las four columns presen resuls: LB denoes he value of he bes lower bound. BIP he value of he bes feasble soluon, Secs he runnng me n seconds (wh a maxmum of 3,600), and GAP = 100 BIP LB / LB he dualy gap on ermnaon. All runs have been carred ou on a Penum 200 Mhz runnng under Wndows NT. Noe ha, apar from he bes feasble soluon for ches5, hese mprove on he resuls n Kang e al. (1999) obaned usng an orgnal approach, conssng 1002 Managemen Scence/Vol. 47, No. 7, July 2001

11 Modellng Praccal Lo-Szng Problems as Mxed-Ineger Programs Table 1 Resuls for Unsolved CHES Problems Insance NI NK NT r c n LB BIP Secs Gap ches % ches % ches ches % of breakng he producon sequence durng a perod no a lmed number of spl sequences each conanng 3 5 ems. By enumeraon, an opmal orderng can be found for each spl sequence. An MIP s hen solved n whch hese spl sequences are joned o produce feasble producon schedules, whle column generaon and heurscs are used o generae neresng spl sequences SmpEreng In Smpson and Erenguc (1998b), a general model for mullevel problems wh assembly produc srucure s proposed, nvolvng produc famles conssng of one or more ems, where each famly can n urn have a fxed cos, a se-up me, or a resource consran assocaed wh. The formulaon proposed s F f a f x + g V f mn h s + c f f f s 1 + x = x + s + d for all x My for all y f y f 0 1 x s 0 for all f wh F f gf g C f for all f for all f where he new varable f akes he value 1 f some em n famly f s se up n perod, s he mmedae successor of produc. F f s he se of ems n famly f. M s a large posve value. C f s he resource avalable for famly f n perod. a f s he rae a whch produc uses he resource assocaed o famly f. V f s he se of famles occurrng n he famly f budge consran. gf s he fxed amoun of he resource of famly f used f famly g s se up. Two observaons can be used o modfy hs formulaon. The frs s o use echelon socks as suggesed n 2. The second s ha he se-up varables y are unnecessary as here are no se-up coss. Ths leads o he followng revsed formulaon wh many fewer neger varables: F f mn h s + c f f f e 1 + x = dq + e for all e e for all x M f for all f wh F f a f x + gf g C f f for all f g V f f 0 1 x s 0 for all f where q s he fnal produc conanng em. Insances s NI v are versons of wo praccal problems wh NI ems, a sngle machne, and NT = 16 me perods. Problem s 78 v s a sorage rack producon problem wh 6 end producs, whereas problem s 80 v s an anmal feed packng problem wh 8 feeds, each packed n 2 conaner ypes, makng 16 end producs. All he daa are provded excep for he demand daa for he end producs for whch dscree probably dsrbuons are proposed. We gve resuls for wo dfferen approaches n Table 2. The frs s o run he reformulaed problem wh he general purpose sysem BC-OPT. The second s o run he orgnal formulaon wh he specalzed lo-szng sysem BC-PROD. We are no able o run he reformulaed problem wh BC-PROD as he sysem does no a presen recognze produc Managemen Scence/Vol. 47, No. 7, July

12 Modellng Praccal Lo-Szng Problems as Mxed-Ineger Programs Table 2 Resuls for Smpson and Erenguc Problems Insance r c n LP XLP BIP Secs Gap s-78-lec s-78-2ec s-80-1ec s-80-2ec s s s s famles. In Table 2, LP denoes he nal LP value, XLP he value afer he addon of cus, BIP he value of he bes feasble soluon found, and GAP = BIP BLB /BIP 100, where BLB s he value of he bes lower bound on ermnaon. The runs have been carred ou on a Penum 350 Mhz runnng under Wndows NT. For he s NI v models he gaps of 10% 20% may appear large. However, n Smpson and Erenguc (1998b), where s suggesed ha local search heursc mehods should replace relaxaon-based mehods, only average resuls are repored and he gaps are much larger L1 Ths problem nvolves capacy consrans, desrable sock levels, backloggng, mnmum run lenghs, and producon of ems from only one famly durng a gven perod on each machne. The coss are for backloggng, socks and for socks fallng below he safey sock level. S s he safey sock level for Iem n Perod, and LB k s a lower bound on he amoun of produced on k n f s produced. The varable bs measures he amoun by whch he sock when posve falls below he safey sock level. s 1 r 1 + k x k = d + s r for all (74) bs + s S for all (75) k x k B k for all k (76) LB k yk x k C k y k f y k for all k (77) fk for all k wh F f (78) fk 1 for all k (79) y k fk 0 1 s r bs xk 0 for all k f Two observaons can be used o modfy hs formulaon. The frs s o change he base sock level, much as n 2.2. Specfcally, f we se he base sock level a S raher han 0 by defnng s = bs + s S 0, and r = bs + r 0, hen s s now he amoun by whch he sock exceeds he safey sock level S, and r s he amoun by whch he sock s below S. Now seng d = d S 1 + S, he flow balance equaon (74) becomes s 1 r 1 + k x k = d + s r and he nequaly ( s 1 + r d 1 ) y k k whch s a generalzaon of (15), may be volaed. The second s algorhmc. Even hough he famly se-up varables do no need o be se o be negral because of he negraly of he se-up varables y k and he forcng consrans (78), s perhaps neresng o make hem neger, and hen use prores o branch on he famly se-up varables before branchng on he se-up varables for he ndvdual ems Managemen Scence/Vol. 47, No. 7, July 2001

13 Modellng Praccal Lo-Szng Problems as Mxed-Ineger Programs Fgure 2 Se and Sales Area v l v l s l x k = l L l l LL l sh l l + bb l for all l (82) d l for all l (83) sh ll +1 for all l (84) = B k y k for all k (85) x 0 y 0 and neger (86) 6. A Producon-Dsrbuon Bg Bucke Problem Ths s a mulse, mulproduc problem n whch producon shpmen and sales mus be deermned over a 3-monh (perod) plannng horzon. The sysem consss of 4 producon ses comprsng 15 producon uns and 150 producs, and 30 sales areas. The flows a he producon ses and sales areas are shown n Fgure 2. Parculares of he problem are producs are socked only a he producon ses, he nvenory a he end of a me perod mus cover a ceran fracon of he sales n he succeedng perod (because long me perods are used and producon s no nsananeous), here s ransporaon beween ceran producon ses and from producon ses o sales areas, bu no beween sales areas, some producs are manufacured n mulples of a fxed bach sze, whereas for ohers here are mnmum producon quanes, lower bounds are gven on he sales per em a each sales area for each perod. A formulaon of a slghly smplfed problem s s l 1 + bl + x k + sh l l k K l l l l = s l + l l l sh ll for all l (80) x k C k for all k (81) wh he followng daa: s he fracon of sales ha mus be n sock n he prevous perod. d l s he lower bound on sales of n sales area l n perod. B k s he bach sze for em n un k. K l s he se of uns k such ha em s produceable on un k a producon se l. L l s he se of producon ses. LL l s he se of sales areas. The varables are s l s he sock of n producon se l a he end of perod. b l s he amoun of bough a producon se l n perod. sh ll s he amoun of shpped from l o l n perod. bb l s he amoun of bough a sales area l n perod. v l s he amoun of sold a sales area l n perod. y k s he number of baches of produced on un k n. Consrans (80) and (82) represen flow conservaon a he producon ses and sales areas, respecvely, (83) represens he lower bound on sales, (84) he sock avalably condon, and (81) he capacy of each producon un. To oban effecve vald nequales, we aggregae ems over ses and sales areas. Summng (80) over l, and nroducng aggregaed varables s sh b, ec., leads o s 1 + x k + b = s + sh (87) k Summng (82) and (83) over l gves where u = l bb l. sh + u d (88) Managemen Scence/Vol. 47, No. 7, July

14 Modellng Praccal Lo-Szng Problems as Mxed-Ineger Programs Aggregang (84) gves s sh +1 (89) Now combnng 87 88, and (89) gves s 1 + b + u + u +1 + k x k d + d +1 (90) Fnally, usng (85) and defnng a varable = s 1 + b +u + u +1 0 and leng D = d + d +1, we oban + B y k D k For hs we oban he mxed-neger roundng nequaly (Nemhauser and Wolsey 1988) ( ) y k k where = D/B, and = D B 1. Smlar nequales are obaned for ems wh mnmum producon runs. In Table 3 we presen compuaonal resuls for four ypcal nsances. We compare resuls usng he nal formulaon (mod1), mod1 plus all he aggregaed MIR nequales (mod2), and mod1 wh he aggregaed MIR nequales acve a he op node (mod3). Resuls were obaned on a Penum Pro 200 Mhz runnng under Wndows NT wh he XPRESS opmser mp-op A me lm of 900 seconds s used for each nsance, and he opmsaon sraegy consss of a bes-bound node selecon sraegy for 127 nodes, followed by he mp-op defaul based on a choce among he wo descendan nodes. Ths lm s mposed n pracce as he model s run durng plannng meengs every monh. 7. Remarks The resuls n Belvaux and Wolsey (2000) and here sugges ha reformulaon and cung planes combned wh mxed-neger programmng provde an effecve ool for acklng a varey of praccal loszng problems. The undoubed srengh of he approach les n he qualy of he dual (lower) bounds provded by he ghened relaxaons. To fnd good feasble soluons gvng prmal (upper) bounds, here are many more possble approaches, and a general MIP sysem does no properly make use of problem srucure eher o branch or o search for feasble soluons. I does, however, have as sarng pon he lnear programmng soluon from he srenghened formulaon, so heurscs ncorporang hs soluon and problem srucure deserve furher exploraon. For dscree lo-szng problems whou realvalued producon varables nvolvng sar-up or changeover coss and mes (Fleschmann 1994), he reformulaons presened here reman mporan. However, he large number of varables needed o model changeovers s an mporan drawback. For such problems research no a combnaon of consran programmng and mxed-neger programmng appears an neresng drecon. Table 3 Resuls for Four Typcal Insances Insance r c LP IP Gap bm1mod bm1mod bm1mod bm2mod bm2mod bm2mod bm3mod bm3mod bm3mod bm4mod bm4mod bm4mod Managemen Scence/Vol. 47, No. 7, July 2001

15 Modellng Praccal Lo-Szng Problems as Mxed-Ineger Programs Acknowledgmens The auhors are graeful o BASF for allowng hem o ce he formulaon b4 and he Producon-Dsrbuon problem from he PAMIPS (1995) and MEMIPS (1997) projecs, respecvely, and o Procer and Gamble for allowng hem o presen he formulaon L1. References Afenaks, P., B. Gavsh Opmal lo-szng for complex produc srucures. Oper. Res Balas, E The prze collecng ravelng salesman problem. Neworks Barany, I., T. J. Van Roy, L. A. Wolsey Srong formulaons for mul-em capacaed lo-szng. Managemen Sc Belvaux, G Modellng and solvng lo-szng problems by mxed-neger programmng. Ph.D. hess, Faculé des Scences Applquées, Unversé Caholque de Louvan, Belgum., L. A. Wolsey BC-PROD: A specalzed branch-andcu sysem for lo-szng problems. Managemen Sc ,. LOTSIZELIB: A lbrary of lo-szng models and marces. Avalable a hp:// Loszel.hm. Carysse, D., J. Maes, L. N. van Wassenhove Se paronng and column generaon heurscs for capacaed dynamc loszng. Eur. J. Oper. Res , M. Salomon. R. Kuk, L. N. van Wassenhove A dual ascen and column generaon heursc for he DLSP wh seup mes. Managemen Sc The CHES problems. Chesapeake Decson Scences, Inc., New Provdence, NJ. Clark, A. J., H. Scarf Opmal polces for mul-echelon nvenory problems. Managemen Sc Consanno, M A polyhedral approach o producon plannng models: Sar-up coss and mes, and lower bounds on producon. Ph.D hess, Deparmen of Mahemacs, Unversé Caholque de Louvan, Belgum A cung plane approach o capacaed lo-szng wh sar-up coss. Mah. Programmng Corder, C., H. Marchand, R. Laundry, L. A. Wolsey bc op: A branch-and-cu code for mxed-neger programs. Mah. Programmng Daby, M., H. C. Bahl, M. H. Karwan, S. Zons A lagrangean relaxaon approach for very large scale capacaed lo-szng. Managemen Sc Dxon, P. S., E. A. Slver A heursc soluon procedure for he mul-em, sngle level, lmed capacy lo-szng problem. J. Oper. Managemen Eppen, G. D., R. K. Marn Solvng mul-em loszng problems usng varable redefnon. Oper. Res Fleschmann, B The dscree lo-szng and schedulng problem wh sequence-dependen seup coss. Eur. J. Oper. Res Kang, S., K. Malk, L. J. Thomas Loszng and schedulng n parallel machnes wh sequence-dependen seup coss. Managemen Sc Karmarkar, U. S., L. S. Schrage The deermnsc dynamc produc cyclng problem. Oper. Res MEMIPS. Model enhanced soluon mehods for neger programmng sofware. Espr Projec 20118, Publc repor reference DR Nemhauser, G. L., L. A. Wolsey Ineger and Combnaoral Opmzaon. John Wley and Sons, New York PAMIPS. Developmen of parallel algorhms and sofware for mxed-neger programmng n ndusral schedulng. Espr Projec 8755, Publc repor reference DR Poche, Y., L. A. Wolsey Solvng mul-em lo-szng problems usng srong cung planes. Managemen Sc , Lo-szng wh consan baches: Formulaon and vald nequales. Mah. Oper. Res , Algorhms and reformulaons for lo-szng problems. W. Cook, L. Lovasz, P. Seymour, eds. Combnaoral Opmzaon. DIMACS Seres n Dscree Mahemacs and Compuer Scence , Addng flexbly n lo-szng models: Mnmum bach szes and cleanng mes. Workshop on Producon Plannng and Conrol, FUCAM, Mons, Belgum. Savelsbergh, M. W. P., G. L. Nemhauser Funconal descrpon of MINTO, a mxed-neger opmzer. Repor COC-91-03A, Georga Insue of Technology, Alana, Georga. Smpson, N. C., S. S. Erenguc. 1998a. Improved heursc mehods for mulple sage producon plannng. Comp. Oper. Res ,. 1998b. Producon plannng n mulple sage manufacurng envronmens wh jon coss, lmed resources and se-up mes. Techncal repor, Deparmen of Managemen Scence and Sysems, Unversy of Buffalo, Buffalo, NY. Tempelmeer, H., M. Dersroff A lagrangean-based heursc for dynamc mullevel mulem consraned loszng wh seup mes. Managemen Sc Thzy, J. M., L. N. van Wassenhove Lagrangean relaxaon for he mul-em capacaed lo-szng problem. IIE Trans Vanderbeck, F Lo-szng wh sar-up mes. Managemen Sc Acceped by Th. M. Leblng; receved February Ths paper was wh he auhors 3 monhs for 1 revson. Managemen Scence/Vol. 47, No. 7, July

Modeling and Solving of Multi-Product Inventory Lot-Sizing with Supplier Selection under Quantity Discounts

Modeling and Solving of Multi-Product Inventory Lot-Sizing with Supplier Selection under Quantity Discounts nernaonal ournal of Appled Engneerng Research SSN 0973-4562 Volume 13, Number 10 (2018) pp. 8708-8713 Modelng and Solvng of Mul-Produc nvenory Lo-Szng wh Suppler Selecon under Quany Dscouns Naapa anchanaruangrong

More information

A Tour of Modeling Techniques

A Tour of Modeling Techniques A Tour of Modelng Technques John Hooker Carnege Mellon Unversy EWO Semnar February 8 Slde Oulne Med neger lnear (MILP) modelng Dsuncve modelng Knapsack modelng Consran programmng models Inegraed Models

More information

V.Abramov - FURTHER ANALYSIS OF CONFIDENCE INTERVALS FOR LARGE CLIENT/SERVER COMPUTER NETWORKS

V.Abramov - FURTHER ANALYSIS OF CONFIDENCE INTERVALS FOR LARGE CLIENT/SERVER COMPUTER NETWORKS R&RATA # Vol.) 8, March FURTHER AALYSIS OF COFIDECE ITERVALS FOR LARGE CLIET/SERVER COMPUTER ETWORKS Vyacheslav Abramov School of Mahemacal Scences, Monash Unversy, Buldng 8, Level 4, Clayon Campus, Wellngon

More information

Online Supplement for Dynamic Multi-Technology. Production-Inventory Problem with Emissions Trading

Online Supplement for Dynamic Multi-Technology. Production-Inventory Problem with Emissions Trading Onlne Supplemen for Dynamc Mul-Technology Producon-Invenory Problem wh Emssons Tradng by We Zhang Zhongsheng Hua Yu Xa and Baofeng Huo Proof of Lemma For any ( qr ) Θ s easy o verfy ha he lnear programmng

More information

The Dynamic Programming Models for Inventory Control System with Time-varying Demand

The Dynamic Programming Models for Inventory Control System with Time-varying Demand The Dynamc Programmng Models for Invenory Conrol Sysem wh Tme-varyng Demand Truong Hong Trnh (Correspondng auhor) The Unversy of Danang, Unversy of Economcs, Venam Tel: 84-236-352-5459 E-mal: rnh.h@due.edu.vn

More information

On One Analytic Method of. Constructing Program Controls

On One Analytic Method of. Constructing Program Controls Appled Mahemacal Scences, Vol. 9, 05, no. 8, 409-407 HIKARI Ld, www.m-hkar.com hp://dx.do.org/0.988/ams.05.54349 On One Analyc Mehod of Consrucng Program Conrols A. N. Kvko, S. V. Chsyakov and Yu. E. Balyna

More information

Solution in semi infinite diffusion couples (error function analysis)

Solution in semi infinite diffusion couples (error function analysis) Soluon n sem nfne dffuson couples (error funcon analyss) Le us consder now he sem nfne dffuson couple of wo blocks wh concenraon of and I means ha, n a A- bnary sysem, s bondng beween wo blocks made of

More information

GENERATING CERTAIN QUINTIC IRREDUCIBLE POLYNOMIALS OVER FINITE FIELDS. Youngwoo Ahn and Kitae Kim

GENERATING CERTAIN QUINTIC IRREDUCIBLE POLYNOMIALS OVER FINITE FIELDS. Youngwoo Ahn and Kitae Kim Korean J. Mah. 19 (2011), No. 3, pp. 263 272 GENERATING CERTAIN QUINTIC IRREDUCIBLE POLYNOMIALS OVER FINITE FIELDS Youngwoo Ahn and Kae Km Absrac. In he paper [1], an explc correspondence beween ceran

More information

Mechanics Physics 151

Mechanics Physics 151 Mechancs Physcs 5 Lecure 0 Canoncal Transformaons (Chaper 9) Wha We Dd Las Tme Hamlon s Prncple n he Hamlonan formalsm Dervaon was smple δi δ Addonal end-pon consrans pq H( q, p, ) d 0 δ q ( ) δq ( ) δ

More information

Variants of Pegasos. December 11, 2009

Variants of Pegasos. December 11, 2009 Inroducon Varans of Pegasos SooWoong Ryu bshboy@sanford.edu December, 009 Youngsoo Cho yc344@sanford.edu Developng a new SVM algorhm s ongong research opc. Among many exng SVM algorhms, we wll focus on

More information

[ ] 2. [ ]3 + (Δx i + Δx i 1 ) / 2. Δx i-1 Δx i Δx i+1. TPG4160 Reservoir Simulation 2018 Lecture note 3. page 1 of 5

[ ] 2. [ ]3 + (Δx i + Δx i 1 ) / 2. Δx i-1 Δx i Δx i+1. TPG4160 Reservoir Simulation 2018 Lecture note 3. page 1 of 5 TPG460 Reservor Smulaon 08 page of 5 DISCRETIZATIO OF THE FOW EQUATIOS As we already have seen, fne dfference appromaons of he paral dervaves appearng n he flow equaons may be obaned from Taylor seres

More information

( ) () we define the interaction representation by the unitary transformation () = ()

( ) () we define the interaction representation by the unitary transformation () = () Hgher Order Perurbaon Theory Mchael Fowler 3/7/6 The neracon Represenaon Recall ha n he frs par of hs course sequence, we dscussed he chrödnger and Hesenberg represenaons of quanum mechancs here n he chrödnger

More information

Solving the multi-period fixed cost transportation problem using LINGO solver

Solving the multi-period fixed cost transportation problem using LINGO solver Inernaonal Journal of Pure and Appled Mahemacs Volume 119 No. 12 2018, 2151-2157 ISSN: 1314-3395 (on-lne verson) url: hp://www.pam.eu Specal Issue pam.eu Solvng he mul-perod fxed cos ransporaon problem

More information

Testing a new idea to solve the P = NP problem with mathematical induction

Testing a new idea to solve the P = NP problem with mathematical induction Tesng a new dea o solve he P = NP problem wh mahemacal nducon Bacground P and NP are wo classes (ses) of languages n Compuer Scence An open problem s wheher P = NP Ths paper ess a new dea o compare he

More information

Notes on the stability of dynamic systems and the use of Eigen Values.

Notes on the stability of dynamic systems and the use of Eigen Values. Noes on he sabl of dnamc ssems and he use of Egen Values. Source: Macro II course noes, Dr. Davd Bessler s Tme Seres course noes, zarads (999) Ineremporal Macroeconomcs chaper 4 & Techncal ppend, and Hamlon

More information

P R = P 0. The system is shown on the next figure:

P R = P 0. The system is shown on the next figure: TPG460 Reservor Smulaon 08 page of INTRODUCTION TO RESERVOIR SIMULATION Analycal and numercal soluons of smple one-dmensonal, one-phase flow equaons As an nroducon o reservor smulaon, we wll revew he smples

More information

European Journal of Operational Research

European Journal of Operational Research European Journal of Operaonal Research 267 2018 86 95 Conens lss avalable a ScenceDrec European Journal of Operaonal Research journal homepage: www.elsever.com/locae/ejor Producon, Manufacurng and Logscs

More information

Mechanics Physics 151

Mechanics Physics 151 Mechancs Physcs 5 Lecure 9 Hamlonan Equaons of Moon (Chaper 8) Wha We Dd Las Tme Consruced Hamlonan formalsm H ( q, p, ) = q p L( q, q, ) H p = q H q = p H = L Equvalen o Lagrangan formalsm Smpler, bu

More information

Dynamic Team Decision Theory. EECS 558 Project Shrutivandana Sharma and David Shuman December 10, 2005

Dynamic Team Decision Theory. EECS 558 Project Shrutivandana Sharma and David Shuman December 10, 2005 Dynamc Team Decson Theory EECS 558 Proec Shruvandana Sharma and Davd Shuman December 0, 005 Oulne Inroducon o Team Decson Theory Decomposon of he Dynamc Team Decson Problem Equvalence of Sac and Dynamc

More information

Mechanics Physics 151

Mechanics Physics 151 Mechancs Physcs 5 Lecure 9 Hamlonan Equaons of Moon (Chaper 8) Wha We Dd Las Tme Consruced Hamlonan formalsm Hqp (,,) = qp Lqq (,,) H p = q H q = p H L = Equvalen o Lagrangan formalsm Smpler, bu wce as

More information

Tight results for Next Fit and Worst Fit with resource augmentation

Tight results for Next Fit and Worst Fit with resource augmentation Tgh resuls for Nex F and Wors F wh resource augmenaon Joan Boyar Leah Epsen Asaf Levn Asrac I s well known ha he wo smple algorhms for he classc n packng prolem, NF and WF oh have an approxmaon rao of

More information

CS434a/541a: Pattern Recognition Prof. Olga Veksler. Lecture 4

CS434a/541a: Pattern Recognition Prof. Olga Veksler. Lecture 4 CS434a/54a: Paern Recognon Prof. Olga Veksler Lecure 4 Oulne Normal Random Varable Properes Dscrmnan funcons Why Normal Random Varables? Analycally racable Works well when observaon comes form a corruped

More information

Graduate Macroeconomics 2 Problem set 5. - Solutions

Graduate Macroeconomics 2 Problem set 5. - Solutions Graduae Macroeconomcs 2 Problem se. - Soluons Queson 1 To answer hs queson we need he frms frs order condons and he equaon ha deermnes he number of frms n equlbrum. The frms frs order condons are: F K

More information

The preemptive resource-constrained project scheduling problem subject to due dates and preemption penalties: An integer programming approach

The preemptive resource-constrained project scheduling problem subject to due dates and preemption penalties: An integer programming approach Journal of Indusral Engneerng 1 (008) 35-39 The preempve resource-consraned projec schedulng problem subjec o due daes and preempon penales An neger programmng approach B. Afshar Nadjaf Deparmen of Indusral

More information

Lecture 18: The Laplace Transform (See Sections and 14.7 in Boas)

Lecture 18: The Laplace Transform (See Sections and 14.7 in Boas) Lecure 8: The Lalace Transform (See Secons 88- and 47 n Boas) Recall ha our bg-cure goal s he analyss of he dfferenal equaon, ax bx cx F, where we emloy varous exansons for he drvng funcon F deendng on

More information

Reactive Methods to Solve the Berth AllocationProblem with Stochastic Arrival and Handling Times

Reactive Methods to Solve the Berth AllocationProblem with Stochastic Arrival and Handling Times Reacve Mehods o Solve he Berh AllocaonProblem wh Sochasc Arrval and Handlng Tmes Nsh Umang* Mchel Berlare* * TRANSP-OR, Ecole Polyechnque Fédérale de Lausanne Frs Workshop on Large Scale Opmzaon November

More information

EEL 6266 Power System Operation and Control. Chapter 5 Unit Commitment

EEL 6266 Power System Operation and Control. Chapter 5 Unit Commitment EEL 6266 Power Sysem Operaon and Conrol Chaper 5 Un Commmen Dynamc programmng chef advanage over enumeraon schemes s he reducon n he dmensonaly of he problem n a src prory order scheme, here are only N

More information

A Principled Approach to MILP Modeling

A Principled Approach to MILP Modeling A Prncpled Approach o MILP Modelng John Hooer Carnege Mellon Unvers Augus 008 Slde Proposal MILP modelng s an ar, bu need no be unprncpled. Slde Proposal MILP modelng s an ar, bu need no be unprncpled.

More information

Comb Filters. Comb Filters

Comb Filters. Comb Filters The smple flers dscussed so far are characered eher by a sngle passband and/or a sngle sopband There are applcaons where flers wh mulple passbands and sopbands are requred Thecomb fler s an example of

More information

THE PREDICTION OF COMPETITIVE ENVIRONMENT IN BUSINESS

THE PREDICTION OF COMPETITIVE ENVIRONMENT IN BUSINESS THE PREICTION OF COMPETITIVE ENVIRONMENT IN BUSINESS INTROUCTION The wo dmensonal paral dfferenal equaons of second order can be used for he smulaon of compeve envronmen n busness The arcle presens he

More information

Relative controllability of nonlinear systems with delays in control

Relative controllability of nonlinear systems with delays in control Relave conrollably o nonlnear sysems wh delays n conrol Jerzy Klamka Insue o Conrol Engneerng, Slesan Techncal Unversy, 44- Glwce, Poland. phone/ax : 48 32 37227, {jklamka}@a.polsl.glwce.pl Keywor: Conrollably.

More information

HIERARCHICAL DECISIONS FOR LINEAR/NON-LINEAR DISJUNCTIVE PROBLEMS

HIERARCHICAL DECISIONS FOR LINEAR/NON-LINEAR DISJUNCTIVE PROBLEMS 2 nd Mercosur Congress on Chemcal Engneerng 4 h Mercosur Congress on Process Sysems Engneerng HIERARCHICAL DECISIONS FOR LINEAR/NON-LINEAR DISJUNCTIVE PROLEMS Jorge M. Monagna and Aldo R. Vecche * INGAR

More information

Outline. Probabilistic Model Learning. Probabilistic Model Learning. Probabilistic Model for Time-series Data: Hidden Markov Model

Outline. Probabilistic Model Learning. Probabilistic Model Learning. Probabilistic Model for Time-series Data: Hidden Markov Model Probablsc Model for Tme-seres Daa: Hdden Markov Model Hrosh Mamsuka Bonformacs Cener Kyoo Unversy Oulne Three Problems for probablsc models n machne learnng. Compung lkelhood 2. Learnng 3. Parsng (predcon

More information

John Geweke a and Gianni Amisano b a Departments of Economics and Statistics, University of Iowa, USA b European Central Bank, Frankfurt, Germany

John Geweke a and Gianni Amisano b a Departments of Economics and Statistics, University of Iowa, USA b European Central Bank, Frankfurt, Germany Herarchcal Markov Normal Mxure models wh Applcaons o Fnancal Asse Reurns Appendx: Proofs of Theorems and Condonal Poseror Dsrbuons John Geweke a and Gann Amsano b a Deparmens of Economcs and Sascs, Unversy

More information

CS286.2 Lecture 14: Quantum de Finetti Theorems II

CS286.2 Lecture 14: Quantum de Finetti Theorems II CS286.2 Lecure 14: Quanum de Fne Theorems II Scrbe: Mara Okounkova 1 Saemen of he heorem Recall he las saemen of he quanum de Fne heorem from he prevous lecure. Theorem 1 Quanum de Fne). Le ρ Dens C 2

More information

Comparison of Differences between Power Means 1

Comparison of Differences between Power Means 1 In. Journal of Mah. Analyss, Vol. 7, 203, no., 5-55 Comparson of Dfferences beween Power Means Chang-An Tan, Guanghua Sh and Fe Zuo College of Mahemacs and Informaon Scence Henan Normal Unversy, 453007,

More information

UNIVERSITAT AUTÒNOMA DE BARCELONA MARCH 2017 EXAMINATION

UNIVERSITAT AUTÒNOMA DE BARCELONA MARCH 2017 EXAMINATION INTERNATIONAL TRADE T. J. KEHOE UNIVERSITAT AUTÒNOMA DE BARCELONA MARCH 27 EXAMINATION Please answer wo of he hree quesons. You can consul class noes, workng papers, and arcles whle you are workng on he

More information

ABSTRACT. KEYWORDS Hybrid, Genetic Algorithm, Shipping, Dispatching, Vehicle, Time Windows INTRODUCTION

ABSTRACT. KEYWORDS Hybrid, Genetic Algorithm, Shipping, Dispatching, Vehicle, Time Windows INTRODUCTION A HYBRID GENETIC ALGORITH FOR A DYNAIC INBOUND ORDERING AND SHIPPING AND OUTBOUND DISPATCHING PROBLE WITH HETEROGENEOUS VEHICLE TYPES AND DELIVERY TIE WINDOWS by Byung Soo Km, Woon-Seek Lee, and Young-Seok

More information

. The geometric multiplicity is dim[ker( λi. number of linearly independent eigenvectors associated with this eigenvalue.

. The geometric multiplicity is dim[ker( λi. number of linearly independent eigenvectors associated with this eigenvalue. Lnear Algebra Lecure # Noes We connue wh he dscusson of egenvalues, egenvecors, and dagonalzably of marces We wan o know, n parcular wha condons wll assure ha a marx can be dagonalzed and wha he obsrucons

More information

A Deterministic Algorithm for Summarizing Asynchronous Streams over a Sliding Window

A Deterministic Algorithm for Summarizing Asynchronous Streams over a Sliding Window A Deermnsc Algorhm for Summarzng Asynchronous Sreams over a Sldng ndow Cosas Busch Rensselaer Polyechnc Insue Srkana Trhapura Iowa Sae Unversy Oulne of Talk Inroducon Algorhm Analyss Tme C Daa sream: 3

More information

. The geometric multiplicity is dim[ker( λi. A )], i.e. the number of linearly independent eigenvectors associated with this eigenvalue.

. The geometric multiplicity is dim[ker( λi. A )], i.e. the number of linearly independent eigenvectors associated with this eigenvalue. Mah E-b Lecure #0 Noes We connue wh he dscusson of egenvalues, egenvecors, and dagonalzably of marces We wan o know, n parcular wha condons wll assure ha a marx can be dagonalzed and wha he obsrucons are

More information

Linear Response Theory: The connection between QFT and experiments

Linear Response Theory: The connection between QFT and experiments Phys540.nb 39 3 Lnear Response Theory: The connecon beween QFT and expermens 3.1. Basc conceps and deas Q: ow do we measure he conducvy of a meal? A: we frs nroduce a weak elecrc feld E, and hen measure

More information

Volatility Interpolation

Volatility Interpolation Volaly Inerpolaon Prelmnary Verson March 00 Jesper Andreasen and Bran Huge Danse Mares, Copenhagen wan.daddy@danseban.com brno@danseban.com Elecronc copy avalable a: hp://ssrn.com/absrac=69497 Inro Local

More information

Robustness Experiments with Two Variance Components

Robustness Experiments with Two Variance Components Naonal Insue of Sandards and Technology (NIST) Informaon Technology Laboraory (ITL) Sascal Engneerng Dvson (SED) Robusness Expermens wh Two Varance Componens by Ana Ivelsse Avlés avles@ns.gov Conference

More information

TSS = SST + SSE An orthogonal partition of the total SS

TSS = SST + SSE An orthogonal partition of the total SS ANOVA: Topc 4. Orhogonal conrass [ST&D p. 183] H 0 : µ 1 = µ =... = µ H 1 : The mean of a leas one reamen group s dfferen To es hs hypohess, a basc ANOVA allocaes he varaon among reamen means (SST) equally

More information

Performance Analysis for a Network having Standby Redundant Unit with Waiting in Repair

Performance Analysis for a Network having Standby Redundant Unit with Waiting in Repair TECHNI Inernaonal Journal of Compung Scence Communcaon Technologes VOL.5 NO. July 22 (ISSN 974-3375 erformance nalyss for a Nework havng Sby edundan Un wh ang n epar Jendra Sngh 2 abns orwal 2 Deparmen

More information

Online Appendix for. Strategic safety stocks in supply chains with evolving forecasts

Online Appendix for. Strategic safety stocks in supply chains with evolving forecasts Onlne Appendx for Sraegc safey socs n supply chans wh evolvng forecass Tor Schoenmeyr Sephen C. Graves Opsolar, Inc. 332 Hunwood Avenue Hayward, CA 94544 A. P. Sloan School of Managemen Massachuses Insue

More information

Existence and Uniqueness Results for Random Impulsive Integro-Differential Equation

Existence and Uniqueness Results for Random Impulsive Integro-Differential Equation Global Journal of Pure and Appled Mahemacs. ISSN 973-768 Volume 4, Number 6 (8), pp. 89-87 Research Inda Publcaons hp://www.rpublcaon.com Exsence and Unqueness Resuls for Random Impulsve Inegro-Dfferenal

More information

5th International Conference on Advanced Design and Manufacturing Engineering (ICADME 2015)

5th International Conference on Advanced Design and Manufacturing Engineering (ICADME 2015) 5h Inernaonal onference on Advanced Desgn and Manufacurng Engneerng (IADME 5 The Falure Rae Expermenal Sudy of Specal N Machne Tool hunshan He, a, *, La Pan,b and Bng Hu 3,c,,3 ollege of Mechancal and

More information

e-journal Reliability: Theory& Applications No 2 (Vol.2) Vyacheslav Abramov

e-journal Reliability: Theory& Applications No 2 (Vol.2) Vyacheslav Abramov June 7 e-ournal Relably: Theory& Applcaons No (Vol. CONFIDENCE INTERVALS ASSOCIATED WITH PERFORMANCE ANALYSIS OF SYMMETRIC LARGE CLOSED CLIENT/SERVER COMPUTER NETWORKS Absrac Vyacheslav Abramov School

More information

HEAT CONDUCTION PROBLEM IN A TWO-LAYERED HOLLOW CYLINDER BY USING THE GREEN S FUNCTION METHOD

HEAT CONDUCTION PROBLEM IN A TWO-LAYERED HOLLOW CYLINDER BY USING THE GREEN S FUNCTION METHOD Journal of Appled Mahemacs and Compuaonal Mechancs 3, (), 45-5 HEAT CONDUCTION PROBLEM IN A TWO-LAYERED HOLLOW CYLINDER BY USING THE GREEN S FUNCTION METHOD Sansław Kukla, Urszula Sedlecka Insue of Mahemacs,

More information

Cubic Bezier Homotopy Function for Solving Exponential Equations

Cubic Bezier Homotopy Function for Solving Exponential Equations Penerb Journal of Advanced Research n Compung and Applcaons ISSN (onlne: 46-97 Vol. 4, No.. Pages -8, 6 omoopy Funcon for Solvng Eponenal Equaons S. S. Raml *,,. Mohamad Nor,a, N. S. Saharzan,b and M.

More information

This document is downloaded from DR-NTU, Nanyang Technological University Library, Singapore.

This document is downloaded from DR-NTU, Nanyang Technological University Library, Singapore. Ths documen s downloaded from DR-NTU, Nanyang Technologcal Unversy Lbrary, Sngapore. Tle A smplfed verb machng algorhm for word paron n vsual speech processng( Acceped verson ) Auhor(s) Foo, Say We; Yong,

More information

M. Y. Adamu Mathematical Sciences Programme, AbubakarTafawaBalewa University, Bauchi, Nigeria

M. Y. Adamu Mathematical Sciences Programme, AbubakarTafawaBalewa University, Bauchi, Nigeria IOSR Journal of Mahemacs (IOSR-JM e-issn: 78-578, p-issn: 9-765X. Volume 0, Issue 4 Ver. IV (Jul-Aug. 04, PP 40-44 Mulple SolonSoluons for a (+-dmensonalhroa-sasuma shallow waer wave equaon UsngPanlevé-Bӓclund

More information

Lecture 11 SVM cont

Lecture 11 SVM cont Lecure SVM con. 0 008 Wha we have done so far We have esalshed ha we wan o fnd a lnear decson oundary whose margn s he larges We know how o measure he margn of a lnear decson oundary Tha s: he mnmum geomerc

More information

Econ107 Applied Econometrics Topic 5: Specification: Choosing Independent Variables (Studenmund, Chapter 6)

Econ107 Applied Econometrics Topic 5: Specification: Choosing Independent Variables (Studenmund, Chapter 6) Econ7 Appled Economercs Topc 5: Specfcaon: Choosng Independen Varables (Sudenmund, Chaper 6 Specfcaon errors ha we wll deal wh: wrong ndependen varable; wrong funconal form. Ths lecure deals wh wrong ndependen

More information

10. A.C CIRCUITS. Theoretically current grows to maximum value after infinite time. But practically it grows to maximum after 5τ. Decay of current :

10. A.C CIRCUITS. Theoretically current grows to maximum value after infinite time. But practically it grows to maximum after 5τ. Decay of current : . A. IUITS Synopss : GOWTH OF UNT IN IUIT : d. When swch S s closed a =; = d. A me, curren = e 3. The consan / has dmensons of me and s called he nducve me consan ( τ ) of he crcu. 4. = τ; =.63, n one

More information

2/20/2013. EE 101 Midterm 2 Review

2/20/2013. EE 101 Midterm 2 Review //3 EE Mderm eew //3 Volage-mplfer Model The npu ressance s he equalen ressance see when lookng no he npu ermnals of he amplfer. o s he oupu ressance. I causes he oupu olage o decrease as he load ressance

More information

2.1 Constitutive Theory

2.1 Constitutive Theory Secon.. Consuve Theory.. Consuve Equaons Governng Equaons The equaons governng he behavour of maerals are (n he spaal form) dρ v & ρ + ρdv v = + ρ = Conservaon of Mass (..a) d x σ j dv dvσ + b = ρ v& +

More information

Dual Approximate Dynamic Programming for Large Scale Hydro Valleys

Dual Approximate Dynamic Programming for Large Scale Hydro Valleys Dual Approxmae Dynamc Programmng for Large Scale Hydro Valleys Perre Carpener and Jean-Phlppe Chanceler 1 ENSTA ParsTech and ENPC ParsTech CMM Workshop, January 2016 1 Jon work wh J.-C. Alas, suppored

More information

Approximate Analytic Solution of (2+1) - Dimensional Zakharov-Kuznetsov(Zk) Equations Using Homotopy

Approximate Analytic Solution of (2+1) - Dimensional Zakharov-Kuznetsov(Zk) Equations Using Homotopy Arcle Inernaonal Journal of Modern Mahemacal Scences, 4, (): - Inernaonal Journal of Modern Mahemacal Scences Journal homepage: www.modernscenfcpress.com/journals/jmms.aspx ISSN: 66-86X Florda, USA Approxmae

More information

Political Economy of Institutions and Development: Problem Set 2 Due Date: Thursday, March 15, 2019.

Political Economy of Institutions and Development: Problem Set 2 Due Date: Thursday, March 15, 2019. Polcal Economy of Insuons and Developmen: 14.773 Problem Se 2 Due Dae: Thursday, March 15, 2019. Please answer Quesons 1, 2 and 3. Queson 1 Consder an nfne-horzon dynamc game beween wo groups, an ele and

More information

DEEP UNFOLDING FOR MULTICHANNEL SOURCE SEPARATION SUPPLEMENTARY MATERIAL

DEEP UNFOLDING FOR MULTICHANNEL SOURCE SEPARATION SUPPLEMENTARY MATERIAL DEEP UNFOLDING FOR MULTICHANNEL SOURCE SEPARATION SUPPLEMENTARY MATERIAL Sco Wsdom, John Hershey 2, Jonahan Le Roux 2, and Shnj Waanabe 2 Deparmen o Elecrcal Engneerng, Unversy o Washngon, Seale, WA, USA

More information

In the complete model, these slopes are ANALYSIS OF VARIANCE FOR THE COMPLETE TWO-WAY MODEL. (! i+1 -! i ) + [(!") i+1,q - [(!

In the complete model, these slopes are ANALYSIS OF VARIANCE FOR THE COMPLETE TWO-WAY MODEL. (! i+1 -! i ) + [(!) i+1,q - [(! ANALYSIS OF VARIANCE FOR THE COMPLETE TWO-WAY MODEL The frs hng o es n wo-way ANOVA: Is here neracon? "No neracon" means: The man effecs model would f. Ths n urn means: In he neracon plo (wh A on he horzonal

More information

Appendix H: Rarefaction and extrapolation of Hill numbers for incidence data

Appendix H: Rarefaction and extrapolation of Hill numbers for incidence data Anne Chao Ncholas J Goell C seh lzabeh L ander K Ma Rober K Colwell and Aaron M llson 03 Rarefacon and erapolaon wh ll numbers: a framewor for samplng and esmaon n speces dversy sudes cology Monographs

More information

Ordinary Differential Equations in Neuroscience with Matlab examples. Aim 1- Gain understanding of how to set up and solve ODE s

Ordinary Differential Equations in Neuroscience with Matlab examples. Aim 1- Gain understanding of how to set up and solve ODE s Ordnary Dfferenal Equaons n Neuroscence wh Malab eamples. Am - Gan undersandng of how o se up and solve ODE s Am Undersand how o se up an solve a smple eample of he Hebb rule n D Our goal a end of class

More information

Chapter 6: AC Circuits

Chapter 6: AC Circuits Chaper 6: AC Crcus Chaper 6: Oulne Phasors and he AC Seady Sae AC Crcus A sable, lnear crcu operang n he seady sae wh snusodal excaon (.e., snusodal seady sae. Complee response forced response naural response.

More information

Part II CONTINUOUS TIME STOCHASTIC PROCESSES

Part II CONTINUOUS TIME STOCHASTIC PROCESSES Par II CONTINUOUS TIME STOCHASTIC PROCESSES 4 Chaper 4 For an advanced analyss of he properes of he Wener process, see: Revus D and Yor M: Connuous marngales and Brownan Moon Karazas I and Shreve S E:

More information

Multi-priority Online Scheduling with Cancellations

Multi-priority Online Scheduling with Cancellations Submed o Operaons Research manuscrp (Please, provde he manuscrp number!) Auhors are encouraged o subm new papers o INFORMS journals by means of a syle fle emplae, whch ncludes he journal le. However, use

More information

Robust and Accurate Cancer Classification with Gene Expression Profiling

Robust and Accurate Cancer Classification with Gene Expression Profiling Robus and Accurae Cancer Classfcaon wh Gene Expresson Proflng (Compuaonal ysems Bology, 2005) Auhor: Hafeng L, Keshu Zhang, ao Jang Oulne Background LDA (lnear dscrmnan analyss) and small sample sze problem

More information

On elements with index of the form 2 a 3 b in a parametric family of biquadratic elds

On elements with index of the form 2 a 3 b in a parametric family of biquadratic elds On elemens wh ndex of he form a 3 b n a paramerc famly of bquadrac elds Bora JadrevĆ Absrac In hs paper we gve some resuls abou prmve negral elemens p(c p n he famly of bcyclc bquadrac elds L c = Q ) c;

More information

How about the more general "linear" scalar functions of scalars (i.e., a 1st degree polynomial of the following form with a constant term )?

How about the more general linear scalar functions of scalars (i.e., a 1st degree polynomial of the following form with a constant term )? lmcd Lnear ransformaon of a vecor he deas presened here are que general hey go beyond he radonal mar-vecor ype seen n lnear algebra Furhermore, hey do no deal wh bass and are equally vald for any se of

More information

On computing differential transform of nonlinear non-autonomous functions and its applications

On computing differential transform of nonlinear non-autonomous functions and its applications On compung dfferenal ransform of nonlnear non-auonomous funcons and s applcaons Essam. R. El-Zahar, and Abdelhalm Ebad Deparmen of Mahemacs, Faculy of Scences and Humanes, Prnce Saam Bn Abdulazz Unversy,

More information

( t) Outline of program: BGC1: Survival and event history analysis Oslo, March-May Recapitulation. The additive regression model

( t) Outline of program: BGC1: Survival and event history analysis Oslo, March-May Recapitulation. The additive regression model BGC1: Survval and even hsory analyss Oslo, March-May 212 Monday May 7h and Tuesday May 8h The addve regresson model Ørnulf Borgan Deparmen of Mahemacs Unversy of Oslo Oulne of program: Recapulaon Counng

More information

Computational results on new staff scheduling benchmark instances

Computational results on new staff scheduling benchmark instances TECHNICAL REPORT Compuaonal resuls on new saff shedulng enhmark nsanes Tm Curos Rong Qu ASAP Researh Group Shool of Compuer Sene Unersy of Nongham NG8 1BB Nongham UK Frs pulshed onlne: 19-Sep-2014 las

More information

SOME NOISELESS CODING THEOREMS OF INACCURACY MEASURE OF ORDER α AND TYPE β

SOME NOISELESS CODING THEOREMS OF INACCURACY MEASURE OF ORDER α AND TYPE β SARAJEVO JOURNAL OF MATHEMATICS Vol.3 (15) (2007), 137 143 SOME NOISELESS CODING THEOREMS OF INACCURACY MEASURE OF ORDER α AND TYPE β M. A. K. BAIG AND RAYEES AHMAD DAR Absrac. In hs paper, we propose

More information

CH.3. COMPATIBILITY EQUATIONS. Continuum Mechanics Course (MMC) - ETSECCPB - UPC

CH.3. COMPATIBILITY EQUATIONS. Continuum Mechanics Course (MMC) - ETSECCPB - UPC CH.3. COMPATIBILITY EQUATIONS Connuum Mechancs Course (MMC) - ETSECCPB - UPC Overvew Compably Condons Compably Equaons of a Poenal Vecor Feld Compably Condons for Infnesmal Srans Inegraon of he Infnesmal

More information

Using Aggregation to Construct Periodic Policies for Routing Jobs to Parallel Servers with Deterministic Service Times

Using Aggregation to Construct Periodic Policies for Routing Jobs to Parallel Servers with Deterministic Service Times Usng Aggregaon o Consruc Perodc Polces for Roung Jobs o Parallel Servers wh Deermnsc Servce Tmes Jeffrey W. errmann A. James Clark School of Engneerng 2181 Marn all Unversy of Maryland College Park, MD

More information

12d Model. Civil and Surveying Software. Drainage Analysis Module Detention/Retention Basins. Owen Thornton BE (Mech), 12d Model Programmer

12d Model. Civil and Surveying Software. Drainage Analysis Module Detention/Retention Basins. Owen Thornton BE (Mech), 12d Model Programmer d Model Cvl and Surveyng Soware Dranage Analyss Module Deenon/Reenon Basns Owen Thornon BE (Mech), d Model Programmer owen.hornon@d.com 4 January 007 Revsed: 04 Aprl 007 9 February 008 (8Cp) Ths documen

More information

Department of Economics University of Toronto

Department of Economics University of Toronto Deparmen of Economcs Unversy of Torono ECO408F M.A. Economercs Lecure Noes on Heeroskedascy Heeroskedascy o Ths lecure nvolves lookng a modfcaons we need o make o deal wh he regresson model when some of

More information

A NEW METHOD OF FMS SCHEDULING USING OPTIMIZATION AND SIMULATION

A NEW METHOD OF FMS SCHEDULING USING OPTIMIZATION AND SIMULATION A NEW METHD F FMS SCHEDULING USING PTIMIZATIN AND SIMULATIN Ezedeen Kodeekha Deparmen of Producon, Informacs, Managemen and Conrol Faculy of Mechancal Engneerng udapes Unversy of Technology and Econcs

More information

APOC #232 Capacity Planning for Fault-Tolerant All-Optical Network

APOC #232 Capacity Planning for Fault-Tolerant All-Optical Network APOC #232 Capacy Plannng for Faul-Toleran All-Opcal Nework Mchael Kwok-Shng Ho and Kwok-wa Cheung Deparmen of Informaon ngneerng The Chnese Unversy of Hong Kong Shan, N.T., Hong Kong SAR, Chna -mal: kwcheung@e.cuhk.edu.hk

More information

Efficient Asynchronous Channel Hopping Design for Cognitive Radio Networks

Efficient Asynchronous Channel Hopping Design for Cognitive Radio Networks Effcen Asynchronous Channel Hoppng Desgn for Cognve Rado Neworks Chh-Mn Chao, Chen-Yu Hsu, and Yun-ng Lng Absrac In a cognve rado nework (CRN), a necessary condon for nodes o communcae wh each oher s ha

More information

J i-1 i. J i i+1. Numerical integration of the diffusion equation (I) Finite difference method. Spatial Discretization. Internal nodes.

J i-1 i. J i i+1. Numerical integration of the diffusion equation (I) Finite difference method. Spatial Discretization. Internal nodes. umercal negraon of he dffuson equaon (I) Fne dfference mehod. Spaal screaon. Inernal nodes. R L V For hermal conducon le s dscree he spaal doman no small fne spans, =,,: Balance of parcles for an nernal

More information

Epistemic Game Theory: Online Appendix

Epistemic Game Theory: Online Appendix Epsemc Game Theory: Onlne Appendx Edde Dekel Lucano Pomao Marcano Snscalch July 18, 2014 Prelmnares Fx a fne ype srucure T I, S, T, β I and a probably µ S T. Le T µ I, S, T µ, βµ I be a ype srucure ha

More information

Advanced Macroeconomics II: Exchange economy

Advanced Macroeconomics II: Exchange economy Advanced Macroeconomcs II: Exchange economy Krzyszof Makarsk 1 Smple deermnsc dynamc model. 1.1 Inroducon Inroducon Smple deermnsc dynamc model. Defnons of equlbrum: Arrow-Debreu Sequenal Recursve Equvalence

More information

Genetic Algorithm in Parameter Estimation of Nonlinear Dynamic Systems

Genetic Algorithm in Parameter Estimation of Nonlinear Dynamic Systems Genec Algorhm n Parameer Esmaon of Nonlnear Dynamc Sysems E. Paeraks manos@egnaa.ee.auh.gr V. Perds perds@vergna.eng.auh.gr Ah. ehagas kehagas@egnaa.ee.auh.gr hp://skron.conrol.ee.auh.gr/kehagas/ndex.hm

More information

Determining the Number of Games Needed to Guarantee an NHL Playoff Spot

Determining the Number of Games Needed to Guarantee an NHL Playoff Spot Deermnng he Number of Games Needed o Guaranee an NHL Playoff Spo Tyrel Russell and Peer van Bee Cheron School of Compuer Scence Unversy of Waerloo {crussel,vanbee}@uwaerloo.ca Absrac. Many spors fans nves

More information

Panel Data Regression Models

Panel Data Regression Models Panel Daa Regresson Models Wha s Panel Daa? () Mulple dmensoned Dmensons, e.g., cross-secon and me node-o-node (c) Pongsa Pornchawseskul, Faculy of Economcs, Chulalongkorn Unversy (c) Pongsa Pornchawseskul,

More information

Density Matrix Description of NMR BCMB/CHEM 8190

Density Matrix Description of NMR BCMB/CHEM 8190 Densy Marx Descrpon of NMR BCMBCHEM 89 Operaors n Marx Noaon Alernae approach o second order specra: ask abou x magnezaon nsead of energes and ranson probables. If we say wh one bass se, properes vary

More information

Sampling Procedure of the Sum of two Binary Markov Process Realizations

Sampling Procedure of the Sum of two Binary Markov Process Realizations Samplng Procedure of he Sum of wo Bnary Markov Process Realzaons YURY GORITSKIY Dep. of Mahemacal Modelng of Moscow Power Insue (Techncal Unversy), Moscow, RUSSIA, E-mal: gorsky@yandex.ru VLADIMIR KAZAKOV

More information

Planar truss bridge optimization by dynamic programming and linear programming

Planar truss bridge optimization by dynamic programming and linear programming IABSE-JSCE Jon Conference on Advances n Brdge Engneerng-III, Augus 1-, 015, Dhaka, Bangladesh. ISBN: 978-984-33-9313-5 Amn, Oku, Bhuyan, Ueda (eds.) www.abse-bd.org Planar russ brdge opmzaon by dynamc

More information

Lecture 6: Learning for Control (Generalised Linear Regression)

Lecture 6: Learning for Control (Generalised Linear Regression) Lecure 6: Learnng for Conrol (Generalsed Lnear Regresson) Conens: Lnear Mehods for Regresson Leas Squares, Gauss Markov heorem Recursve Leas Squares Lecure 6: RLSC - Prof. Sehu Vjayakumar Lnear Regresson

More information

NPTEL Project. Econometric Modelling. Module23: Granger Causality Test. Lecture35: Granger Causality Test. Vinod Gupta School of Management

NPTEL Project. Econometric Modelling. Module23: Granger Causality Test. Lecture35: Granger Causality Test. Vinod Gupta School of Management P age NPTEL Proec Economerc Modellng Vnod Gua School of Managemen Module23: Granger Causaly Tes Lecure35: Granger Causaly Tes Rudra P. Pradhan Vnod Gua School of Managemen Indan Insue of Technology Kharagur,

More information

Survival Analysis and Reliability. A Note on the Mean Residual Life Function of a Parallel System

Survival Analysis and Reliability. A Note on the Mean Residual Life Function of a Parallel System Communcaons n Sascs Theory and Mehods, 34: 475 484, 2005 Copyrgh Taylor & Francs, Inc. ISSN: 0361-0926 prn/1532-415x onlne DOI: 10.1081/STA-200047430 Survval Analyss and Relably A Noe on he Mean Resdual

More information

CS 268: Packet Scheduling

CS 268: Packet Scheduling Pace Schedulng Decde when and wha pace o send on oupu ln - Usually mplemened a oupu nerface CS 68: Pace Schedulng flow Ion Soca March 9, 004 Classfer flow flow n Buffer managemen Scheduler soca@cs.bereley.edu

More information

Hidden Markov Models Following a lecture by Andrew W. Moore Carnegie Mellon University

Hidden Markov Models Following a lecture by Andrew W. Moore Carnegie Mellon University Hdden Markov Models Followng a lecure by Andrew W. Moore Carnege Mellon Unversy www.cs.cmu.edu/~awm/uorals A Markov Sysem Has N saes, called s, s 2.. s N s 2 There are dscree meseps, 0,, s s 3 N 3 0 Hdden

More information

Including the ordinary differential of distance with time as velocity makes a system of ordinary differential equations.

Including the ordinary differential of distance with time as velocity makes a system of ordinary differential equations. Soluons o Ordnary Derenal Equaons An ordnary derenal equaon has only one ndependen varable. A sysem o ordnary derenal equaons consss o several derenal equaons each wh he same ndependen varable. An eample

More information

3. OVERVIEW OF NUMERICAL METHODS

3. OVERVIEW OF NUMERICAL METHODS 3 OVERVIEW OF NUMERICAL METHODS 3 Inroducory remarks Ths chaper summarzes hose numercal echnques whose knowledge s ndspensable for he undersandng of he dfferen dscree elemen mehods: he Newon-Raphson-mehod,

More information

Lecture VI Regression

Lecture VI Regression Lecure VI Regresson (Lnear Mehods for Regresson) Conens: Lnear Mehods for Regresson Leas Squares, Gauss Markov heorem Recursve Leas Squares Lecure VI: MLSC - Dr. Sehu Vjayakumar Lnear Regresson Model M

More information