Variants of Quantified Linear Programming and Quantified Linear Implication

Size: px
Start display at page:

Download "Variants of Quantified Linear Programming and Quantified Linear Implication"

Transcription

1 Varants of Quantfed Lnear Programmng and Quantfed Lnear Implcaton Potr Wojcechowsk LDCSEE West Vrgna Unversty Morgantown, WV Pavlos Ernaks DMST Athens Unversty of Economcs and Busness Athens, Greece K. Subraman LDCSEE West Vrgna Unversty Morgantown, WV Abstract A Quantfed Lnear Program (QLP) conssts of a set of lnear nequaltes and a correspondng quantfer strng, n whch each unversally quantfed varable s bounded. By extendng the quantfcaton of varables to mplcatons of two lnear systems, we explore Quantfed Lnear Implcatons (QLIs). QLPs and QLIs offer a rch language that s deal for expressng specfcatons n real-tme schedulng and for modellng reactve systems. In ths paper, we show that the varants of QLP and QLI that arse when the unversally quantfed varables are partally bounded or unbounded can be decded n polynomal tme. Moreover, we show that for each class of the polynomal herarchy (PH), there exsts a form of QLI that s complete for that class, thus provdng a contnuous analogue of the way Quantfed Boolean Formulas cover the PH. We also prove that the generc QLI problem s PSPACE-complete and solve some open problems on QLIs wth one quantfer alternaton. 1 Introducton Quantfed lnear programmng s the problem of checkng whether a lnear system s satsfable wth respect to a gven quantfer strng that specfes whch varables are exstentally quantfed and whch are unversally quantfed. Hence, quantfed lnear programmng s a (non-trval) generalzaton of lnear programmng. Accordngly, a Quantfed Lnear Program (QLP) conssts of a set of lnear nequaltes and a correspondng quantfer strng, n whch a bounded regon of values s also specfed for each unversally quantfed varable (Subraman, 2007). A study on QLPs wth unbounded unversally quantfed varables appears n (Rugger et al., 2013). The problem that arses by extendng the quantfcaton of varables to mplcatons of Ths research was supported n part by the Natonal Scence Foundaton through Award CCF and Award CCF , and Ar Force Offce of Scentfc Research through Award FA Ths research has been co-fnanced by the European Unon (European Socal Fund ESF) and Greek natonal funds through the Operatonal Program Educaton and Lfelong Learnng of the Natonal Strategc Reference Framework (NSRF) - Research Fundng Program: Thales. Investng n knowledge socety through the European Socal Fund. two lnear systems s called Quantfed Lnear Implcaton (QLI) (Ernaks et al., 2012a; Ernaks et al., 2013). QLPs represent a rch language that s deal for expressng schedulablty specfcatons n real-tme schedulng (Gerber et al., 1995; Cho and Agrawala, 2000). In real-tme schedulng, however, t may be the case that the dspatcher has already obtaned a schedule (soluton) but then some constrants are slghtly altered. QLIs can be then utlzed to decde whether the dspatcher needs to recompute a soluton or can stll use the current one. Moreover, QLPs and QLIs can be used to model reactve systems (Koo et al., 1999; Pftzmann and Wadner, 2000; Kam et al., 2001; Hall, 2002), where the unversally quantfed varables represent the envronmental nput, whle the exstentally quantfed varables represent the system s response. In ths paper, we examne some new varants of QLPs and QLIs, and establsh ther computatonal complextes. We also settle several open questons n the correspondng lterature on QLIs. More specfcally, we prove that the varants of QLP and QLI that arse when the unversally quantfed varables are partally bounded or unbounded can be decded n polynomal tme. Moreover, wth respect to each class of the polynomal herarchy (PH), we show that there exsts some nstantaton to the QLI framework that s complete for ths class. Ths s nterestng due to the fact that QLIs cover the PH usng only contnuous (and not dscrete) varables, thus provdng a contnuous analogue to the results of Stockmeyer (1977), where the PH s generated usng Quantfed Boolean Formulas (QBFs). We also show that the generc QLI problem s PSPACE-complete. Fnally, we examne the computatonal complextes of some classes of QLIs wth one quantfer alternaton not dentfed by Ernaks et al. (2013). The rest of ths paper s organzed as follows: Secton 2 formally descrbes the problems that we consder n ths paper. Sectons 3 and 4 examne polynomally-solvable varants of QLPs. Secton 5 answers several open questons concernng QLIs, whle Secton 6 concludes the paper. 2 Statement of Problems A QLP s a conjunctve system of lnear constrants n whch each varable s ether exstentally or unversally quantfed accordng to a gven quantfer strng: x 1 y 1 [l 1, u 1 ]... x n y n [l n, u n ] A x + N y b

2 where x 1... x n s a partton of x wth, possbly, x 1 empty; y 1... y n s a partton of y wth, possbly, y n empty; and l, u are lower and upper bounds n R for y, = 1,..., n. Note that n a QLP each unversally quantfed varable s bounded from above and below. Let us ntroduce two varants of QLP both of whch change the nature of the bounds on the unversal varables. A Partally bounded Quantfed Lnear Program (PQLP) s a QLP n whch each unversally quantfed varable s only bounded on one sde. Wthout loss of generalty, we can assume ths sngle bound forces each such varable to be nonnegatve: x 1 y 1 [0, )... x n y n [0, ) A x + N y b An Unbounded Quantfed Lnear Program (UQLP), also studed by Rugger et al. (2013), s a QLP where there are no bounds on any unversal varable: x 1 y 1... x n y n A x + N y b QLIs extend the noton of ncluson of two lnear systems to arbtrary quantfers: x 1 y 1... x n y n [A x + N y b C x + M y d] We say that a QLI holds f t s true as a frst-order formula over the doman of the reals. The decson problem for a QLI conssts of checkng whether t holds or not. A nomenclature s ntroduced n (Ernaks et al., 2012a) to represent the classes of QLIs. Consder a trple A, Q, R. Let A denote the number of quantfer alternatons n the quantfer strng and Q the frst quantfer. Also, let R be an (A + 1)-character strng, whch specfes whether each quantfed set of varables n the quantfer strng appears on the Left, on the Rght, or on Both sdes of the mplcaton. For nstance, 1,, BL ndcates a problem descrbed by: y x [A x + N y b M y d] We extend the noton of partally bounded and unbounded unversally quantfed varables to QLIs. Note that due to the nature of mplcatons, only the constrants n the Left Hand Sde (LHS) restrct the values that unversally quantfed varables can take. Hence, a Partally bounded Quantfed Lnear Implcaton (PQLI) s a QLI n whch each unversally quantfed varable s only bounded by a sngle absolute constrant. Smlarly to PQLPs, we can assume (wthout loss of generalty) that ths sngle constrant s a nonnegatvty constrant. Thus, a PQLI has the followng form: x 1 y 1... x n y n [(A x b y 0) C x + M y d] Accordngly, an Unbounded Quantfed Lnear Implcaton (UQLI) s a QLI n whch each unversally quantfed varable does not appear n the LHS of the mplcaton at all. Thus, a UQLI has the followng form: x 1 y 1... x n y n [A x b C x + M y d] It s mportant to note that 2-person game semantcs have been proposed both for QLPs (Subraman, 2007) and QLIs (Ernaks et al., 2013). More specfcally, for QLPs such a game ncludes an exstental player X, who chooses values for the exstentally quantfed varables, and a unversal player Y, who chooses values for the unversally quantfed varables. X and Y make ther choces accordng to the order of the varables n the quantfer strng. If, at the end, the nstantated lnear system n the QLP s true, then X wns the game (and we say that X has a wnnng strategy). Otherwse, Y wns the game (and we say that Y has a wnnng strategy). In the case of QLIs, an exstental player X and a unversal player Y also choose ther moves accordng to the order of the varables n the quantfer strng. In any game of ths form, the goals of the players are the followng: X selects the values of the exstentally quantfed varables so as to volate the constrants n the LHS or to satsfy the constrants n the Rght-Hand Sde (RHS) of the mplcaton. On the other hand, Y selects the values of the unversally quantfed varables so as to satsfy the constrants of the LHS and at the same tme to volate the constrants of the RHS of the mplcaton. X wns the game (and has a wnnng strategy) f at the end of the game the nstantated mplcaton n the QLI s true. Otherwse, Y wns the game (and has a wnnng strategy). In both cases, the QLP or the QLI holds precsely when X has a wnnng strategy. It s apparent that these game semantcs can be readly appled to the aforementoned varants. 3 Partally bounded varants In ths secton, we utlze the noton of drecton of a convex set (Ignzo and Cavaler, 1993) to show that both PQLP and PQLI are n P. Frst, we obtan an ntermedate result for PQLPs. Theorem 3.1 PQLP wth a fxed number of quantfer alternatons s n P. Proof: Let L n denote the followng PQLP problem: x n y n [0, ) x n 1 y n 1 [0, )... x 1 n n y 1 [0, ) x 0 (A x ) + (B y ) c =0 For n = 0, L 0 s a Lnear Program (LP), hence n P (Khachyan, 1979). Moreover, assume that L k s n P and consder the problem L : x y [0, ) x k y k [0, )... x 1 y 1 [0, ) x 0 (A x ) + (B y ) c (1) =0 We examne whether t s possble to choose n P an x that makes (1) feasble. Such an x needs to handle the case where y = 0. Thus, the followng nstance of L k needs to be feasble: x y x k y k [0, )... x 1 y 1 [0, ) (A x ) + (B y ) c, y = 0 x 0 =0

3 Moreover, y needs to be unbounded. Let m denote the dmenson of y and each e j, j = 1... m, denote a drecton of the system after x s chosen. Note that the drectons of that system do not depend on the value of x. Ths s because the choce of x only changes the value of the constant term and hence does not affect whether y = e j s a drecton or not. So we only need to check whether the followng m nstances of L k are feasble: y x k y k [0, )... x 1 y 1 [0, ) x 0 k (A x ) + (B y ) 0, y = e j : j = 1... m =0 Thus, the feasblty of (1) can be decded based on the feasblty of m + 1 nstances of L k, whch have been assumed to be n P; so, L s also n P. The nductve proof of Theorem 3.1 mples a recursve procedure for decdng any PQLP wth a fnte number of quantfer alternatons n P. However, f the number of alternatons s unbounded, an exponental number of sub-problems s generated. However, many of these subproblems are superfluous and hence need not be calculated. Next, we show that even f the number of alternatons s unbounded, only polynomally-many LPs are generated. Theorem 3.2 PQLP s n P. Proof: Let L be the followng PQLP problem: x n y n [0, ) x n 1 y n 1 [0, )... x 1 y 1 [0, ) x 0 A x + B y c (2) Usng the recursve method mpled by the proof of Theorem 3.1, L would take exponental tme to solve snce 2 n sub-problems are generated. We show that many of these sub-problems are the same and so they can be precomputed. Each tme we check whether y = 1 s a drecton of the system, we always compute: y x 1 y 1 [0, )... x 1 y 1 [0, ) x 0 A x + B y 0 : y = 1 (3) Ths s because pror choces only change the value of the constant term and hence do not affect whether y = 1 s a drecton or not. Checkng whether y = 1 s a drecton wll nvolve subsequently checkng whether y 1 = 1 through y 1 = 1 are also drectons; ths can be handled by performng these computatons frst. Thus, we frst check whether y 1 = 1 s a drecton by solvng: y 1 x 0 A x + B y 0 : y 1 = 1 Then, when computng whether y 2 = 1 s a drecton, we need to solve: y 2 x 1 y 1 [0, ) x 0 A x + B y 0 : y 2 = 1 Snce we already know that y 1 = 1 s a drecton, we only need to solve: y 2 x 1 y 1 x 0 A x + B y 0 : y 2 = 1, y 1 = 0 Now assume that we know that y 1 = 1 through y 1 = 1 are drectons. Recall that n order to check whether y = 1 s a drecton, we need to solve (3). Snce y 1 = 1 s a drecton, we only need to solve: y x 1 y 1... x 1 y 1 [0, ) x 0 A x + B y 0 : y = 1, y 1 = 0 Applyng the same for y 2 = 1 through y 1 = 1, we get: y x 1 y 1... x 1 y 1 x 0 A x + B y 0 : y = 1, y 1,..., y 1 = 0 Thus, to show that each y = 1 s a drecton, we only need to solve n LPs, each of the form above. Let us now show how L can be solved n P. Snce y n = 1 s a drecton, we only need to check: x n y n x n 1 y n 1 [0, )... x 1 y 1 [0, ) x 0 A x + B y c : y n = 0 Snce y n 1 = 1 through y 1 = 1 are also drectons, the problem can be smplfed down to: x n y n x n 1 y n 1... x 1 y 1 x 0 A x + B y c : y = 0 (4) Thus, to solve L, we only need to solve n + 1 LPs. Furthermore, we can use these LPs to generate the functon whch determnes the value of each x. Let ˆx n contan the values of x that solve (4), whch s equvalent to ˆx n satsfyng: x n x n 1... x 1 x 0 A x c (5) Moreover, for each = 1,..., n, let ˆx 1 contan the values of x that solve: y x 1 y 1... x 1 y 1 x 0 A x + B y 0 : y = 1, y 1... y 1 = 0 whch s equvalent to ˆx 1 satsfyng: x 1... x 1 x 0 A x + B 0 (6) Now we expand the ˆx 1 s wth zeroes so that for each ˆx 1, = 1,..., n, we have ˆx 1 = x. Then, the followng functon can be used to determne the values of x: n x = ˆx n + y ˆx 1 (7) Note that the value that each x takes depends only on the y j s that precede t n the quantfer strng. But then: ( ) n n A x + B y = A ˆx n + y ˆx 1 + B y = A ˆx n + n y (A ˆx 1 + B ) c + n y 0 = c snce A ˆx n c by (5) and A ˆx 1 + B 0 for any = 1,..., n by (6). But ths means that the orgnal PQLP (2) can also be satsfed by the values gven to x by (7). Note that f nstead of varables y n (2), we have a vectors y, then when checkng for drectons, nstead of solvng one LP, we need to solve y LPs. Thus, n that stuaton the total number of LPs needed to be solved s y +1, whch can also be done n P. We wll utlze ths result to show that PQLI s also n P.

4 Corollary 3.1 PQLI s n P. Proof: Consder the followng PQLI problem: x 1 y 1... x n y n [A x b, y 0 C x + D y f] (8) If x A x b s nfeasble (.e., f A x b contans even one constrant), then (8) s trvally satsfed. Otherwse, (8) s equvalent to: x 1 y 1... x n y n [y 0 C x + D y f] whch can be rewrtten as x 1 y 1 [0, )... x n y n [0, ) C x + D y f The latter s a PQLP and hence n P by Theorem Unbounded varants In ths secton, we show that UQLP and UQLI are also n P. Theorem 4.1 UQLP s n P. Proof: Consder the followng UQLP problem: x n y n x n 1 y n 1... x 1 y 1 x 0 A x + B y c (9) To show that ths can be solved n polynomal tme, we wll reduce t to a PQLP. We start by creatng addtonal vectors y and y for each y. Then, for each y, we add constrants y = y y. By applyng unversal quantfcaton to y and y and exstental quantfcaton to y, we construct the followng PQLP: x n y n [0, ) y n [0, ) y n x n 1 y n 1 [0, ) y n 1 [0, ) y n 1... x 1 y 1 [0, ) y 1 [0, ) y 1 x 0 A x + B y c, y = y y (10) where y 1,..., y n s a partton of y and y 1,..., y n s a partton of y. Now consder the correspondng 2-person game for (10). The unversal player has full control over the values assumed by y, (whch can take any value n the range (, )), snce y = y y and both y and y are n the range [0, ). Thus, the quantfer sequence y [0, ) y [0, ) y s equvalent to y. Therefore, (10) s equvalent to (9), whch means that UQLP s n P. Observe that the above theorem has been proven ndependently and usng dfferent technques n (Rugger et al., 2013). We wll use ths result to show that UQLI s also solvable n polynomal tme. Corollary 4.1 UQLI s n P. Proof: Consder the followng UQLI problem: x 1 y 1... x n y n [A x b C x + D y f] (11) Smlarly to the proof of Corollary 3.1, f x A x b s nfeasble, then (11) s trvally satsfed. Otherwse, (11) s equvalent to: x 1 y 1... x n y n C x + D y f whch s a UQLP and hence n P by Theorem QLI and the polynomal herarchy In ths secton, we prove that for each class of the PH, there exsts a form of QLI that s complete for that class. Ths s nterestng, snce QLIs use contnuous varables and not dscrete. Hence, we provde a contnuous analogue to the results n Stockmeyer (1977), where the PH s generated usng QBFs. Moreover, we show the generc QLI problem tself s PSPACE-complete. Let B denote the strng B }. {{.. B }. The followng results were obtaned n (Ernaks et al., 2013, Theorems 14 and 15): 1. k,, B wth k odd s Σ P k -hard. 2. k,, B wth k even s Π P k -hard. To establsh the computatonal complextes of k,, B when k s even and k,, B when k s odd, we frst provde a reducton from Q3DNF (see Appendx A) to QLI. Theorem 5.1 Q3DNF can be reduced to QLI. Proof: We wll reduce the Q3DNF problem to a QLI. Consder a Q3DNF nstance Q(x, y) φ(x, y), where Q(x, y) represents the quantfer strng, x s the set of exstentally quantfed varables, y s the set of unversally quantfed varables, and φ s a dsjuncton of 3-lteral terms. We want to produce a correspondng QLI whch wll hold f and only f Q(x, y) φ(x, y) s satsfable. Let E represent the set of constrants on the LHS of the mplcaton and F the set of constrants on the RHS of the constructed mplcaton. For each exstentally quantfed varable x n the nstance of Q3DNF, we add an exstentally quantfed varable x and a unversally quantfed varable r. We also add the constrants r x and r 1 x to E and the constrants r 0 to F. Note that these constrants are equvalent to r mn(x, 1 x ) r 0. Moreover, we add 0 x 1 to F. For each unversally quantfed varable y n the nstance of Q3DNF, we add an exstentally quantfed varable s and a unversally quantfed varable y. We also add the constrants 0 y 1 to E and the constrants 0 s 1, 2y 1 s, and s 2y to F. Note that these are the only constrants that use y varables, snce the clause constrants wll only use x and s varables. For each clause φ j n the nstance of Q3DNF, we add the exstentally quantfed varable w j, and 3 correspondng constrants to F. These constrants ask for w j to be less than or equal to the exstental varables correspondng to the lterals of φ j. Note that these constrants contan only exstental varables. Even n the case of a unversal varable y n φ j, the constrant contans the exstental varable s of the QLI. Negated lterals are also treated approprately. For example, f φ j = (x, y k, x l ), we add w j x, w j s k, and w j 1 x l to F, whle f φ j = (x, ȳ k, x l ), we add w j x, w j 1 s k, and w j 1 x l to F. Furthermore, we add w 1 + w w m 1 to F. We create the quantfer strng of the QLI accordng to Q(x, y): For x n Q(x, y), we ntroduce x r ; for y n Q(x, y), we ntroduce y s ; fnally, we ntroduce w.

5 To complete the proof, we wll show that the constructed QLI holds f and only f Q(x, y) φ(x, y) s satsfable. We start by establshng that all varables that partcpate n the constrants correspondng to the clauses of the Q3DNF nstance are effectvely restrcted to values 0 or 1. To do so, we utlze the game semantcs ntroduced n Secton 2. Frst note that the exstental player X wll only choose x from the set {0, 1}. Ths s because f x [0, 1], then at least one of the constrants n F would be volated. But then the mplcaton would not hold (hence the unversal player Y would wn the game), snce X cannot cause any constrant n E to be volated. On the other hand, f x (0, 1), then Y could choose to set r = mn(x, 1 x ) > 0, whch would cause the mplcaton not to hold (causng the exstental player to lose the game). To sum up, any choce of x {0, 1} would cause X to lose the game. The unversal player Y wll also choose y from the set {0, 1}. We wll show why ths assumpton s not restrctve. Suppose that Y can wn by choosng y [0, 1]. Then at least one constrant of E s volated (.e., 0 y 1) and the mplcaton holds (.e., a contradcton). Suppose now that Y can wn by choosng y (0, 1). If y (0, 1 2 ], then we have that s [0, 2y ]. However, f nstead Y had chosen y = 0, then s {0} [0, 2y ], thus restrctng the possble responses of X. Snce y only appears n the constrants descrbed above, we have that ths s a strctly better move for Y. Smlarly, choosng y = 1 s strctly better for Y than choosng y [ 1 2, 1). To sum up, we can safely assume that the unversal player would only choose y {0, 1}. Snce y s n the set {0, 1}, we have that the exstental player s forced to set s = y. Any other choce of s would volate at least one constrant of F, causng (agan) the exstental player to lose the game. Hence, s varables are also restrcted n the set {0, 1}. Let us show that for the QLI obtaned from a Q3DNF nstance of the form Q(x, y) φ(x, y), the exstental player has a wnnng strategy for the QLI f and only f Q(x, y) φ(x, y) s satsfable. Only-f part. Assume that the exstental player X has a wnnng strategy U for the constructed QLI. Ths means that for every sequence of moves V made by the unversal player Y, the mplcaton holds. From constrant w 1 + w w m 1, we must have that at least one w j > 0. Ths w j corresponds to the clause φ j of the orgnal Q3DNF, and we can assume wlog that t has the form (x, y k, x l ). Snce the x s and s s are restrcted to the set {0, 1}, the constrants w j x, w j s k, and w j x l force each varable to be 1. Thus at least one term of the orgnal Q3DNF formula s satsfed, meanng that the entre formula s. If part. Assume that Q(x, y) φ(x, y) s satsfable. Then there exst values x for x such that x = [c 1, f 1 (y 1 ), f 2 (y 1, y 2 ),..., f n 1 (y 1, y 2,..., y n 1 ] T and for any values y = [y 1, y 2,..., y n ] T gven to y, the Q3SAT expresson s satsfed. Note that f () are Skolem functons and are used to represent that the values of the elements of x depend on the values of the correspondng elements of y. Snce the Q3DNF expresson s satsfed, at least one clause, say φ j, must be satsfed. Consder the constrants constructed from φ j, assumng wlog that t s of the form (x, y k, x l ). Snce φ j s satsfed, we must have that x, y k, x l are all true. Ths means that x = s k = x l = 1. Thus, X can set w j = 1 and the other clause varables to 0, satsfyng the constrant w 1 +w 2 + +w m 1. Snce the x and y varables can be restrcted to the set {0, 1} and snce s = y, we have that for each, r x and r 1 x mply that r 0 (thus satsfyng the correspondng constrant of F ). For the same reason, for any the constrants 0 x 1, 0 y 1, 0 s 1, 2y 1 s, and s 2y are all satsfed. Hence, E F s satsfed and so X has a wnnng strategy for the correspondng QLI. Snce the last quantfer of a Q3DNF formula s always, ths reducton ncreases the number of quantfer alternatons by one and the produced QLI always has as the fnal quantfer. Ths allows us to obtan the followng two results. Corollary 5.1 k,, B wth k even s Σ P k -hard. Proof: Consder the class of Q3DNF formulas wth k quantfers startng wth an exstental one,.e., wth a quantfer strng of the form.... Such a class s Σ k P- complete (the assumpton that k s even s essental). The prevous proof reduces such a class to a QLI wth a quantfer strng obtaned by addng an exstental quantfer at the end, namely to a k,, B formula. Hence, the result. Corollary 5.2 k,, B wth k odd s Π P k -hard. Proof: Consder the class of Q3DNF formulas wth k quantfers startng wth a unversal one,.e., wth the quantfer strng of the form.... Such a class s Π k P- complete (the assumpton that k s odd s essental). The prevous proof reduces such a class to a QLI wth a quantfer strng obtaned by addng an exstental quantfer at the end, namely to a k,, B formula. Hence, the result. Corollares 5.1 and 5.2 when pared wth (Ernaks et al., 2013, Theorems 14 and 15) show the mportance of the fnal quantfer n a QLI. If the fnal quantfer s, then the QLI corresponds to a Q3SAT nstance wth one fewer quantfer alternaton and so covers the correspondng class n the polynomal herarchy. Smlarly, f the fnal quantfer s, then the QLI corresponds to a Q3DNF nstance wth one fewer quantfer alternaton and so covers the correspondng class n the polynomal herarchy. However these results cover only the hardness of these forms of QLI; to show completeness, we need prove to that each of these problems s also contaned wthn ts level of the polynomal herarchy. Theorem 5.2 k,, B s n Σ P k, k,, B s n Π P k. Proof: Sontag (1985) showed that decdng an arbtrary boolean combnaton of lnear constrants under a gven quantfer strng s contaned wthn PSPACE. QLIs are a sub-problem of ths (snce t can be rewrtten as a quantfed conjuncton of the RHS constrants and the negaton of the LHS constrants) and thus also n PSPACE. Hence, QLIs can be decded even f each varable s restrcted to values that are polynomally-szed wth respect to the nput. Consder problem k,, B. After k rounds of each player choosng polynomaly-szed values, the remanng

6 QLI, ether 0,, B or 0,, B, can be solved n P (Ernaks et al., 2013, Theorems 1 and 6). Thus, k,, B s n Σ P k. Smlarly, k,, B s n Π P k. The varous forms of QLI cover the polynomal herarchy as shown n Fgure 1. Moreover, snce QLIs are n general PSPACE-hard (Ernaks et al., 2013, Theorem 5), the proof of Theorem 5.2 has the followng mplcaton. Corollary 5.3 QLI s PSPACE-complete. Π P 2 2,, BBB conp 1,, BB PH. P 0,, B 0,, B Σ P 2 2,, BBB NP 1,, BB Fgure 1: QLIs and the Polynomal Herarchy 5.1 Related Problems In ths secton, we examne the computatonal complextes of varous classes of QLIs wth one quantfer alternaton that were left open n (Ernaks et al., 2013). Corollary 5.4 1,, RB,.e., decdng whether y x [M x n A x + B y c] holds, s n P. Proof: Frst, we check whether x M x n s nfeasble (.e., whether M x n contans even one constrant). If t s, then the QLI trvally holds. If t s not, then y x [M x n A x+b y c] reduces to y x A x+b y c, whch s a UQLP and hence n P by Theorem 4.1. The decson problem for formula x y [A x + B y c M y n] s NP-complete (Ernaks et al., 2012a). However, by movng the exstentally quantfed varables x to the RHS, the problem becomes tractable. Corollary 5.5 1,, BL,.e., decdng whether x y [A x + B y c M x n] holds, s n P. Proof: Frst, we check whether x M x n holds, whch can be done n P (Khachyan, 1979). If x M x n holds, then x y [A x+b y c M x n] trvally holds. If x M x n does not hold, x y [A x + B y c M x n] reduces to x y A x + B y c, whch can be checked to hold n P by Theorem 4.1. Let us turn our attenton to the class 1,, BR and ts super-class 1,, BB, shown conp-hard n (Ernaks et al., 2013). Note that these classes are also n conp by Theorem 5.2. We provde an alternatve proof for the latter. Lemma 5.1 1,, BR and 1,, BB are n conp. Proof: Problems n 1,, BR are descrbed by: y x [N y b C x + M y d] (12) We consder two cases: N y b beng (a) bounded and (b) unbounded. (a) Let N y b be bounded. Then, f (12) does not hold, there wll exst some y for whch x [N y b C x + M y d] does not hold. Note that ths means that y satsfes N y b (otherwse the mplcaton would be trvally true). But then, snce N y b s bounded, there wll also exst some extreme pont y of N y b, such that x [N y b C x + M y d] does not hold. Our clam follows from the fact that y s an extreme pont of N y b and hence polynomally-szed. (b) Now let N y b be unbounded. For each soluton of N y b to satsfy C x + M y d, we need to have that the extreme ponts of N y b satsfy C x + M y d and that the extreme drectons of N y b are drectons of C x + M y d. For the latter, we need to check whether each soluton of N y 0 also satsfes C x + M y 0,.e., we need to decde y x [N y 0 C x + M y 0]. But snce we are searchng for extreme drectons, we can restrct the values of y to the nterval [ 1, 1]. Hence, consder a case where (12) does not hold. We would frst decde y x [N y 0, 1 y 1 C x + M y 0], whch s n conp, snce the LHS s bounded (case (a)). If ths QLI does not hold, we are done. Otherwse, there wll exst a (polynomally-szed) extreme pont y of N y b such that x [N y b C x + M y d] does not hold (case (a)). Let us now consder problem 1,, BB,.e., y x [A x + N y b C x + M y d]. If A 0, ths QLI s trvally true; for any choce of values for the unversally quantfed varables, there wll exst a choce of the exstentally quantfed varables such that the LHS of the QLI s false. Otherwse (f A = 0), problem 1,, BB reduces to 1,, BR, whch was shown above to be n conp. Corollares 5.4 and 5.5, and Lemma 5.1 solve some open problems on the computatonal complextes of QLIs wth one quantfer alternaton (Ernaks et al., 2013). A complete representaton of all classes wth one quantfer alternaton QLIs startng wth an exstental or wth a unversal quantfer s gven n Appendx B, n Fgure 2 and Fgure 3 respectvely. The symmetry s apparent. 6 Concluson In ths paper, we examned several varants of QLP and QLI that arse when the unversally quantfed varables are partally bounded or unbounded and showed that all these varants are n P. Furthermore, we proved that k,, B s Σ P k -complete and k,, B s Π P k -complete. Hence, we showed that any class of the PH can be represented by some nstantaton to the QLI framework (usng only contnuous varables), thus provdng a contnuous analogue to the way QBFs cover the PH. Moreover, we proved that the

7 generc QLI problem s PSPACE-complete. Fnally, we answered several open questons on the computatonal complextes of classes of QLIs wth one quantfer alternaton, thus completng the map of the computatonal complexty of all such classes of QLIs. References Cho, S. and Agrawala, A Dynamc dspatchng of cyclc real-tme tasks wth relatve tmng constrants. Real-Tme Systems 19(1): Gerber, R.; Pugh, W.; and Saksena, M Parametrc dspatchng of hard real-tme tasks. IEEE Transactons on Computng 44(3): Ernaks, P; Rugger, S; Subraman, K; and Wojcechowsk, P Computatonal complextes of ncluson queres over polyhedral sets. Internatonal Symposum on Artfcal Intellgence and Mathematcs, Fort Lauderdale, FL. Ernaks, P; Rugger, S; Subraman, K; and Wojcechowsk, P Quantfed Lnear Implcatons. Annals of Mathematcs and Artfcal Intellgence (n press) - DOI: /s Hall, R Specfcaton, valdaton, and synthess of emal agent controllers: A case study n functon rch reactve system desgn. Automated Software Engneerng 9(3): Ignzo, J.P. and Cavaler, T.P Lnear Programmng. Prentce Hall Kam, N.; Cohen, I.R.; and Harel, D The mmune system as a reactve system: modelng T cell actvaton wth statecharts. In Proceedngs of the 2001 IEEE Symposa on Human-Centrc Computng Languages and Envronments, Khachyan, L.G A polynomal algorthm n lnear programmng. Doklady Akadema Nauk SSSR 224: Koo, T; Snopol, B.; Sangovann-Vncentell, A.; and Sastry, S A formal approach to reactve system desgn: unmanned aeral vehcle flght management system desgn example. In Proceedngs of the 1999 IEEE Internatonal Symposum on Computer Aded Control System Desgn, Pftzmann, B. and Wadner, M Composton and ntegrty preservaton of secure reactve systems. In Proceedngs of the 2000 ACM Conference on Computer Communcatons Securty, Rugger, S; Ernaks, P; Subraman, K; and Wojcechowsk, P On the complexty of quantfed lnear systems. Theoretcal Computer Scence (n press) - DOI: /j.tcs Sontag, E.D Real Addton and the Polynomal Herarchy. Informaton Processng Letters 20(3): Stockmeyer, L.J The Polynomal-Tme Herarchy. Theoretcal Computer Scence 3: Subraman, K On a decson procedure for quantfed lnear programs. Annals of Mathematcs and Artfcal Intellgence 51(1): Appendx A A quantfed boolean expresson s an nstance of Q3DNF f t s a dsjuncton of terms each of whch s a conjuncton of three lterals. Thus, the followng problem s an nstance of Q3DNF. x 1 y 1 x 2 y 2 x 3 y 3 (x 1 y 3 x 2 ) (x 3 y 2 x 1 ) (x 2 y 1 x 3 ) We can defne the two player game semantcs for ths problem as follows. The exstental and unversal players make moves accordng to the quantfer strng. The exstental player wns a term f all the lterals n the term are set to True. Conversely the unversal player wns the term f at least on lteral s set to False. The exstental player wns the game f he wns at least one term n the dsjuncton, thus causng the expresson to evaluate to True. The unversal player wns f he wns every term n the dsjuncton, thus causng the expresson to evaluate to False. Ths can be consdered as a reversal of the players objectves n an nstance of Q3SAT, where the exstental player needs to wn all the clauses, and the unversal player only needs to wn a sngle clause. Ths occurs because the Q3DNF and Q3SAT problems are complements of each other. 1,, BL 1,, LL Appendx B 1,, BB (NP-complete) 1,, LB (NP-complete) 1,, LR 6 1,, BR 1,, RL 1,, RB 1,, RR Fgure 2: Complexty of classes of QLIs. Arrows denote nclusons.

8 1,, BL 1,, LL 1,, BB (conp-complete) 1,, LB 1,, LR 1,, RL 6 1,, BR 1,, RB (conp-complete) 1,, RR Fgure 3: Complexty of classes of QLIs. Arrows denote nclusons.

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

Calculation of time complexity (3%)

Calculation of time complexity (3%) Problem 1. (30%) Calculaton of tme complexty (3%) Gven n ctes, usng exhaust search to see every result takes O(n!). Calculaton of tme needed to solve the problem (2%) 40 ctes:40! dfferent tours 40 add

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

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

Bézier curves. Michael S. Floater. September 10, These notes provide an introduction to Bézier curves. i=0

Bézier curves. Michael S. Floater. September 10, These notes provide an introduction to Bézier curves. i=0 Bézer curves Mchael S. Floater September 1, 215 These notes provde an ntroducton to Bézer curves. 1 Bernsten polynomals Recall that a real polynomal of a real varable x R, wth degree n, s a functon of

More information

Computing Correlated Equilibria in Multi-Player Games

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

More information

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

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

Appendix B: Resampling Algorithms

Appendix B: Resampling Algorithms 407 Appendx B: Resamplng Algorthms A common problem of all partcle flters s the degeneracy of weghts, whch conssts of the unbounded ncrease of the varance of the mportance weghts ω [ ] of the partcles

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

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

arxiv: v1 [cs.ds] 30 Aug 2016

arxiv: v1 [cs.ds] 30 Aug 2016 Dynamc Controllablty of Condtonal Smple Temporal Networks s PSPACE-complete arxv:1608.08545v1 [cs.ds] 30 Aug 2016 Massmo Caro Mathematcs Department Unversty of Trento Trento, Italy massmo.caro@untn.t Abstract

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

Games of Threats. Elon Kohlberg Abraham Neyman. Working Paper

Games of Threats. Elon Kohlberg Abraham Neyman. Working Paper Games of Threats Elon Kohlberg Abraham Neyman Workng Paper 18-023 Games of Threats Elon Kohlberg Harvard Busness School Abraham Neyman The Hebrew Unversty of Jerusalem Workng Paper 18-023 Copyrght 2017

More information

20. Mon, Oct. 13 What we have done so far corresponds roughly to Chapters 2 & 3 of Lee. Now we turn to Chapter 4. The first idea is connectedness.

20. Mon, Oct. 13 What we have done so far corresponds roughly to Chapters 2 & 3 of Lee. Now we turn to Chapter 4. The first idea is connectedness. 20. Mon, Oct. 13 What we have done so far corresponds roughly to Chapters 2 & 3 of Lee. Now we turn to Chapter 4. The frst dea s connectedness. Essentally, we want to say that a space cannot be decomposed

More information

Bezier curves. Michael S. Floater. August 25, These notes provide an introduction to Bezier curves. i=0

Bezier curves. Michael S. Floater. August 25, These notes provide an introduction to Bezier curves. i=0 Bezer curves Mchael S. Floater August 25, 211 These notes provde an ntroducton to Bezer curves. 1 Bernsten polynomals Recall that a real polynomal of a real varable x R, wth degree n, s a functon of the

More information

The Second Anti-Mathima on Game Theory

The Second Anti-Mathima on Game Theory The Second Ant-Mathma on Game Theory Ath. Kehagas December 1 2006 1 Introducton In ths note we wll examne the noton of game equlbrum for three types of games 1. 2-player 2-acton zero-sum games 2. 2-player

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

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

Single-machine scheduling with trade-off between number of tardy jobs and compression cost

Single-machine scheduling with trade-off between number of tardy jobs and compression cost Ths s the Pre-Publshed Verson. Sngle-machne schedulng wth trade-off between number of tardy jobs and compresson cost 1, 2, Yong He 1 Department of Mathematcs, Zhejang Unversty, Hangzhou 310027, P.R. Chna

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

A 2D Bounded Linear Program (H,c) 2D Linear Programming

A 2D Bounded Linear Program (H,c) 2D Linear Programming A 2D Bounded Lnear Program (H,c) h 3 v h 8 h 5 c h 4 h h 6 h 7 h 2 2D Lnear Programmng C s a polygonal regon, the ntersecton of n halfplanes. (H, c) s nfeasble, as C s empty. Feasble regon C s unbounded

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

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

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

Canonical transformations

Canonical transformations Canoncal transformatons November 23, 2014 Recall that we have defned a symplectc transformaton to be any lnear transformaton M A B leavng the symplectc form nvarant, Ω AB M A CM B DΩ CD Coordnate transformatons,

More information

Lecture Space-Bounded Derandomization

Lecture Space-Bounded Derandomization Notes on Complexty Theory Last updated: October, 2008 Jonathan Katz Lecture Space-Bounded Derandomzaton 1 Space-Bounded Derandomzaton We now dscuss derandomzaton of space-bounded algorthms. Here non-trval

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

(1 ) (1 ) 0 (1 ) (1 ) 0

(1 ) (1 ) 0 (1 ) (1 ) 0 Appendx A Appendx A contans proofs for resubmsson "Contractng Informaton Securty n the Presence of Double oral Hazard" Proof of Lemma 1: Assume that, to the contrary, BS efforts are achevable under a blateral

More information

The lower and upper bounds on Perron root of nonnegative irreducible matrices

The lower and upper bounds on Perron root of nonnegative irreducible matrices Journal of Computatonal Appled Mathematcs 217 (2008) 259 267 wwwelsevercom/locate/cam The lower upper bounds on Perron root of nonnegatve rreducble matrces Guang-Xn Huang a,, Feng Yn b,keguo a a College

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

arxiv: v1 [math.co] 1 Mar 2014

arxiv: v1 [math.co] 1 Mar 2014 Unon-ntersectng set systems Gyula O.H. Katona and Dánel T. Nagy March 4, 014 arxv:1403.0088v1 [math.co] 1 Mar 014 Abstract Three ntersecton theorems are proved. Frst, we determne the sze of the largest

More information

Convexity preserving interpolation by splines of arbitrary degree

Convexity preserving interpolation by splines of arbitrary degree Computer Scence Journal of Moldova, vol.18, no.1(52), 2010 Convexty preservng nterpolaton by splnes of arbtrary degree Igor Verlan Abstract In the present paper an algorthm of C 2 nterpolaton of dscrete

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

Finding Primitive Roots Pseudo-Deterministically

Finding Primitive Roots Pseudo-Deterministically Electronc Colloquum on Computatonal Complexty, Report No 207 (205) Fndng Prmtve Roots Pseudo-Determnstcally Ofer Grossman December 22, 205 Abstract Pseudo-determnstc algorthms are randomzed search algorthms

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

Chapter 2 A Class of Robust Solution for Linear Bilevel Programming

Chapter 2 A Class of Robust Solution for Linear Bilevel Programming Chapter 2 A Class of Robust Soluton for Lnear Blevel Programmng Bo Lu, Bo L and Yan L Abstract Under the way of the centralzed decson-makng, the lnear b-level programmng (BLP) whose coeffcents are supposed

More information

On the Repeating Group Finding Problem

On the Repeating Group Finding Problem The 9th Workshop on Combnatoral Mathematcs and Computaton Theory On the Repeatng Group Fndng Problem Bo-Ren Kung, Wen-Hsen Chen, R.C.T Lee Graduate Insttute of Informaton Technology and Management Takmng

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

Economics 101. Lecture 4 - Equilibrium and Efficiency

Economics 101. Lecture 4 - Equilibrium and Efficiency Economcs 0 Lecture 4 - Equlbrum and Effcency Intro As dscussed n the prevous lecture, we wll now move from an envronment where we looed at consumers mang decsons n solaton to analyzng economes full of

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

The equation of motion of a dynamical system is given by a set of differential equations. That is (1)

The equation of motion of a dynamical system is given by a set of differential equations. That is (1) Dynamcal Systems Many engneerng and natural systems are dynamcal systems. For example a pendulum s a dynamcal system. State l The state of the dynamcal system specfes t condtons. For a pendulum n the absence

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

Fuzzy Boundaries of Sample Selection Model

Fuzzy Boundaries of Sample Selection Model Proceedngs of the 9th WSES Internatonal Conference on ppled Mathematcs, Istanbul, Turkey, May 7-9, 006 (pp309-34) Fuzzy Boundares of Sample Selecton Model L. MUHMD SFIIH, NTON BDULBSH KMIL, M. T. BU OSMN

More information

A Local Variational Problem of Second Order for a Class of Optimal Control Problems with Nonsmooth Objective Function

A Local Variational Problem of Second Order for a Class of Optimal Control Problems with Nonsmooth Objective Function A Local Varatonal Problem of Second Order for a Class of Optmal Control Problems wth Nonsmooth Objectve Functon Alexander P. Afanasev Insttute for Informaton Transmsson Problems, Russan Academy of Scences,

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

Game Theory. Lecture Notes By Y. Narahari. Department of Computer Science and Automation Indian Institute of Science Bangalore, India February 2008

Game Theory. Lecture Notes By Y. Narahari. Department of Computer Science and Automation Indian Institute of Science Bangalore, India February 2008 Game Theory Lecture Notes By Y. Narahar Department of Computer Scence and Automaton Indan Insttute of Scence Bangalore, Inda February 2008 Chapter 10: Two Person Zero Sum Games Note: Ths s a only a draft

More information

The Synchronous 8th-Order Differential Attack on 12 Rounds of the Block Cipher HyRAL

The Synchronous 8th-Order Differential Attack on 12 Rounds of the Block Cipher HyRAL The Synchronous 8th-Order Dfferental Attack on 12 Rounds of the Block Cpher HyRAL Yasutaka Igarash, Sej Fukushma, and Tomohro Hachno Kagoshma Unversty, Kagoshma, Japan Emal: {garash, fukushma, hachno}@eee.kagoshma-u.ac.jp

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

NUMERICAL DIFFERENTIATION

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

More information

The Geometry of Logit and Probit

The Geometry of Logit and Probit The Geometry of Logt and Probt Ths short note s meant as a supplement to Chapters and 3 of Spatal Models of Parlamentary Votng and the notaton and reference to fgures n the text below s to those two chapters.

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

Report on Image warping

Report on Image warping Report on Image warpng Xuan Ne, Dec. 20, 2004 Ths document summarzed the algorthms of our mage warpng soluton for further study, and there s a detaled descrpton about the mplementaton of these algorthms.

More information

Errors for Linear Systems

Errors for Linear Systems Errors for Lnear Systems When we solve a lnear system Ax b we often do not know A and b exactly, but have only approxmatons  and ˆb avalable. Then the best thng we can do s to solve ˆx ˆb exactly whch

More information

Learning Theory: Lecture Notes

Learning Theory: Lecture Notes Learnng Theory: Lecture Notes Lecturer: Kamalka Chaudhur Scrbe: Qush Wang October 27, 2012 1 The Agnostc PAC Model Recall that one of the constrants of the PAC model s that the data dstrbuton has to be

More information

CS286r Assign One. Answer Key

CS286r Assign One. Answer Key CS286r Assgn One Answer Key 1 Game theory 1.1 1.1.1 Let off-equlbrum strateges also be that people contnue to play n Nash equlbrum. Devatng from any Nash equlbrum s a weakly domnated strategy. That s,

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

Stanford University CS359G: Graph Partitioning and Expanders Handout 4 Luca Trevisan January 13, 2011

Stanford University CS359G: Graph Partitioning and Expanders Handout 4 Luca Trevisan January 13, 2011 Stanford Unversty CS359G: Graph Parttonng and Expanders Handout 4 Luca Trevsan January 3, 0 Lecture 4 In whch we prove the dffcult drecton of Cheeger s nequalty. As n the past lectures, consder an undrected

More information

arxiv:cs.cv/ Jun 2000

arxiv:cs.cv/ Jun 2000 Correlaton over Decomposed Sgnals: A Non-Lnear Approach to Fast and Effectve Sequences Comparson Lucano da Fontoura Costa arxv:cs.cv/0006040 28 Jun 2000 Cybernetc Vson Research Group IFSC Unversty of São

More information

The Study of Teaching-learning-based Optimization Algorithm

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

More information

= 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

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

U.C. Berkeley CS278: Computational Complexity Professor Luca Trevisan 2/21/2008. Notes for Lecture 8

U.C. Berkeley CS278: Computational Complexity Professor Luca Trevisan 2/21/2008. Notes for Lecture 8 U.C. Berkeley CS278: Computatonal Complexty Handout N8 Professor Luca Trevsan 2/21/2008 Notes for Lecture 8 1 Undrected Connectvty In the undrected s t connectvty problem (abbrevated ST-UCONN) we are gven

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

An Admission Control Algorithm in Cloud Computing Systems

An Admission Control Algorithm in Cloud Computing Systems An Admsson Control Algorthm n Cloud Computng Systems Authors: Frank Yeong-Sung Ln Department of Informaton Management Natonal Tawan Unversty Tape, Tawan, R.O.C. ysln@m.ntu.edu.tw Yngje Lan Management Scence

More information

and problem sheet 2

and problem sheet 2 -8 and 5-5 problem sheet Solutons to the followng seven exercses and optonal bonus problem are to be submtted through gradescope by :0PM on Wednesday th September 08. There are also some practce problems,

More information

Time-Varying Systems and Computations Lecture 6

Time-Varying Systems and Computations Lecture 6 Tme-Varyng Systems and Computatons Lecture 6 Klaus Depold 14. Januar 2014 The Kalman Flter The Kalman estmaton flter attempts to estmate the actual state of an unknown dscrete dynamcal system, gven nosy

More information

CHAPTER-5 INFORMATION MEASURE OF FUZZY MATRIX AND FUZZY BINARY RELATION

CHAPTER-5 INFORMATION MEASURE OF FUZZY MATRIX AND FUZZY BINARY RELATION CAPTER- INFORMATION MEASURE OF FUZZY MATRI AN FUZZY BINARY RELATION Introducton The basc concept of the fuzz matr theor s ver smple and can be appled to socal and natural stuatons A branch of fuzz matr

More information

Research Article Relative Smooth Topological Spaces

Research Article Relative Smooth Topological Spaces Advances n Fuzzy Systems Volume 2009, Artcle ID 172917, 5 pages do:10.1155/2009/172917 Research Artcle Relatve Smooth Topologcal Spaces B. Ghazanfar Department of Mathematcs, Faculty of Scence, Lorestan

More information

Notes on Frequency Estimation in Data Streams

Notes on Frequency Estimation in Data Streams Notes on Frequency Estmaton n Data Streams In (one of) the data streamng model(s), the data s a sequence of arrvals a 1, a 2,..., a m of the form a j = (, v) where s the dentty of the tem and belongs to

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

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

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

More information

A CLASS OF RECURSIVE SETS. Florentin Smarandache University of New Mexico 200 College Road Gallup, NM 87301, USA

A CLASS OF RECURSIVE SETS. Florentin Smarandache University of New Mexico 200 College Road Gallup, NM 87301, USA A CLASS OF RECURSIVE SETS Florentn Smarandache Unversty of New Mexco 200 College Road Gallup, NM 87301, USA E-mal: smarand@unmedu In ths artcle one bulds a class of recursve sets, one establshes propertes

More information

Uncertainty and auto-correlation in. Measurement

Uncertainty and auto-correlation in. Measurement Uncertanty and auto-correlaton n arxv:1707.03276v2 [physcs.data-an] 30 Dec 2017 Measurement Markus Schebl Federal Offce of Metrology and Surveyng (BEV), 1160 Venna, Austra E-mal: markus.schebl@bev.gv.at

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

A New Refinement of Jacobi Method for Solution of Linear System Equations AX=b

A New Refinement of Jacobi Method for Solution of Linear System Equations AX=b Int J Contemp Math Scences, Vol 3, 28, no 17, 819-827 A New Refnement of Jacob Method for Soluton of Lnear System Equatons AX=b F Naem Dafchah Department of Mathematcs, Faculty of Scences Unversty of Gulan,

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

arxiv: v1 [cs.gt] 14 Mar 2019

arxiv: v1 [cs.gt] 14 Mar 2019 Stable Roommates wth Narcssstc, Sngle-Peaked, and Sngle-Crossng Preferences Robert Bredereck 1, Jehua Chen 2, Ugo Paavo Fnnendahl 1, and Rolf Nedermeer 1 arxv:1903.05975v1 [cs.gt] 14 Mar 2019 1 TU Berln,

More information

Using T.O.M to Estimate Parameter of distributions that have not Single Exponential Family

Using T.O.M to Estimate Parameter of distributions that have not Single Exponential Family IOSR Journal of Mathematcs IOSR-JM) ISSN: 2278-5728. Volume 3, Issue 3 Sep-Oct. 202), PP 44-48 www.osrjournals.org Usng T.O.M to Estmate Parameter of dstrbutons that have not Sngle Exponental Famly Jubran

More information

Mixed-integer vertex covers on bipartite graphs

Mixed-integer vertex covers on bipartite graphs Mxed-nteger vertex covers on bpartte graphs Mchele Confort, Bert Gerards, Gacomo Zambell November, 2006 Abstract Let A be the edge-node ncdence matrx of a bpartte graph G = (U, V ; E), I be a subset the

More information

n ). This is tight for all admissible values of t, k and n. k t + + n t

n ). This is tight for all admissible values of t, k and n. k t + + n t MAXIMIZING THE NUMBER OF NONNEGATIVE SUBSETS NOGA ALON, HAROUT AYDINIAN, AND HAO HUANG Abstract. Gven a set of n real numbers, f the sum of elements of every subset of sze larger than k s negatve, what

More information

Genericity of Critical Types

Genericity of Critical Types Genercty of Crtcal Types Y-Chun Chen Alfredo D Tllo Eduardo Fangold Syang Xong September 2008 Abstract Ely and Pesk 2008 offers an nsghtful characterzaton of crtcal types: a type s crtcal f and only f

More information

THE WEIGHTED WEAK TYPE INEQUALITY FOR THE STRONG MAXIMAL FUNCTION

THE WEIGHTED WEAK TYPE INEQUALITY FOR THE STRONG MAXIMAL FUNCTION THE WEIGHTED WEAK TYPE INEQUALITY FO THE STONG MAXIMAL FUNCTION THEMIS MITSIS Abstract. We prove the natural Fefferman-Sten weak type nequalty for the strong maxmal functon n the plane, under the assumpton

More information

Appendix B. Criterion of Riemann-Stieltjes Integrability

Appendix B. Criterion of Riemann-Stieltjes Integrability Appendx B. Crteron of Remann-Steltes Integrablty Ths note s complementary to [R, Ch. 6] and [T, Sec. 3.5]. The man result of ths note s Theorem B.3, whch provdes the necessary and suffcent condtons for

More information