Incremental and Encoding Formulations for Mixed Integer Programming

Size: px
Start display at page:

Download "Incremental and Encoding Formulations for Mixed Integer Programming"

Transcription

1 Incremental and Encodng Formulatons for Mxed Integer Programmng Sercan Yıldız a, Juan Pablo Velma b,c, a Tepper School of Busness, Carnege Mellon Unversty, 5000 Forbes Ave, Pttsburgh, PA 523, Unted States b Sloan School of Management, Massachusetts Insttute of Technology, 77 Massachusetts Ave., Cambrdge, MA 0239, Unted States c Department of Industral Engneerng, Unversty of Pttsburgh, 3700 O Hara Street, Pttsburgh, PA 526, Unted States Abstract The standard way to represent a choce between n alternatves n Mxed Integer Programmng s through n bnary varables that add up to one. Unfortunately, ths approach commonly leads to unbalanced branch-and-bound trees and dmnshed solver performance. In ths paper, we present an encodng formulaton framework that encompasses and expands exstng approaches to mtgate ths behavor. Through ths framework, we generalze the ncremental formulaton for pecewse lnear functons to any fnte unon of polyhedra wth dentcal recesson cones. Keywords: Mxed Integer Programmng, Dsunctve Programmng, Formulatons. Introducton A textbook approach to selectng among n dscrete alternatves n a Mxed Integer Programmng (MIP) problem s to utlze n bnary varables {y } n wth the addtonal requrement that they add up to one. In partcular, ths s captured by the set D n := n \ {0, } n where n := y 2 R n + : P n y =. For nstance, gven a fnte set {a,...,a n } R, we may model the constrant x 2 {a,...,a n } Correspondng author. Tel.: Emal addresses: syldz@andrew.cmu.edu (Sercan Yıldız), velma@mt.edu (Juan Pablo Velma) Preprnt submtted to Elsever September 3, 203

2 va the MIP formulaton gven by x = a y, y 2 D n. () Ths approach s qute smple and, as we dscuss n Secton 3, t can be easly extended to more general settngs. Unfortunately, although the approach s usually qute e ectve, t has some unfavorable propertes that can slow down branch-andbound based MIP solvers. These propertes stem from the way the formulaton s a ected by standard varable branchng n a branch-and-bound algorthm. Whle fxng a gven bnary varable y to one (usually denoted up-branchng) fxes all other varables to zero, fxng the same varable to zero (usually denoted downbranchng) does not mpose any constrants on the other varables (unless n = 2). It s well-known that ths behavor leads to unbalanced branch-and-bound trees and may result n long soluton tmes. To ths date, there have been three central approaches to dealng wth the branchng ssues that arse n formulatons based on D n. The frst approach replaces the usual varable branchng strategy wth a more sophstcated constrant branchng scheme []. For nstance, Beale and Tomln [2] ntroduced SOS branchng as a very e ectve branchng scheme for D n and other related sets. Unfortunately, f SOS branchng s not mplemented usng modern technques, t can be slower than varable branchng n state-of-the-art solvers [3]. One way to crcumvent ths problem s to use bnary varables to smulate a specalzed constrant branchng scheme [4, 5]. The second approach s normally employed n schedulng problems where selectng opton corresponds to executng a certan actvty n perod. In ths settng, we can use an alternatve set of bnary varables that nstead ndcate whether the actvty s executed n perod or before. These new varables are sometmes called by-varables and lead to balanced branch-and-bound trees that can sgnfcantly reduce soluton tmes [6 2]. The thrd approach proposes to model n alternatves wth a number of bnary varables that s logarthmc n n [5, 3 7]. Ths approach tends to generate more balanced branch-and-bound trees, but, as we wll see n Secton 2, they are not mmune to mbalance ssues ether. Nevertheless, formulatons that are bult usng a logarthmc number of bnary varables usually provde a sgnfcant computatonal advantage. The encodng formulatons that we descrbe n Secton 2 provde a general framework n whch all three approaches can be vewed. In Secton 3, we use ths framework to come up wth an ncremental formulaton for the fnte unon 2

3 of polyhedra wth dentcal recesson cones. We end the paper wth a note on the selecton of encodngs n Secton 4. In the remander, we let [n] := {,...,n} wth the understandng that [0] = ;. We use 0 n 2 R n to refer to the n-dmensonal vector of all zeros and e,n 2 R n to refer to the th n-dmensonal standard unt vector. 2. Encodng Formulatons and Branch-and-Bound The bass for our analyss wll be a reformulaton of D n whch has been part of the mathematcal programmng folklore snce at least 972 [3] and whch has recently been re-dscovered by several authors [5, 5, 7]. Ths formulaton requres a set of vectors n b o n {0, }m such that b, b for any,. For any such set, t models y 2 D n as y =, u = b y, u 2 {0, } m, y 2 R n +. (2) A key theoretcal strength of (2) s the tghtness of ts LP relaxaton. In ths regard, a MIP formulaton has the strongest possble property when all extreme ponts of ts LP relaxaton obey the correspondng ntegralty requrements. Such formulatons are referred to as locally deal by Padberg [8] and Padberg and Ral [9]. We note here that (2) s locally deal for any choce of n b o n {0, }m as long as b, b for any, [6, 7]. Formulaton (2) may seem wasteful as t contans more varables and constrants n o than the orgnal defnton of D n. However, dependng on the choce of n b, (2) can present a computatonal advantage by sgnfcantly reducng the number of bnary varables or by leadng to more balanced branch-and-bound trees. One reason for ths advantage stems from the fact that varable branchng on u nduces a constrant branchng scheme on y n the sprt of [4, 5]. More specfcally, we have that up-branchng on the varable u fxes y = 0 for all such that b = 0 whle down-branchng on u fxes y = 0 for all such that b =. The e ectveness of ths nduced branchng scheme wll of course depend on the choce of n b o n. We now descrbe three such choces and analyze the branchng schemes they nduce. We wll refer to the resultng formulatons as encodng formulatons snce one can thnk of n b o n as an encodng of a selecton among the alternatves n the sense that u = b f and only f alternatve s selected. 3

4 We obtan the smplest encodng when we set m = n and b = e,n. In ths case, (2) reduces to y =, u = y, 8 2 [n], (3) u 2 {0, } n, y 2 R n +, whch s nothng more than a redundant reformulaton of D n. We can easly see that the nduced branchng scheme s ust the standard varable branchng on y and, hence, we do not obtan any benefts. We refer to ths choce as the unary encodng formulaton. A more meanngful choce appears n the development of logarthmc-szed formulatons. For ease of exposton, let us assume that n = 2 k for some postve nteger k 2 Z + although the approach can easly be adapted to more general cases as well. In ths settng, we can smply let m = k and n b o n = {0, }m. Ths choce transforms (2) nto a formulaton wth a logarthmc number of bnary varables whch was the orgnal case studed n [5, 3, 5, 7]. The nduced constrant branchng scheme s not qute clear, but we can easly see that t creates balanced branch-and-bound trees by observng that we have n 2 [n] :b 0 = 0, b = u, 8 2 J o = n 2 [n] :b 0 =, b = u, 8 2 J o for all 0 2 [m], J [m] \ { 0 } and u 2 {0, } m. Here, J can be consdered as the subset of n u o m whch has been fxed to some value at a partcular node of = the branch-and-bound tree. Together wth the reduced number of varables, ths property can lead to a sgnfcant computatonal advantage. We refer to ths choce as the bnary encodng formulaton. Whle the bnary encodng formulaton nduces a non-trval branchng scheme, as noted n [5], t does not nduce the tradtonal SOS constrant branchng scheme. To construct a formulaton that nduces SOS branchng on y, we can set m = n, b = 0 m and b = P = e,m for all 2 {2,...,n}. In ths case, we have that up-branchng on the varable u fxes y = 0 for all apple whle downbranchng on u fxes y = 0 for all >. Thus, we recover precsely the SOS branchng scheme of Beale and Tomln [2]. We refer to ths choce as the ncremental encodng formulaton because t can be used to construct a formulaton for pecewse-lnear functons that s known as the ncremental model n the lterature. 4

5 To llustrate the potental advantage of the ncremental encodng formulaton, we study the behavor of a smple branch-and-bound algorthm on the d erent formulatons of the followng smple problem. For a gven a 2 (Z \ {0}) n such that a < a + for all 2 [n ], consder the problem mn x s.t. x 2 {a,...,a n }. (4a) (4b) We can use (2) to model the dscrete alternatves constrant (4b) and usual lnear programmng trcks to lnearze the obectve functon. Applyng these technques leads us to the MIP formulaton of (4) gven by mn t s.t. (5a) x apple t, x apple t (5b) x = a y, y =, y 2 R n +, (5c) u = b y, u 2 {0, } m. (5d) Note that, n (5), we have modeled the dscrete alternatves constrant and lnearzed the obectve functon ndependently. Whle t s possble to construct a stronger formulaton for ths problem by consderng both aspects at the same tme, we refran from ths because our ntent here s to evaluate the e ectveness of the formulatons as they would be used n practce where d erent portons of an optmzaton problem are modeled ndependently and then combned. The followng proposton shows that the ncremental encodng formulaton never requres branchng on more than a sngle varable to solve ths problem. In contrast, there are choces of n and a for whch the unary and bnary encodng formulatons may need branchng on up to n/2 and log 2 n varables, respectvely. Proposton 2... For any n and choce of a 2 (Z \ {0}) n, the ncremental encodng verson of (5) can be solved by a branch-and-bound algorthm by branchng on at most one varable. 5

6 2. If n s even, a < 0 for all apple n/2 and a > 0 for all > n/2, then a branchand-bound algorthm solvng the unary encodng verson of (5) requres branchng on at least n/2 varables. 3. If n s a power of two, a < 0 < a n for all apple n, then a branch-and-bound algorthm solvng the bnary encodng verson of (5) requres branchng on at least log 2 n varables. Proof. Pror to any branchng, the LP relaxaton of (5) equals the nterval [a, a n ] n the x-space. When a n < 0 or a > 0, the optmal soluton of the LP relaxaton occurs at one of the endponts of ths nterval and satsfes (4b). Therefore, no branchng s necessary. Otherwse, there exsts 2 [n] such that a < 0 < a +. It s clear that the optmal soluton to (4) s ether x = a or x = a +. Recall that, n the ncremental encodng formulaton, we have b = 0 for all apple and b = for all >. Hence, up-branchng on the sngle varable u fxes y = 0 for all apple whle down-branchng on u fxes y = 0 for all >. When proected on to the x-space, the LP relaxaton of (5) n the frst branch corresponds to the nterval [a +, a n ] and has the optmal soluton x = a +. Smlarly, the LP relaxaton n the second branch proects down to [a, a ] and has the optmal soluton x = a. Ths proves Clam. Now, suppose a branch-and-bound algorthm can solve the unary encodng verson of (5) by branchng on the varables {u } 2K where K [n] and K < n/2. As we have observed before, up-branchng on the varable u fxes y = and y = 0 for all, n the unary encodng formulaton. Hence, the optmal value of (5) at any leaf reached by up-branchng on a varable u must equal a > 0. On the other hand, the LP relaxaton of (5) at the leaf reached by down-branchng on all {u } 2K has the optmal value 0 snce there exst dstnct 0, 0 2 [n] \ K such that 0 apple n/2 and 0 > n/2 and the proecton of the LP relaxaton on to the x-space contans 0. Ths leaf could have been pruned by nether ntegralty nor bound and must stll be actve. Ths contradcton proves Clam 2. The proof of Clam 3 follows an outlne smlar to that of Clam 2. Let l := log 2 n and suppose a branch-and-bound algorthm can solve the bnary encodng verson of (5) by branchng on the varables {u } 2K where K [l]. At termnaton, the branch-and-bound tree has a leaf at whch u = b n s a feasble assgnment. Let 0 2 [l] \ K. Let 0 2 [n] be such that b 0 = b n 0 and b 0 0 = b n for all, 0. It follows that u = b 0 s a feasble assgnment at ths leaf as well. The fact that a 0 < 0 mples that the proecton of the LP relaxaton of (5) on to the x-space at ths leaf contans 0. Agan, we conclude that ths leaf could have been pruned by nether ntegralty nor bound. Ths completes the proof of Clam 3. 6

7 3. Incremental Formulatons An nterestng behavor of the ncremental encodng formulaton s that t naturally nduces the order u u 2... u n. Ths orderng condton s also shared by the by-varable formulatons [6 2], whch explans why they lead to balanced branch-and-bound trees. Moreover, t s explctly mposed n other more complcated ad-hoc formulatons for pecewse-lnear functons [20 23] and probablstc constrants [24, 25]. We now show that all these formulatons can be obtaned through a generc procedure. The frst step n ths procedure s to note that we can use D n to formulate more general dscrete alternatves by usng the followng formulaton ntroduced by Lowe [26] and Jeroslow and Lowe [27]. Proposton 3. ([26, 27]). Let {P } n be a fnte famly of polyhedra such that there exsts n r ko R k=0 Rd and n o v V =0 Rd for 2 [n] such that P := conv n o v V + =0 cone n r ko R for all 2 [n]. Then a MIP formulaton of x 2 S n k=0 P s gven by x = VX VX =0 =0 XR v + k=0 = y, 8 2 [n], µ k r k, (6a) (6b) µ 0, 0, 8 2 [n], y 2 D n. (6c) Formulaton (6) s locally deal [26 28]. We can combne ths standard formulaton wth the generc encodng formulaton (2) of D n to obtan the followng generalzaton of a formulaton for pecewse-lnear functons ntroduced n [6]. Corollary 3.2. Let {P } n be a fnte famly of polyhedra such that there exsts n o R r k R d and n o v V k=0 R d for 2 [n] such that P := conv n o v V =0 + =0 cone n r ko R for all 2 [n]. Then a MIP formulaton of x 2 S n k=0 P s gven by 7

8 x = VX =0 VX =0 XR v + =, u = k=0 µ k r k, (7a) VX =0 b, (7b) µ 0, 0, 8 2 [n], u 2 {0, } m. (7c) The fact that (7) s locally deal follows drectly from the smlar result on (2). For the specfc case of the ncremental encodng, we can use smple algebrac manpulatons to obtan the followng formulaton whch explctly consders the order property among the u varables. Proposton 3.3. Let {P } n be a fnte famly of polyhedra such that there exsts n r ko R k=0 Rd and n o v V R d for 2 [n] such that P := conv n o v V =0 + =0 cone n r ko R for all 2 [n]. Then a MIP formulaton of x 2 S n k=0 P s gven by Xn x = v 0 + u v 0 + v V + VX = v v 0 XR + µ k r k, k=0 (8a) u apple V, 8 2 [n ], (8b) VX + = + apple u, 8 2 [n ], (8c) VX = apple, (8d) µ 0, 0, 8 2 [n], u 2 {0, } n. (8e) Proof. We exhbt a becton whch preserves the values of the x varables between ponts satsfyng (8) and the ncremental encodng verson of (7). The clam then follows from Corollary 3.2. In the rest of the proof, we refer to the ncremental encodng verson of (7) as smply (7). Let (x, u,, µ) be a soluton to (7). Defne u 0 := and u n := 0 and observe that (7b) can be rewrtten as u = P n l=+ 8 P Vl =0 l

9 for all 2 {0,,...,n } when the vectors n b o n are chosen accordng to ncremental encodng. Let := for all 2 [V 2] and V := u + V for all 2 [n]. The valdty of (8b) and (8e) for (x, u,, µ) follows from the defnton of and the non-negatvty of. Smlarly, (8c) and (8d) are satsfed because the nonnegatvty of mples P V + = + = u + + P V + = + apple u + + P V + =0 + = u for all 2 [n ] and P V = = u + P V = apple u + P V =0 =. To verfy that (8a) also holds, let r = P R k=0 µ k r k and observe VX x = v + r = = =0 00 VX VX B@ B@ u u CA v0 + v CA + r = = 0 VX B@ (u u )v 0 + (v v 0 ) CA + r Xn = v 0 + u (v 0 + Xn = v 0 + u (v 0 + v 0 ) + 0 VX 2 B@ = = VX v 0 ) + = (v v 0 ) + r (v v 0 ) + ( V u )(v V v 0 ) CA + r Xn = v 0 + u (v 0 + v V ) + VX = (v v 0 ) + r. To prove the reverse ncluson, let ( x, ũ,, µ) be a soluton to (8). Defne ũ 0 := and ũ n := 0 and let := for all 2 [V 2], V := V ũ and 0 P := ũ V = for all 2 [n]. It s not d cult to see that these defntons mply ũ = P n P Vl l=+ =0 l for all 2 {0,,...,n }. Furthermore, the equaltes n the frst part of the proof contnue to hold. Thus, ( x, ũ,, µ) satsfes (7a)-(7b). To complete the proof, t s enough to show that s non-negatve. However, ths follows drectly from ts defnton and (8b)-(8c). Formulaton (8) s locally deal. To see ths, observe that the lnear mappng 9

10 defned n the proof of Proposton 3.3 s n fact a becton between ponts satsfyng the LP relaxatons of (8) and the ncremental encodng verson of (7). Gven any soluton to the LP relaxaton of (8) n whch the u varables have fractonal values, ths mappng assocates wth t a soluton to the LP relaxaton of (7) wth the same u values. However, ths cannot be an extreme pont soluton to the LP relaxaton of (7) by the smple fact that (7) s locally deal. Hence, t can be expressed as the convex combnaton of other solutons to the LP relaxaton of (7). Mappng these solutons back to (8) and usng the contnuty of lnear mappngs shows that the LP relaxaton of (8) cannot have any extreme pont solutons that volate the ntegralty restrctons. Formulaton (8) generalzes a formulaton ntroduced by Wlson [23] for pecewselnear functons n two ways. Frst, t presents a drect extenson from pecewselnear functons to the unon of a fnte number of arbtrary polyhedra wth dentcal recesson cones. Second, t enoys a broader scope of applcablty by elmnatng the need for a topologcal condton that was requred for the valdty of Wlson s formulaton. We may obtan other ncremental formulatons through the followng rather straghtforward lemma. Lemma 3.4. The ncremental encodng formulaton for y 2 D n s equvalent to u 2 {0, } n, u u +, 8 2 [n 2], (9a) y = u, y n = u n, (9b) y = u u, 8 2 {2,...,n 2}. (9c) Gven any formulaton that requres y 2 D n, we can replace ths constrant wth (9a) and elmnate every occurrence of y through the relatonshps (9b)-(9c). For nstance, followng ths procedure leads us to the formulaton for probablstc constrants studed n [24, 28] from a standard formulaton ntroduced n [29 3]. We refer the readers to [28] for detals on ths specfc transformaton. 4. Creatng and Selectng Encodngs In ths paper, we have concentrated on the unary, bnary and ncremental encodngs. Whle formulatons that use the bnary and ncremental encodngs usually outperform those that use the unary encodng, t can stll be hard to predct whch encodng wll perform best n a specfc class of nstances. For example, [5, 6] present computatonal results whch show that formulatons for pecewselnear functons based on the bnary encodng can sgnfcantly outperform those 0

11 based on the unary and ncremental encodngs. In contrast, n a d erent set of problems, [32] reports that the ncremental formulaton performs better than the bnary encodng formulaton. Formulaton (7) o ers a convenent way to prelmnarly evaluate the performance of these three encodngs wthout havng to hard-code the specfc formulatons derved from them (e.g., (8) for the ncremental n o encodng). Furthermore, (7) s vald for any selecton of dstnct bnary vectors n b and hence can be used to construct alternatve formulatons that could outperform all three prevously consdered encodngs. Developng technques for constructng such alternatve encodngs s beyond the scope of ths paper, but to llustrate ther potental, we present a herarchy of encodngs that combne the propertes of the ncremental and unary encodngs. Such encodngs could be useful n settngs where the ncremental encodng works best n certan portons of the problem whle the bnary encodng works best n others. To smplfy presentaton, we assume n the rest of ths secton that n s a power of 2. The herarchy we propose s ndexed by h 2 0,,...,log 2 n, whch can be consdered as the number of bts of n b o n that behave as n the bnary encodng. Let t l := n/2 l. For a gven value of h, we set m = h + t h and defne the vectors n b o n Rm n the encodng as follows: For all l 2 [h], let b l = f there exsts p 2 [d2 l e] such that 2 {2(p )t l +,...,(2p )t l } and b l = 0 otherwse. For all s 2 [t h ], let b h+s = f there exsts p 2 [d2 h e] such that 2 {2(p )t h + s +,...,(2p )t h + s} and b,h+s = 0 otherwse. Fgure shows ths herarchy of encodngs n matrx form (we present a matrx whose columns are n b o n ) for n = 8 and h 2 {0,, 2, 3}. Note that p ndexes the blocks of contguous s n each row n the defnton above. It can be seen that settng h = 0 produces the ncremental encodng whereas settng h = log 2 n (or even h = log 2 n ) produces the bnary encodng. However, usng d erent values of h yelds hybrd encodngs that share aspects of both the bnary and ncremental encodngs. For nstance, choosng h = n Fgure yelds an encodng where the selecton between the sets of optons {,...,4} and {5,...,8} s done as n the bnary encodng whle the selecton among the optons n {,...,4} or {5,...,8} s done as n the ncremental encodng. References [] D. M. Ryan, B. A. Foster, An nteger programmng approach to schedulng, n: A. Wren (Ed.), Computer Schedulng of Publc Transport Urban Passen-

12 B A C A (a) h = B A (b) h = B A (c) h = 2 (d) h = 3 Fgure : Encodngs n the herarchy for n = 8. ger Vehcle and Crew Schedulng, North-Holland, 98, pp [2] E. M. L. Beale, J. A. Tomln, Specal facltes n a general mathematcal programmng system for non-convex problems usng ordered sets of varables, n: OR 69: Proceedngs of the Ffth Internatonal Conference on Operatonal Research, pp [3] J. P. Velma, A. B. Keha, G. L. Nemhauser, Nonconvex, lower semcontnuous pecewse lnear optmzaton, Dscrete Optmzaton 5 (2008) [4] J. A. Appleget, R. K. Wood, Explct-constrant branchng for solvng mxed-nteger programs, volume 2 of Operatons Research / Computer Scence Interfaces Seres, Kluwer, 2000, pp [5] J. P. Velma, G. L. Nemhauser, Modelng dsunctve constrants wth a logarthmc number of bnary varables and constrants, Mathematcal Programmng 28 (20) [6] D. L. Brcker, Reformulaton of specal ordered sets for mplct enumeraton algorthms wth applcatons n nonconvex separable programmng, AIIE Transactons 9 (977)

13 [7] E. Y. H. Ln, D. L. Brcker, On the calculaton of true and pseudo penaltes n multple choce nteger programmng, European Journal of Operatonal Research 55 (99) [8] W. Ogryczak, A note on modelng multple choce requrements for smple mxed nteger programmng solvers, Computers & Operatons Research 23 (996) [9] C. C. Souza, L. A. Wolsey, Schedulng Proects wth Labour Constrants, Relatóro Técnco IC-97-22, IC - UNICAMP, 997. [0] C. C. B. Cavalcante, C. C. de Souza, M. W. P. Savelsbergh, Y. Wang, L. A. Wolsey, Schedulng proects wth labor constrants, Dscrete Appled Mathematcs 2 (200) [] R. H. Möhrng, A. S. Schulz, F. Stork, M. Uetz, On proect schedulng wth rregular startng tme costs, Operatons Research Letters 28 (200) [2] E. Y. H. Ln, D. L. Brcker, Connectng specal ordered nequaltes and transformaton and reformulaton technque n multple choce programmng, Computers & Operatons Research 29 (2002) [3] D. C. Sommer, Computatonal experence wth the ophele mxed nteger code, 972. Talk presented at the Internatonal TIMS Conference, Houston. [4] T. Ibarak, Integer programmng formulaton of combnatoral optmzaton problems, Dscrete Mathematcs 6 (976) [5] H. L. L, H. C. Lu, Global optmzaton for generalzed geometrc programs wth mxed free-sgn varables, Operatons Research 57 (2009) [6] J. P. Velma, S. Ahmed, G. L. Nemhauser, Mxed-nteger models for nonseparable pecewse-lnear optmzaton: Unfyng framework and extensons, Operatons Research 58 (200) [7] W. P. Adams, S. M. Henry, Base-2 expansons for lnearzng products of functons of dscrete varables, Operatons Research 60 (202) [8] M. Padberg, Approxmatng separable nonlnear functons va mxed zeroone programs, Operatons Research Letters 27 (2000) 5. 3

14 [9] M. W. Padberg, M. P. Ral, Locaton, Schedulng, Desgn and Integer Programmng, volume 3 of Internatonal Seres n Operatons Research & Management Scence, Sprnger, 996. [20] H. Markowtz, A. Manne, On the soluton of dscrete programmng problems, Econometrca 25 (957) [2] G. B. Dantzg, On the sgnfcance of solvng lnear-programmng problems wth some nteger varables, Econometrca 28 (960) [22] G. B. Dantzg, Lnear Programmng and Extensons, Prnceton Unversty Press, Prnceton, 963. [23] D. L. Wlson, Polyhedral Methods for Pecewse-Lnear Functons, Ph.D. thess, Unversty of Kentucky, Lexngton, KY, USA, 998. [24] J. Luedtke, S. Ahmed, G. L. Nemhauser, An nteger programmng approach for lnear programs wth probablstc constrants, Mathematcal Programmng 22 (200) [25] J. P. Velma, S. Ahmed, G. L. Nemhauser, Mxed nteger lnear programmng formulatons for probablstc constrants, Operatons Research Letters 40 (202) [26] J. K. Lowe, Modellng wth Integer Varables, Ph.D. thess, Georga Insttute of Technology, 984. [27] R. G. Jeroslow, J. K. Lowe, Modelng wth nteger varables, Mathematcal Programmng Studes 22 (984) [28] J. P. Velma, Mxed nteger lnear programmng formulaton technques, Optmzaton Onlne (202). HTML/202/07/3539.html. [29] E. Balas, On the convex-hull of the unon of certan polyhedra, Operatons Research Letters 7 (988) [30] R. G. Jeroslow, A smplfcaton for some dsunctve formulatons, European Journal of Operatonal Research 36 (988) 6 2. [3] C. Blar, Representaton for multple rght-hand sdes, Mathematcal Programmng 49 (990) 5. 4

15 [32] B. Geßler, A. Martn, A. Mors, L. Schewe, Usng pecewse lnear functons for solvng MINLPs, n: J. Lee, S. Ley er (Eds.), Mxed Integer Nonlnear Programmng, volume 54 of The IMA Volumes n Mathematcs and ts Applcatons, Sprnger, 202, pp

Incremental and Encoding Formulations for Mixed Integer Programming

Incremental and Encoding Formulations for Mixed Integer Programming Incremental and Encodng Formulatons for Mxed Integer Programmng Sercan Yıldız a, Juan Pablo Velma b,c, a Tepper School of Busness, Carnege Mellon Unversty, 5000 Forbes Ave., Pttsburgh, PA 15213, Unted

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

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

princeton univ. F 17 cos 521: Advanced Algorithm Design Lecture 7: LP Duality Lecturer: Matt Weinberg

princeton univ. F 17 cos 521: Advanced Algorithm Design Lecture 7: LP Duality Lecturer: Matt Weinberg prnceton unv. F 17 cos 521: Advanced Algorthm Desgn Lecture 7: LP Dualty Lecturer: Matt Wenberg Scrbe: LP Dualty s an extremely useful tool for analyzng structural propertes of lnear programs. Whle there

More information

Inner Product. Euclidean Space. Orthonormal Basis. Orthogonal

Inner Product. Euclidean Space. Orthonormal Basis. Orthogonal Inner Product Defnton 1 () A Eucldean space s a fnte-dmensonal vector space over the reals R, wth an nner product,. Defnton 2 (Inner Product) An nner product, on a real vector space X s a symmetrc, blnear,

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

COS 521: Advanced Algorithms Game Theory and Linear Programming

COS 521: Advanced Algorithms Game Theory and Linear Programming COS 521: Advanced Algorthms Game Theory and Lnear Programmng Moses Charkar February 27, 2013 In these notes, we ntroduce some basc concepts n game theory and lnear programmng (LP). We show a connecton

More information

More metrics on cartesian products

More metrics on cartesian products More metrcs on cartesan products If (X, d ) are metrc spaces for 1 n, then n Secton II4 of the lecture notes we defned three metrcs on X whose underlyng topologes are the product topology The purpose of

More information

Lecture 20: Lift and Project, SDP Duality. Today we will study the Lift and Project method. Then we will prove the SDP duality theorem.

Lecture 20: Lift and Project, SDP Duality. Today we will study the Lift and Project method. Then we will prove the SDP duality theorem. prnceton u. sp 02 cos 598B: algorthms and complexty Lecture 20: Lft and Project, SDP Dualty Lecturer: Sanjeev Arora Scrbe:Yury Makarychev Today we wll study the Lft and Project method. Then we wll prove

More information

3.1 Expectation of Functions of Several Random Variables. )' be a k-dimensional discrete or continuous random vector, with joint PMF p (, E X E X1 E X

3.1 Expectation of Functions of Several Random Variables. )' be a k-dimensional discrete or continuous random vector, with joint PMF p (, E X E X1 E X Statstcs 1: Probablty Theory II 37 3 EPECTATION OF SEVERAL RANDOM VARIABLES As n Probablty Theory I, the nterest n most stuatons les not on the actual dstrbuton of a random vector, but rather on a number

More information

The Order Relation and Trace Inequalities for. Hermitian Operators

The Order Relation and Trace Inequalities for. Hermitian Operators Internatonal Mathematcal Forum, Vol 3, 08, no, 507-57 HIKARI Ltd, wwwm-hkarcom https://doorg/0988/mf088055 The Order Relaton and Trace Inequaltes for Hermtan Operators Y Huang School of Informaton Scence

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

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

Yong Joon Ryang. 1. Introduction Consider the multicommodity transportation problem with convex quadratic cost function. 1 2 (x x0 ) T Q(x x 0 )

Yong Joon Ryang. 1. Introduction Consider the multicommodity transportation problem with convex quadratic cost function. 1 2 (x x0 ) T Q(x x 0 ) Kangweon-Kyungk Math. Jour. 4 1996), No. 1, pp. 7 16 AN ITERATIVE ROW-ACTION METHOD FOR MULTICOMMODITY TRANSPORTATION PROBLEMS Yong Joon Ryang Abstract. The optmzaton problems wth quadratc constrants often

More information

Difference Equations

Difference Equations Dfference Equatons c Jan Vrbk 1 Bascs Suppose a sequence of numbers, say a 0,a 1,a,a 3,... s defned by a certan general relatonshp between, say, three consecutve values of the sequence, e.g. a + +3a +1

More information

VARIATION OF CONSTANT SUM CONSTRAINT FOR INTEGER MODEL WITH NON UNIFORM VARIABLES

VARIATION OF CONSTANT SUM CONSTRAINT FOR INTEGER MODEL WITH NON UNIFORM VARIABLES VARIATION OF CONSTANT SUM CONSTRAINT FOR INTEGER MODEL WITH NON UNIFORM VARIABLES BÂRZĂ, Slvu Faculty of Mathematcs-Informatcs Spru Haret Unversty barza_slvu@yahoo.com Abstract Ths paper wants to contnue

More information

NP-Completeness : Proofs

NP-Completeness : Proofs NP-Completeness : Proofs Proof Methods A method to show a decson problem Π NP-complete s as follows. (1) Show Π NP. (2) Choose an NP-complete problem Π. (3) Show Π Π. A method to show an optmzaton problem

More information

Lecture 12: Discrete Laplacian

Lecture 12: Discrete Laplacian Lecture 12: Dscrete Laplacan Scrbe: Tanye Lu Our goal s to come up wth a dscrete verson of Laplacan operator for trangulated surfaces, so that we can use t n practce to solve related problems We are mostly

More information

Maximizing the number of nonnegative subsets

Maximizing the number of nonnegative subsets Maxmzng the number of nonnegatve subsets Noga Alon Hao Huang December 1, 213 Abstract Gven a set of n real numbers, f the sum of elements of every subset of sze larger than k s negatve, what s the maxmum

More information

Problem Set 9 Solutions

Problem Set 9 Solutions Desgn and Analyss of Algorthms May 4, 2015 Massachusetts Insttute of Technology 6.046J/18.410J Profs. Erk Demane, Srn Devadas, and Nancy Lynch Problem Set 9 Solutons Problem Set 9 Solutons Ths problem

More information

Global Optimization of Truss. Structure Design INFORMS J. N. Hooker. Tallys Yunes. Slide 1

Global Optimization of Truss. Structure Design INFORMS J. N. Hooker. Tallys Yunes. Slide 1 Slde 1 Global Optmzaton of Truss Structure Desgn J. N. Hooker Tallys Yunes INFORMS 2010 Truss Structure Desgn Select sze of each bar (possbly zero) to support the load whle mnmzng weght. Bar szes are dscrete.

More information

APPENDIX A Some Linear Algebra

APPENDIX A Some Linear Algebra APPENDIX A Some Lnear Algebra The collecton of m, n matrces A.1 Matrces a 1,1,..., a 1,n A = a m,1,..., a m,n wth real elements a,j s denoted by R m,n. If n = 1 then A s called a column vector. Smlarly,

More information

Randić Energy and Randić Estrada Index of a Graph

Randić Energy and Randić Estrada Index of a Graph EUROPEAN JOURNAL OF PURE AND APPLIED MATHEMATICS Vol. 5, No., 202, 88-96 ISSN 307-5543 www.ejpam.com SPECIAL ISSUE FOR THE INTERNATIONAL CONFERENCE ON APPLIED ANALYSIS AND ALGEBRA 29 JUNE -02JULY 20, ISTANBUL

More information

Simultaneous Optimization of Berth Allocation, Quay Crane Assignment and Quay Crane Scheduling Problems in Container Terminals

Simultaneous Optimization of Berth Allocation, Quay Crane Assignment and Quay Crane Scheduling Problems in Container Terminals Smultaneous Optmzaton of Berth Allocaton, Quay Crane Assgnment and Quay Crane Schedulng Problems n Contaner Termnals Necat Aras, Yavuz Türkoğulları, Z. Caner Taşkın, Kuban Altınel Abstract In ths work,

More information

THE SUMMATION NOTATION Ʃ

THE SUMMATION NOTATION Ʃ Sngle Subscrpt otaton THE SUMMATIO OTATIO Ʃ Most of the calculatons we perform n statstcs are repettve operatons on lsts of numbers. For example, we compute the sum of a set of numbers, or the sum of the

More information

8.4 COMPLEX VECTOR SPACES AND INNER PRODUCTS

8.4 COMPLEX VECTOR SPACES AND INNER PRODUCTS SECTION 8.4 COMPLEX VECTOR SPACES AND INNER PRODUCTS 493 8.4 COMPLEX VECTOR SPACES AND INNER PRODUCTS All the vector spaces you have studed thus far n the text are real vector spaces because the scalars

More information

CHAPTER 5 NUMERICAL EVALUATION OF DYNAMIC RESPONSE

CHAPTER 5 NUMERICAL EVALUATION OF DYNAMIC RESPONSE CHAPTER 5 NUMERICAL EVALUATION OF DYNAMIC RESPONSE Analytcal soluton s usually not possble when exctaton vares arbtrarly wth tme or f the system s nonlnear. Such problems can be solved by numercal tmesteppng

More information

Combining Constraint Programming and Integer Programming

Combining Constraint Programming and Integer Programming Combnng Constrant Programmng and Integer Programmng GLOBAL CONSTRAINT OPTIMIZATION COMPONENT Specal Purpose Algorthm mn c T x +(x- 0 ) x( + ()) =1 x( - ()) =1 FILTERING ALGORITHM COST-BASED FILTERING ALGORITHM

More information

Feature Selection: Part 1

Feature Selection: Part 1 CSE 546: Machne Learnng Lecture 5 Feature Selecton: Part 1 Instructor: Sham Kakade 1 Regresson n the hgh dmensonal settng How do we learn when the number of features d s greater than the sample sze n?

More information

Introduction to Vapor/Liquid Equilibrium, part 2. Raoult s Law:

Introduction to Vapor/Liquid Equilibrium, part 2. Raoult s Law: CE304, Sprng 2004 Lecture 4 Introducton to Vapor/Lqud Equlbrum, part 2 Raoult s Law: The smplest model that allows us do VLE calculatons s obtaned when we assume that the vapor phase s an deal gas, and

More information

Some modelling aspects for the Matlab implementation of MMA

Some modelling aspects for the Matlab implementation of MMA Some modellng aspects for the Matlab mplementaton of MMA Krster Svanberg krlle@math.kth.se Optmzaton and Systems Theory Department of Mathematcs KTH, SE 10044 Stockholm September 2004 1. Consdered optmzaton

More information

Fundamental loop-current method using virtual voltage sources technique for special cases

Fundamental loop-current method using virtual voltage sources technique for special cases Fundamental loop-current method usng vrtual voltage sources technque for specal cases George E. Chatzaraks, 1 Marna D. Tortorel 1 and Anastasos D. Tzolas 1 Electrcal and Electroncs Engneerng Departments,

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

U.C. Berkeley CS294: Beyond Worst-Case Analysis Luca Trevisan September 5, 2017

U.C. Berkeley CS294: Beyond Worst-Case Analysis Luca Trevisan September 5, 2017 U.C. Berkeley CS94: Beyond Worst-Case Analyss Handout 4s Luca Trevsan September 5, 07 Summary of Lecture 4 In whch we ntroduce semdefnte programmng and apply t to Max Cut. Semdefnte Programmng Recall that

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

Affine transformations and convexity

Affine transformations and convexity Affne transformatons and convexty The purpose of ths document s to prove some basc propertes of affne transformatons nvolvng convex sets. Here are a few onlne references for background nformaton: http://math.ucr.edu/

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

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

Lectures - Week 4 Matrix norms, Conditioning, Vector Spaces, Linear Independence, Spanning sets and Basis, Null space and Range of a Matrix

Lectures - Week 4 Matrix norms, Conditioning, Vector Spaces, Linear Independence, Spanning sets and Basis, Null space and Range of a Matrix Lectures - Week 4 Matrx norms, Condtonng, Vector Spaces, Lnear Independence, Spannng sets and Bass, Null space and Range of a Matrx Matrx Norms Now we turn to assocatng a number to each matrx. We could

More information

Numerical Heat and Mass Transfer

Numerical Heat and Mass Transfer Master degree n Mechancal Engneerng Numercal Heat and Mass Transfer 06-Fnte-Dfference Method (One-dmensonal, steady state heat conducton) Fausto Arpno f.arpno@uncas.t Introducton Why we use models 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

Solutions to exam in SF1811 Optimization, Jan 14, 2015

Solutions to exam in SF1811 Optimization, Jan 14, 2015 Solutons to exam n SF8 Optmzaton, Jan 4, 25 3 3 O------O -4 \ / \ / The network: \/ where all lnks go from left to rght. /\ / \ / \ 6 O------O -5 2 4.(a) Let x = ( x 3, x 4, x 23, x 24 ) T, where the varable

More information

CSci 6974 and ECSE 6966 Math. Tech. for Vision, Graphics and Robotics Lecture 21, April 17, 2006 Estimating A Plane Homography

CSci 6974 and ECSE 6966 Math. Tech. for Vision, Graphics and Robotics Lecture 21, April 17, 2006 Estimating A Plane Homography CSc 6974 and ECSE 6966 Math. Tech. for Vson, Graphcs and Robotcs Lecture 21, Aprl 17, 2006 Estmatng A Plane Homography Overvew We contnue wth a dscusson of the major ssues, usng estmaton of plane projectve

More information

Interactive Bi-Level Multi-Objective Integer. Non-linear Programming Problem

Interactive Bi-Level Multi-Objective Integer. Non-linear Programming Problem Appled Mathematcal Scences Vol 5 0 no 65 3 33 Interactve B-Level Mult-Objectve Integer Non-lnear Programmng Problem O E Emam Department of Informaton Systems aculty of Computer Scence and nformaton Helwan

More information

A new construction of 3-separable matrices via an improved decoding of Macula s construction

A new construction of 3-separable matrices via an improved decoding of Macula s construction Dscrete Optmzaton 5 008 700 704 Contents lsts avalable at ScenceDrect Dscrete Optmzaton journal homepage: wwwelsevercom/locate/dsopt A new constructon of 3-separable matrces va an mproved decodng of Macula

More information

Assortment Optimization under MNL

Assortment Optimization under MNL Assortment Optmzaton under MNL Haotan Song Aprl 30, 2017 1 Introducton The assortment optmzaton problem ams to fnd the revenue-maxmzng assortment of products to offer when the prces of products are fxed.

More information

Formulas for the Determinant

Formulas for the Determinant page 224 224 CHAPTER 3 Determnants e t te t e 2t 38 A = e t 2te t e 2t e t te t 2e 2t 39 If 123 A = 345, 456 compute the matrx product A adj(a) What can you conclude about det(a)? For Problems 40 43, use

More information

Additional Codes using Finite Difference Method. 1 HJB Equation for Consumption-Saving Problem Without Uncertainty

Additional Codes using Finite Difference Method. 1 HJB Equation for Consumption-Saving Problem Without Uncertainty Addtonal Codes usng Fnte Dfference Method Benamn Moll 1 HJB Equaton for Consumpton-Savng Problem Wthout Uncertanty Before consderng the case wth stochastc ncome n http://www.prnceton.edu/~moll/ HACTproect/HACT_Numercal_Appendx.pdf,

More information

n α j x j = 0 j=1 has a nontrivial solution. Here A is the n k matrix whose jth column is the vector for all t j=0

n α j x j = 0 j=1 has a nontrivial solution. Here A is the n k matrix whose jth column is the vector for all t j=0 MODULE 2 Topcs: Lnear ndependence, bass and dmenson We have seen that f n a set of vectors one vector s a lnear combnaton of the remanng vectors n the set then the span of the set s unchanged f that vector

More information

Graph Reconstruction by Permutations

Graph Reconstruction by Permutations Graph Reconstructon by Permutatons Perre Ille and Wllam Kocay* Insttut de Mathémathques de Lumny CNRS UMR 6206 163 avenue de Lumny, Case 907 13288 Marselle Cedex 9, France e-mal: lle@ml.unv-mrs.fr Computer

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

FACTORIZATION IN KRULL MONOIDS WITH INFINITE CLASS GROUP

FACTORIZATION IN KRULL MONOIDS WITH INFINITE CLASS GROUP C O L L O Q U I U M M A T H E M A T I C U M VOL. 80 1999 NO. 1 FACTORIZATION IN KRULL MONOIDS WITH INFINITE CLASS GROUP BY FLORIAN K A I N R A T H (GRAZ) Abstract. Let H be a Krull monod wth nfnte class

More information

Section 8.3 Polar Form of Complex Numbers

Section 8.3 Polar Form of Complex Numbers 80 Chapter 8 Secton 8 Polar Form of Complex Numbers From prevous classes, you may have encountered magnary numbers the square roots of negatve numbers and, more generally, complex numbers whch are the

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

Linear, affine, and convex sets and hulls In the sequel, unless otherwise specified, X will denote a real vector space.

Linear, affine, and convex sets and hulls In the sequel, unless otherwise specified, X will denote a real vector space. Lnear, affne, and convex sets and hulls In the sequel, unless otherwse specfed, X wll denote a real vector space. Lnes and segments. Gven two ponts x, y X, we defne xy = {x + t(y x) : t R} = {(1 t)x +

More information

Lecture 17 : Stochastic Processes II

Lecture 17 : Stochastic Processes II : Stochastc Processes II 1 Contnuous-tme stochastc process So far we have studed dscrete-tme stochastc processes. We studed the concept of Makov chans and martngales, tme seres analyss, and regresson analyss

More information

Integrals and Invariants of Euler-Lagrange Equations

Integrals and Invariants of Euler-Lagrange Equations Lecture 16 Integrals and Invarants of Euler-Lagrange Equatons ME 256 at the Indan Insttute of Scence, Bengaluru Varatonal Methods and Structural Optmzaton G. K. Ananthasuresh Professor, Mechancal Engneerng,

More information

THE CHINESE REMAINDER THEOREM. We should thank the Chinese for their wonderful remainder theorem. Glenn Stevens

THE CHINESE REMAINDER THEOREM. We should thank the Chinese for their wonderful remainder theorem. Glenn Stevens THE CHINESE REMAINDER THEOREM KEITH CONRAD We should thank the Chnese for ther wonderful remander theorem. Glenn Stevens 1. Introducton The Chnese remander theorem says we can unquely solve any par of

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

Vapnik-Chervonenkis theory

Vapnik-Chervonenkis theory Vapnk-Chervonenks theory Rs Kondor June 13, 2008 For the purposes of ths lecture, we restrct ourselves to the bnary supervsed batch learnng settng. We assume that we have an nput space X, and an unknown

More information

Edge Isoperimetric Inequalities

Edge Isoperimetric Inequalities November 7, 2005 Ross M. Rchardson Edge Isopermetrc Inequaltes 1 Four Questons Recall that n the last lecture we looked at the problem of sopermetrc nequaltes n the hypercube, Q n. Our noton of boundary

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

Example: (13320, 22140) =? Solution #1: The divisors of are 1, 2, 3, 4, 5, 6, 9, 10, 12, 15, 18, 20, 27, 30, 36, 41,

Example: (13320, 22140) =? Solution #1: The divisors of are 1, 2, 3, 4, 5, 6, 9, 10, 12, 15, 18, 20, 27, 30, 36, 41, The greatest common dvsor of two ntegers a and b (not both zero) s the largest nteger whch s a common factor of both a and b. We denote ths number by gcd(a, b), or smply (a, b) when there s no confuson

More information

The Multiple Classical Linear Regression Model (CLRM): Specification and Assumptions. 1. Introduction

The Multiple Classical Linear Regression Model (CLRM): Specification and Assumptions. 1. Introduction ECONOMICS 5* -- NOTE (Summary) ECON 5* -- NOTE The Multple Classcal Lnear Regresson Model (CLRM): Specfcaton and Assumptons. Introducton CLRM stands for the Classcal Lnear Regresson Model. The CLRM s also

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

The Quadratic Trigonometric Bézier Curve with Single Shape Parameter

The Quadratic Trigonometric Bézier Curve with Single Shape Parameter J. Basc. Appl. Sc. Res., (3541-546, 01 01, TextRoad Publcaton ISSN 090-4304 Journal of Basc and Appled Scentfc Research www.textroad.com The Quadratc Trgonometrc Bézer Curve wth Sngle Shape Parameter Uzma

More information

= z 20 z n. (k 20) + 4 z k = 4

= z 20 z n. (k 20) + 4 z k = 4 Problem Set #7 solutons 7.2.. (a Fnd the coeffcent of z k n (z + z 5 + z 6 + z 7 + 5, k 20. We use the known seres expanson ( n+l ( z l l z n below: (z + z 5 + z 6 + z 7 + 5 (z 5 ( + z + z 2 + z + 5 5

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

ON A DETERMINATION OF THE INITIAL FUNCTIONS FROM THE OBSERVED VALUES OF THE BOUNDARY FUNCTIONS FOR THE SECOND-ORDER HYPERBOLIC EQUATION

ON A DETERMINATION OF THE INITIAL FUNCTIONS FROM THE OBSERVED VALUES OF THE BOUNDARY FUNCTIONS FOR THE SECOND-ORDER HYPERBOLIC EQUATION Advanced Mathematcal Models & Applcatons Vol.3, No.3, 2018, pp.215-222 ON A DETERMINATION OF THE INITIAL FUNCTIONS FROM THE OBSERVED VALUES OF THE BOUNDARY FUNCTIONS FOR THE SECOND-ORDER HYPERBOLIC EUATION

More information

Geometry of Müntz Spaces

Geometry of Müntz Spaces WDS'12 Proceedngs of Contrbuted Papers, Part I, 31 35, 212. ISBN 978-8-7378-224-5 MATFYZPRESS Geometry of Müntz Spaces P. Petráček Charles Unversty, Faculty of Mathematcs and Physcs, Prague, Czech Republc.

More information

On the correction of the h-index for career length

On the correction of the h-index for career length 1 On the correcton of the h-ndex for career length by L. Egghe Unverstet Hasselt (UHasselt), Campus Depenbeek, Agoralaan, B-3590 Depenbeek, Belgum 1 and Unverstet Antwerpen (UA), IBW, Stadscampus, Venusstraat

More information

Global Optimization of Bilinear Generalized Disjunctive Programs

Global Optimization of Bilinear Generalized Disjunctive Programs Global Optmzaton o Blnear Generalzed Dsunctve Programs Juan Pablo Ruz Ignaco E. Grossmann Department o Chemcal Engneerng Center or Advanced Process Decson-mang Unversty Pttsburgh, PA 15213 1 Non-Convex

More information

Appendix for Causal Interaction in Factorial Experiments: Application to Conjoint Analysis

Appendix for Causal Interaction in Factorial Experiments: Application to Conjoint Analysis A Appendx for Causal Interacton n Factoral Experments: Applcaton to Conjont Analyss Mathematcal Appendx: Proofs of Theorems A. Lemmas Below, we descrbe all the lemmas, whch are used to prove the man theorems

More information

MASSACHUSETTS INSTITUTE OF TECHNOLOGY 6.265/15.070J Fall 2013 Lecture 12 10/21/2013. Martingale Concentration Inequalities and Applications

MASSACHUSETTS INSTITUTE OF TECHNOLOGY 6.265/15.070J Fall 2013 Lecture 12 10/21/2013. Martingale Concentration Inequalities and Applications MASSACHUSETTS INSTITUTE OF TECHNOLOGY 6.65/15.070J Fall 013 Lecture 1 10/1/013 Martngale Concentraton Inequaltes and Applcatons Content. 1. Exponental concentraton for martngales wth bounded ncrements.

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

Supplementary Notes for Chapter 9 Mixture Thermodynamics

Supplementary Notes for Chapter 9 Mixture Thermodynamics Supplementary Notes for Chapter 9 Mxture Thermodynamcs Key ponts Nne major topcs of Chapter 9 are revewed below: 1. Notaton and operatonal equatons for mxtures 2. PVTN EOSs for mxtures 3. General effects

More information

CS : Algorithms and Uncertainty Lecture 17 Date: October 26, 2016

CS : Algorithms and Uncertainty Lecture 17 Date: October 26, 2016 CS 29-128: Algorthms and Uncertanty Lecture 17 Date: October 26, 2016 Instructor: Nkhl Bansal Scrbe: Mchael Denns 1 Introducton In ths lecture we wll be lookng nto the secretary problem, and an nterestng

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

System in Weibull Distribution

System in Weibull Distribution Internatonal Matheatcal Foru 4 9 no. 9 94-95 Relablty Equvalence Factors of a Seres-Parallel Syste n Webull Dstrbuton M. A. El-Dacese Matheatcs Departent Faculty of Scence Tanta Unversty Tanta Egypt eldacese@yahoo.co

More information

LOW BIAS INTEGRATED PATH ESTIMATORS. James M. Calvin

LOW BIAS INTEGRATED PATH ESTIMATORS. James M. Calvin Proceedngs of the 007 Wnter Smulaton Conference S G Henderson, B Bller, M-H Hseh, J Shortle, J D Tew, and R R Barton, eds LOW BIAS INTEGRATED PATH ESTIMATORS James M Calvn Department of Computer Scence

More information

Perron Vectors of an Irreducible Nonnegative Interval Matrix

Perron Vectors of an Irreducible Nonnegative Interval Matrix Perron Vectors of an Irreducble Nonnegatve Interval Matrx Jr Rohn August 4 2005 Abstract As s well known an rreducble nonnegatve matrx possesses a unquely determned Perron vector. As the man result of

More information

Perfect Competition and the Nash Bargaining Solution

Perfect Competition and the Nash Bargaining Solution Perfect Competton and the Nash Barganng Soluton Renhard John Department of Economcs Unversty of Bonn Adenauerallee 24-42 53113 Bonn, Germany emal: rohn@un-bonn.de May 2005 Abstract For a lnear exchange

More information

Chapter 5 Multilevel Models

Chapter 5 Multilevel Models Chapter 5 Multlevel Models 5.1 Cross-sectonal multlevel models 5.1.1 Two-level models 5.1.2 Multple level models 5.1.3 Multple level modelng n other felds 5.2 Longtudnal multlevel models 5.2.1 Two-level

More information

CHAPTER III Neural Networks as Associative Memory

CHAPTER III Neural Networks as Associative Memory CHAPTER III Neural Networs as Assocatve Memory Introducton One of the prmary functons of the bran s assocatve memory. We assocate the faces wth names, letters wth sounds, or we can recognze the people

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

Valuated Binary Tree: A New Approach in Study of Integers

Valuated Binary Tree: A New Approach in Study of Integers Internatonal Journal of Scentfc Innovatve Mathematcal Research (IJSIMR) Volume 4, Issue 3, March 6, PP 63-67 ISS 347-37X (Prnt) & ISS 347-34 (Onlne) wwwarcournalsorg Valuated Bnary Tree: A ew Approach

More information

Introductory Cardinality Theory Alan Kaylor Cline

Introductory Cardinality Theory Alan Kaylor Cline Introductory Cardnalty Theory lan Kaylor Clne lthough by name the theory of set cardnalty may seem to be an offshoot of combnatorcs, the central nterest s actually nfnte sets. Combnatorcs deals wth fnte

More information

Optimization Methods for Engineering Design. Logic-Based. John Hooker. Turkish Operational Research Society. Carnegie Mellon University

Optimization Methods for Engineering Design. Logic-Based. John Hooker. Turkish Operational Research Society. Carnegie Mellon University Logc-Based Optmzaton Methods for Engneerng Desgn John Hooker Carnege Mellon Unerst Turksh Operatonal Research Socet Ankara June 1999 Jont work wth: Srnas Bollapragada General Electrc R&D Omar Ghattas Cl

More information

Bayesian predictive Configural Frequency Analysis

Bayesian predictive Configural Frequency Analysis Psychologcal Test and Assessment Modelng, Volume 54, 2012 (3), 285-292 Bayesan predctve Confgural Frequency Analyss Eduardo Gutérrez-Peña 1 Abstract Confgural Frequency Analyss s a method for cell-wse

More information

REGULAR POSITIVE TERNARY QUADRATIC FORMS. 1. Introduction

REGULAR POSITIVE TERNARY QUADRATIC FORMS. 1. Introduction REGULAR POSITIVE TERNARY QUADRATIC FORMS BYEONG-KWEON OH Abstract. A postve defnte quadratc form f s sad to be regular f t globally represents all ntegers that are represented by the genus of f. In 997

More information

arxiv: v1 [math.ho] 18 May 2008

arxiv: v1 [math.ho] 18 May 2008 Recurrence Formulas for Fbonacc Sums Adlson J. V. Brandão, João L. Martns 2 arxv:0805.2707v [math.ho] 8 May 2008 Abstract. In ths artcle we present a new recurrence formula for a fnte sum nvolvng the Fbonacc

More information

Research Article Green s Theorem for Sign Data

Research Article Green s Theorem for Sign Data Internatonal Scholarly Research Network ISRN Appled Mathematcs Volume 2012, Artcle ID 539359, 10 pages do:10.5402/2012/539359 Research Artcle Green s Theorem for Sgn Data Lous M. Houston The Unversty of

More information

Foundations of Arithmetic

Foundations of Arithmetic Foundatons of Arthmetc Notaton We shall denote the sum and product of numbers n the usual notaton as a 2 + a 2 + a 3 + + a = a, a 1 a 2 a 3 a = a The notaton a b means a dvdes b,.e. ac = b where c s an

More information

Linear Regression Analysis: Terminology and Notation

Linear Regression Analysis: Terminology and Notation ECON 35* -- Secton : Basc Concepts of Regresson Analyss (Page ) Lnear Regresson Analyss: Termnology and Notaton Consder the generc verson of the smple (two-varable) lnear regresson model. It s represented

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

Solutions HW #2. minimize. Ax = b. Give the dual problem, and make the implicit equality constraints explicit. Solution.

Solutions HW #2. minimize. Ax = b. Give the dual problem, and make the implicit equality constraints explicit. Solution. Solutons HW #2 Dual of general LP. Fnd the dual functon of the LP mnmze subject to c T x Gx h Ax = b. Gve the dual problem, and make the mplct equalty constrants explct. Soluton. 1. The Lagrangan s L(x,

More information

2.3 Nilpotent endomorphisms

2.3 Nilpotent endomorphisms s a block dagonal matrx, wth A Mat dm U (C) In fact, we can assume that B = B 1 B k, wth B an ordered bass of U, and that A = [f U ] B, where f U : U U s the restrcton of f to U 40 23 Nlpotent endomorphsms

More information

2E Pattern Recognition Solutions to Introduction to Pattern Recognition, Chapter 2: Bayesian pattern classification

2E Pattern Recognition Solutions to Introduction to Pattern Recognition, Chapter 2: Bayesian pattern classification E395 - Pattern Recognton Solutons to Introducton to Pattern Recognton, Chapter : Bayesan pattern classfcaton Preface Ths document s a soluton manual for selected exercses from Introducton to Pattern Recognton

More information

find (x): given element x, return the canonical element of the set containing x;

find (x): given element x, return the canonical element of the set containing x; COS 43 Sprng, 009 Dsjont Set Unon Problem: Mantan a collecton of dsjont sets. Two operatons: fnd the set contanng a gven element; unte two sets nto one (destructvely). Approach: Canoncal element method:

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