Worst-Case Robust Multiuser Transmit Beamforming Using Semidefinite Relaxation: Duality and Implications

Size: px
Start display at page:

Download "Worst-Case Robust Multiuser Transmit Beamforming Using Semidefinite Relaxation: Duality and Implications"

Transcription

1 Worst-Case Robust Multuser Transmt Beamforg Usng Semdefnte Relaxaton: Dualty and mplcatons Tsung-Hu Chang, Wng-Kn Ma, and Chong-Yung Ch nsttute of Commun. Eng. & Dept. of Elect. Eng. Natonal Tsng Hua Unversty, Hsnchu, Tawan E-mal: Department of Electronc Engneerng The Chnese Unversty of Hong Kong, Shatn, N.T., Hong Kong E-mal: Abstract Ths paper studes a downlnk multuser transmt beamforg desgn under sphercal channel uncertantes, usng a worst-case robust formulaton. Ths robust desgn problem s nonconvex. Recently, a convex approxmaton formulaton based on semdefnte relaxaton (SDR) has been proposed to handle the problem. Curously, smulaton results have consstently ndcated that SDR can attan the global optmum of the robust desgn problem. Ths paper ntends to provde some theoretcal nsghts nto ths mportant emprcal fndng. Our man result s a dual representaton of the SDR formulaton, whch reveals an nterestng lnkage to a dfferent robust desgn problem, and the possblty of SDR optmalty.. NTRODUCTON Ths paper focuses on a standard wreless multuser uncast system where a multple-antenna transmtter broadcasts ndependent data streams to multple sngle-antenna recevers usng transmt beamforg [1. n ths context, the effcacy of beamforg desgns reles on knowledge of the channel state nformaton (CS) of all the recevers. However, the transmtter often has some uncertantes on the CS, due to ssues such as fnte-length tranng and fnte-rate feedback [2. CS uncertantes at the transmtter can result n sgnfcant performance outage, f not taken nto consderaton n the beamforg desgns. The CS uncertanty problem has motvated consderable research endeavors n robust transmt beamforg desgn technques. Ths ncludes the chance constraned robust desgns [3, [4, where the CS uncertantes are modeled as random varables, and the worst-case robust desgns [5 [8, where the CS uncertantes are modeled as bounded unknowns wthn a predetered, small error set. Our problem of nterest s the worst-case sgnal-tonterference-plus-nose rato (SNR) constraned robust transmt beamforg desgn problem under sphercally bounded CS uncertantes, whch has drawn much nterest recently [6 [8. Presently avalable beamforg solutons for ths worstcase robust problem are based on approxmaton methods, ether restrcton [6, [7 or relaxaton [8, and t s now not clear whether the worst-case robust problem can be optmally (and effcently) solved. However, smulatons seem to have provded the answer to the latter the semdefnte relaxaton (SDR) method [8. SDR s a convex relaxaton technque for a certan class of hard (nonconvex) optmzaton problems, and has recently ganed popularty owng to ts wde scope of applcablty [9, [10. For a general applcaton, SDR s consdered a suboptmal solver; however, for the worstcase robust beamforg problem, smulaton results have ndcated that SDR should to be a globally optmal solver, whch s a rather surprsng emprcal fndng. As such, beng able to provde a theoretcal analyss provng whether SDR s optmal would be of much sgnfcance. A recent result [11 has partally addressed ths open queston, where the SDR optmalty under suffcently small error rad s analyzed. Ths paper ntends to address the mystery of SDR optmalty n worst-case robust transmt beamforg optmzaton usng a dfferent analyss approach. We show that the worst-case robust problem has a close relatonshp to a dfferent robust beamforg problem, n form of - optmzaton. n partcular, we prove that ther SDR problems are dual, or equvalent, to each other. Ths new, ntrgung, dualty relatonshp provdes a new perspectve and useful nsghts explanng the optmalty of SDR. n partcular, we wll gve a condton under whch SDR provdes globally optmal solutons to the worst-case robust problem.. SGNAL MODEL AND BACKGROUND Consder a wreless downlnk system where a transmtter, equpped wth N t antennas, wants to communcate wth K sngle-antenna recevers usng transmt beamforg. The problem formulaton follows a standard uncast settng [1: Let h C Nt denote the channel vector of recever, andlet w C Nt be the assocated beamforg vector for recever. The SNR of recever s gven by h H SNR (w 1,...,w K, h )= w 2 K k hh w, (1) k 2 + σ 2 where σ 2 > 0 s the nose power at recever, for all =1,...,K. Our goal s to desgn the beamforg vectors {w } K such that each recever acheves a desred SNR level. Conventonally, transmt beamforg desgns requre full channel state nformaton (CS) at the transmtter;.e., knowledge of {h } K. n wreless communcatons, however, t /11/$ EEE 1579 Aslomar 2011

2 s dffcult for the transmtter to acqure accurate CS, due to mperfect channel estmaton and fnte rate feedback [2. Hence there are channel uncertantes at the transmtter;.e., h = h + e, =1,...,K, (2) where h denotes the channel estmate avalable at the transmtter, and e C Nt represents the channel uncertanty. n ths work, we focus on sphercally bounded channel uncertantes: e 2 r 2, =1,...,K, (3) where denotes the Eucldean norm, and r > 0 s the radus of the uncertanty ball. We study the followng worstcase robust beamforg desgn [6, [8: w C N t,,...,k w 2 (4a) s.t. SNR (w 1,...,w K, h + e ) γ e 2 r 2, =1,...,K, (4b) where γ > 0 s the SNR requrement of recever, whch must be fulflled even under worst possble CS uncertantes. The challenge of solvng the worst-case robust problem (4) les n the worst-case SNR constrants n (4b), each of whch corresponds to an nfnte number of nonconvex quadratc constrants. As mentoned, there are several approxmaton methods for managng problem (4) [6 [8, and here we focus on the SDR method [8. The development of SDR conssts of two steps. The frst step, whch s standard (see, e.g. [10), s to substtute W = w w H, k =1,...,K, nto (4b), and then replace W = w w H by W 0 (.e., W beng postve semdefnte (PSD)) to obtan a relaxed problem W H N t,,...,k Tr(W ) s.t. ( h + e ) H 1 W γ k ( h + e ) σ 2 (5a) e 2 r 2, k =1,...,K, (5b) W 1,...,W K 0, (5c) where H Nt s the set of all N t by N t Hermtan matrces, and Tr(W ) denotes the trace of W. The motvaton of ths step s to lnearze the nonconvex constrants. The second step s to turn (5b) to fnte numbers of constrants, thereby enablng effcent mplementatons. By applyng S-lemma (see [12) to (5b), we obtan the followng SDR formulaton of (4): W H N t,λ R,...,K Tr(W ) (6a) s.t. Ψ (W 1,...,W K,λ ) 0, =1,...,K, (6b) W 1,...,W K 0, λ 1,...,λ K 0, where the matrx functons Ψ (W 1,...,W K,λ ) are defned as [ Ψ (W 1,...,W K,λ ) 1 h H W [ γ h k [ λ σ 2 λ r 2, =1,...,K, (7) where s the N t by N t dentty matrx. Note that the SDR problem (6) s a semdefnte program (SDP), whch s convex and tractable. The SDR problem (6) s methodologcally an approxmaton to the worst-case robust problem (4) because the ranks of W are not constraned. However, f the optmal soluton of the SDR problem (6), denoted by (W1,...,W K ), s of rank one;.e., W = w (w )H for all =1,...,K, then t can be verfed that (w1,...,w K ) s a globally optmal soluton to the worst-case robust formulaton (4). Rather surprsngly, t s found through smulatons that SDR yelds rank-one soluton automatcally, and t happens seegly all the tme [8, [11 (see also [4). Our endeavor n the subsequent secton s to provde a dual formulaton of the SDR problem (6) that may shed lght nto ths emprcal fndng. Before we proceed to the man result, let us present some smulaton results to further strengthen the motvaton of the rased analyss problem. Specfcally, we benchmark the SDR method aganst other concurrent approxmaton methods, namely, the robust SOCP-based method n [6, and the MMSEbased SDP method n [7. The smulaton settngs are: N t =4, K = 4, γ γ 1 = = γ K,σ1 2 = = σk 2 = 0.1, r r 1 = = r K =0.1, and( h 1,..., h K ) beng ndependent and dentcally dstrbuted complex Gaussan random varables wth zero mean and unt varance. The result s shown n Fg. 1, where we see that the SDR method outperforms the other two methods. Moreover, we should emphasze that the SDR method yelded rank-one soluton n all the trals ran.. DUALTY OF WORST-CASE ROBUST SDR Consder the followng - optmzaton problem w 2 e C N t,,...,k w C N t,,...,k s.t. SNR (w 1,...,w K, h + e ) γ, =1,...,K, s.t. e 2 r 2, =1,...,K. (8) At frst look, problem (8) s dfferent from the worst-case robust problem n (4). n (8), the nner mzaton s a standard non-robust beamforg desgn problem [1 whch fnds the most power effcent desgn gven a presumed CS { h + e } K. The outer mzaton, however, targets to fnd a worst set of CS uncertantes {e } K that mzes the nner-mum transmt power. We should also note that problem (8) has a flavor of two-player zero-sum game. We are partcularly nterested n applyng SDR to (8). Lke SDR for the worst-case robust problem, we replace each 1580

3 Average transmsson power (db) Robust SOCP va structured SDP [6 MMSE-based SDP [7 SDR n (5) γ (db) Fg. 1: Smulaton results of average transmsson power versus target SNR γ, for uncertanty radus r =0.1. w w H wth a PSD matrx W, and each [e H 1[e H 1 H wth a PSD matrx V, to obtan the followng problem Tr(W ) W H N t s.t. Tr 1 W R σ 2 γ, V H N t +1,,...,K k =1,...,K, W 1,...,W K 0. s.t. Tr(V ) (1 + r 2 ), =1,...,K, [V Nt+1 =1, =1,...,K, V 1,...,V K 0, (9) where [V Nt+1 s the (N t +1,N t +1)th entry of V and R = [ h [ H V h, =1,...,K. An mportant observaton of problem (9) s that there always exsts a rank-one soluton for the nner mzaton of problem (9): Fact 1 [1 Consder the followng SDP: W H N t,,...,k Tr(W ) (10) s.t. Tr 1 W γ k W 0, =1,...,K, R σ 2, =1,...,K, where R 1,...,R K 0. Suppose that (10) s feasble. Then there exsts an optmal soluton (W1,...,WK ) for whch rank(w )=1for all. Fact 1 mples that the SDR of (W 1,...,W K ) s always tght for the - SDR problem (9). Fact 1 rases an ntrgung queston What s the relatonshp between the - SDR problem (9) and the robust SDR problem (6)? f the optmal solutons of (W 1,...,W K ) of the two problems are dentcal, then Fact 1 mmedately mples that (6) has a rank-one optmal soluton and hence SDR s tght to (6) as well. A. Man Result t turns out that problems (9) and (6) are strongly connected: Proposton 1 Suppose that problem (6) s feasble. Then problems (9) and (6) attan the same optmal objectve value. Moreover, f (W1,...,WK,λ 1,...,λ K ) s an optmal soluton of problem (6), then there exsts (V1,...,V K ) such that (V1,...,VK, W 1,...,WK ) s an outer-nner soluton of problem (9). As the man contrbuton of ths paper, Proposton 1 provdes a soluton correspondence between problems (9) and (6), showng that problem (9) s actually a dual representaton of problem (6). To prove that problems (9) and (6) attan the same optmal objectve value, we show that the Lagrangan dual of problem (6) s equvalent to the Lagrangan dual of problem (9). The former can be shown to be A H N t +1,,...,K σ 2 [A Nt+1 (11) s.t. Y (A 1,...,A K ) 0, =1,...,K, Tr(A ) (1 + r 2 )[A Nt+1, =1,...,K, A 1,...,A K 0, where A 1,...,A K H Nt+1 are the (Lagrangan) dual varables assocated wth constrants (6b), and Y (A 1,...,A K ) 1 [ [ h A γ + h H [ [ hk Ak h H, =1,...,K. (12) k Now let us consder the Lagrangan dual of the nner mzaton problem of (9), whch can be shown to be μ σ 2 (13) μ 1,...,μ K 0 s.t. μ γ R + μ k R 0, =1,...,K, where μ 1,...,μ K are the dual varables assocated wth the trace nequalty constrants of the nner problem of (9). Replacng the nner problem of (9) wth ts dual (13), we obtan the followng problem μ σ 2 μ 0,,...,K (14) V H N t +1,,...,K s.t. Y (μ 1 V 1,...,μ K V K ) 0, =1,...,K, s.t. Tr(V ) (1 + r 2 ), =1,...,K, [V Nt+1 =1, =1,...,K, V 1,...,V K 0. Snce strong dualty holds for the nner parts of (9) and (14), the two problems have the same optmal objectve value. 1581

4 One may observe a connecton between (11) and (14): [A Nt+1 = μ, A = μ V, =1,...,K. (15) n fact, (11) and (14) are equvalent problems, as we show n Appendx the followng lemma: Lemma 1 f (A 1,...,A K ) s an optmal soluton of (11), then (V1,...,V K )=(A 1 /[A 1 N t+1,...,a K /[A K N t+1), (μ 1,...,μ K)=([A 1 Nt+1,...,[A K Nt+1) (16) s an optmal outer-nner soluton par of (14). f (V1,...,V K,μ 1,...,μ K ) s an optmal outer-nner soluton of (14), then(a 1,...,A K )=(μ 1 V 1,...,μ K V K ) s optmal to (11). Lemma 1 shows that (V1,...,V K ) of (14) only dffers from (A 1,...,A K ) of (11) up to a postve scalar. Hence, (14) and (11) attan the same optmal objectve value, mplyng that (9) and (6) attan the same optmal objectve value. By Lemma 1, one can further show that (W1,...,W K ), the optmal prmal soluton of (6), s also optmal to (9). The detaled proof s presented n Appendx. B. mplcaton and Concludng Remark To show that the robust SDR problem (6) has a rank-one soluton, we stll need to prove that the optmal (W 1,...,W K ) of (9) s also optmal to (6). Now, let us assume: Condton 1 The optmal soluton of the nner mzaton of problem (9), (W1,...,W K ), s unque. Condton 1 s consdered mld; by numercal experence, Condton 1 s found to hold all the tme. Under Condton 1, we can nfer from Fact 1 and Proposton 1 that the SDR problem (6) has a rank-one soluton. Hence, we conclude that Clam 1 Under Condton 1, the SDR problem (6) solves the worst-cast robust problem (4) optmally. Our analyss above narrows down the SDR optmalty queston to the proof of unque rank-one soluton of the nner mzaton problem of (9). As a future research drecton, t would be nterestng to nvestgate suffcent condtons under whch Condton 1 holds true. V. APPENDX KKT condtons of (6) The KKT condtons of (6) and (11) can be shown to be W 1,...,W K 0,λ 1,...,λ K 0, A 1,...,A K 0, (17a) Ψ (W 1,...,W K,λ ) 0, =1,...,K, (17b) Y (A 1,...,A K ) 0, =1,...,K, (17c) Ψ (W 1,...,W K,λ )A = 0, =1,...,K, (17d) Y (A 1,...,A K )W = 0, =1,...,K, (17e) Tr(A ) (1 + r 2 )[A Nt+1, =1,...,K, (17f) ( Tr(A ) (1 + r 2 )[A ) Nt+1 λ =0, =1,...,K, (17g) where Ψ ( ) and Y ( ) are defned n (7) and (12), respectvely. Proof of Lemma 1: Lemma 1 can be easly proved by nspecton of (14) and (11). What remans s to show that μ > 0 and [A N t+1 > 0 for all = 1,...,K. The former has been proved n [13, Proposton 4.2; whle the latter can be proved as follows. One can observe from (17a) and (17f) that [A N t+1 = 0 results n A = 0. n ths case, Y (A 1,...,A K ) n (12) s postve defnte,.e., Y (A 1,...,A K ) 0. By the complementary slackness (17e), ths leads to the prmal soluton W = 0, whch however volates (17b) [see (7) due to σ 2 > 0. Proof of Proposton 1: Here we prove that (W1,...,WK ), the optmal prmal soluton of (6), s also optmal to (9). By Lemma 1 whch shows that (A 1/[A 1 Nt+1,...,A K /[A K N t+1) s an optmal outer mzer of (9), t suffces to show that (W1,...,W K ) s optmal to the followng problem W 1,...,W K 0 Tr(W ) (18) s.t. Tr 1 W γ σ 2 [A Nt+1, =1,...,K. [ h A [ h H Ths can be shown by exang that (W 1,...,W K ) satsfes the KKT condtons of (18), whch are gven as follows: W 1,...,W K 0, μ 1,...,μ K 0, (19a) Y (μ 1 (A 1 /[A 1 N t+1),...,μ K (A K /[A K N t+1)) 0, (19b) Y (μ 1 (A 1/[A 1 Nt+1),...,μ K (A K/[A K Nt+1)) W = 0, (19c) Tr 1 W [ [ γ h A = h H σ 2 [A N t+1, (19d) for =1,...,K. Snce (W1,...,WK,λ 1,...,λ K ) and (A 1,...,A K ) satsfy the KKT condtons n (17a), (17c) and (17e), (W1,...,WK ) and (μ 1,...,μ K ) ([A 1 N t+1,...,[a K N t+1) satsfy (19a), (19b) and (19c). To show that (W1,...,W K ) also fulflls (19d), let us consder an alternatve representaton of (6): Lemma 2 Problem (6) can be equvalently expressed as the followng problem W 0,,...,K s.t. Tr(W ) Tr 1 W V V γ σ 2, =1,...,K, (20a) [ [ h V h H (20b) 1582

5 where V = {V H Nt+1 Tr(V ) (1 + r 2 ), [V Nt+1 = 1, V 0}. t s easy to verty that, for (W1,...,WK ), Tr 1 K W γ Wk V V [ h V [ h H = σ 2, =1,...,K, (21).e., the nequalty constrants n (20b) are all actve for the optmal soluton (W1,...,WK ). Hence, to show that (19d) s also fulflled by (W1,...,W K ), t s suffcent to prove that A /[A N t+1 = arg Tr 1 K W V 0 γ W [ [ k h V h H s.t. Tr(V ) (1 + r 2 ), (22a) [V Nt+1 =1, (22b) for all = 1,...,K. Let ξ and τ be the dual varables assocated wth the constrants n (22a) and (22b), respectvely, and defne Ψ (W1,...,WK,ξ,τ ) [ h H 1 W γ W k [ [ h ξ ξ + τ The KKT condtons of the mzaton problem n (22) can be obtaned as Tr(V ) (1 + r 2 ), [V Nt+1 =1,V 0,ξ 0,τ R, Ψ (W1,...,W K,ξ,τ ) 0, Ψ (W1,...,WK,ξ,τ ) V = 0, ( ξ Tr(V ) (1 + r 2 )) =0, τ ([V Nt+1 1) = 0. (23a) (23b) (23c) (23d) For each {1,...,K}, letv = A /[A N t+1, ξ = λ, τ = σ 2 (1 + r 2)λ. t follows from the KKT condtons n (17a), (17b), (17d) and (17g) that (V,ξ,τ ) satsfes all the condtons n (23). Thus (22) s true for all =1,...,K. The proof s then completed. Proof of Lemma 2: t suffces to show that (6b) s equvalent to (20b). Note that (6b) s equvalent to e 2 r 2 ( h + e ) H 1 W ( h + e ) γ σ 2, =1,...,K. (24) (the equvalence s owng to the S-Lemma; see [8, [11). Note that the mzaton problem on the left-hand sde of (24) may not be convex wth respect to (e 1,...,e K ) because the matrx ( 1 γ W K k ) may not be postve semdefnte. Nevertheless, SDR can be appled. Through the same procedure as n obtanng (9), one can obtan the SDR problem of the mzaton problem n (24) as Tr 1 W [ [ γ h V h H. (25) V V Whle (25) s obtaned by relaxaton of the rank of V,theSDR problem (25) s actually tght and optmal to the mzaton problem n (24); see [14, Lemma 3.1. We thus obtan (20b) by substtutng (25) nto (24). V. ACKNOWLEDGEMENTS Ths work s supported n part by Natonal Scence Councl, R.O.C., under Grant NSC E MY3, by a General Research Fund of Hong Kong Research Grant Councl (CUHK ), and by a Drect Grant awarded by the Chnese Unversty of Hong Kong (Project Code ). REFERENCES [1 M. Bengtsson and B. Ottersten, Optmal and suboptmal transmt beamforg, Chapter 18 n Handbook of Antennas n Wreless Communcatons, L. C. Godara, Ed., CRC Press, Aug [2 D. J. Love, R. W. Heath, V. K. N. Lau, D. Gesbert, B. D. Rao, and M. Andrews, An overvew of lmted feedback n wreless communcaton systems, EEE Journal on Sel. Areas n Comm., vol. 26, pp , Oct [3 M. B. Shenouda and T. N. Davdson, Probablstcally-constraned approaches to the desgn of the multple antenna downlnk, n Proc. EEE Aslomar Conf. Sgnals, Systems and Computers, Pacfc Grove, Oct , 2008, pp [4 K.-Y. Wang, A. M.-C. So, T.-H. Chang, W.-K. Ma, and C.-Y. Ch, Outage constraned robust transmt optmzaton for multuser MSO downlnks: Tractable approxmatons by conc optmzaton, submtted to EEE Trans. Sgnal Process., 2011, abrdged versons publshed n EUSPCO 2010 and CASSP [5 M. B. Shenouda and T. N. Davdson, Nonlnear and lnear broadcastng wth QoS requrements: Tractable approaches for bounded channel uncertantes, EEE Trans. Sgnal Process., vol. 57, no. 5, pp , May [6, Convex conc formulatons of robust downlnk precoder desgns wth qualty of servce constrants, EEE J. Sel. Topcs n Sgnal Process., vol. 1, pp , Dec [7 N. Vu cć and H. Boche, Robust QoS-constraned optmzaton of downlnk multuser MSO systems, EEE Trans. Sgnal Process., vol. 57, pp , Feb [8 G. Zheng, K.-K. Wong, and T.-S. Ng, Robust lnear MMO n the downlnk: A worst-case optmzaton wth ellpsodal uncertanty regons, EURASP Journal on Advances n Sgnal Process., vol. 2008, pp. 1 15, June 2008, Artcle D [9 Z.-Q. Luo and T.-H. Chang, SDP relaxaton of homogeneous quadratc optmzaton: Approxmaton bounds and applcatons, Chapter 4 n Convex Optmzaton n Sgnal Processng and Communcatons, D.P. Palomar and Y. Eldar, Eds., UK: Cambrdge Unversty, [10 Z.-Q. Luo, W.-K. Ma, A. M.-C. So, Y. Ye, and S. Zhang, Semdefnte relaxaton of quadratc optmzaton problems, EEE Sgnal Process. Mag., pp , May [11 E. Song, Q. Sh, M. Sanjab, R. Sun, and Z.-Q. Luo, Robust SNRconstraned MSO downlnk beamforg: When s semdefnte programg relaxaton tght? n Proc. EEE CASSP, Prague, Czech, May 22-27, 2011, pp [12 S. Boyd and L. Vandenberghe, Convex Optmzaton. Cambrdge, UK: Cambrdge Unversty Press, [13 Y. Huang and D. P. Palomar, A dual perspectve on separable semdefnte programg wth applcatons to optmal downlnk beamforg, EEE Trans. Sgnal Process., vol. 58, no. 8, pp , August [14, Rank-constraned separable semdefnte program wth applcatons to optmal beamforg, EEE Trans. Sgnal Process., vol. 58, no. 2, pp , Feb

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

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

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

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

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

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

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

Convex Optimization. Optimality conditions. (EE227BT: UC Berkeley) Lecture 9 (Optimality; Conic duality) 9/25/14. Laurent El Ghaoui.

Convex Optimization. Optimality conditions. (EE227BT: UC Berkeley) Lecture 9 (Optimality; Conic duality) 9/25/14. Laurent El Ghaoui. Convex Optmzaton (EE227BT: UC Berkeley) Lecture 9 (Optmalty; Conc dualty) 9/25/14 Laurent El Ghaou Organsatonal Mdterm: 10/7/14 (1.5 hours, n class, double-sded cheat sheet allowed) Project: Intal proposal

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

ECE559VV Project Report

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

More information

Lagrange Multipliers Kernel Trick

Lagrange Multipliers Kernel Trick Lagrange Multplers Kernel Trck Ncholas Ruozz Unversty of Texas at Dallas Based roughly on the sldes of Davd Sontag General Optmzaton A mathematcal detour, we ll come back to SVMs soon! subject to: f x

More information

Lecture 10 Support Vector Machines II

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

More information

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

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

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

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

Power Allocation/Beamforming for DF MIMO Two-Way Relaying: Relay and Network Optimization

Power Allocation/Beamforming for DF MIMO Two-Way Relaying: Relay and Network Optimization Power Allocaton/Beamformng for DF MIMO Two-Way Relayng: Relay and Network Optmzaton Je Gao, Janshu Zhang, Sergy A. Vorobyov, Ha Jang, and Martn Haardt Dept. of Electrcal & Computer Engneerng, Unversty

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

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

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

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

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

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

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

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

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

6.854J / J Advanced Algorithms Fall 2008

6.854J / J Advanced Algorithms Fall 2008 MIT OpenCourseWare http://ocw.mt.edu 6.854J / 18.415J Advanced Algorthms Fall 2008 For nformaton about ctng these materals or our Terms of Use, vst: http://ocw.mt.edu/terms. 18.415/6.854 Advanced Algorthms

More information

Outage Constrained Robust Transmit Optimization for Multiuser MISO Downlinks: Tractable Approximations by Conic Optimization

Outage Constrained Robust Transmit Optimization for Multiuser MISO Downlinks: Tractable Approximations by Conic Optimization Outage Constraned Robust Transmt Optmzaton for Multuser MISO Downlnks: Tractable Approxmatons by Conc Optmzaton Kun-Yu Wang, Anthony Man-Cho So, Tsung-Hu Chang, Wng-Kn Ma, and Chong-Yung Ch Abstract In

More information

University of Alberta. Library Release Form. Title of Thesis: Joint Bandwidth and Power Allocation in Wireless Communication Networks

University of Alberta. Library Release Form. Title of Thesis: Joint Bandwidth and Power Allocation in Wireless Communication Networks Unversty of Alberta Lbrary Release Form Name of Author: Xaowen Gong Ttle of Thess: Jont Bandwdth and Power Allocaton n Wreless Communcaton Networks Degree: Master of Scence Year ths Degree Granted: 2010

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

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

Communication Complexity 16:198: February Lecture 4. x ij y ij

Communication Complexity 16:198: February Lecture 4. x ij y ij Communcaton Complexty 16:198:671 09 February 2010 Lecture 4 Lecturer: Troy Lee Scrbe: Rajat Mttal 1 Homework problem : Trbes We wll solve the thrd queston n the homework. The goal s to show that the nondetermnstc

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

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

Econ107 Applied Econometrics Topic 3: Classical Model (Studenmund, Chapter 4)

Econ107 Applied Econometrics Topic 3: Classical Model (Studenmund, Chapter 4) I. Classcal Assumptons Econ7 Appled Econometrcs Topc 3: Classcal Model (Studenmund, Chapter 4) We have defned OLS and studed some algebrac propertes of OLS. In ths topc we wll study statstcal propertes

More information

General viscosity iterative method for a sequence of quasi-nonexpansive mappings

General viscosity iterative method for a sequence of quasi-nonexpansive mappings Avalable onlne at www.tjnsa.com J. Nonlnear Sc. Appl. 9 (2016), 5672 5682 Research Artcle General vscosty teratve method for a sequence of quas-nonexpansve mappngs Cuje Zhang, Ynan Wang College of Scence,

More information

FUll-duplex (FD) is a promising technique to increase

FUll-duplex (FD) is a promising technique to increase 1 Jont Relay-User Beamformng Desgn n Full-Duplex Two-Way Relay Channel Zhgang Wen, Shua Wang, Xaoqng Lu, and Junwe Zou arxv:1610.04974v1 cs.it] 17 Oct 2016 Abstract A full-duplex two-way relay channel

More information

Matrix Approximation via Sampling, Subspace Embedding. 1 Solving Linear Systems Using SVD

Matrix Approximation via Sampling, Subspace Embedding. 1 Solving Linear Systems Using SVD Matrx Approxmaton va Samplng, Subspace Embeddng Lecturer: Anup Rao Scrbe: Rashth Sharma, Peng Zhang 0/01/016 1 Solvng Lnear Systems Usng SVD Two applcatons of SVD have been covered so far. Today we loo

More information

Two-Way and Multiple-Access Energy Harvesting Systems with Energy Cooperation

Two-Way and Multiple-Access Energy Harvesting Systems with Energy Cooperation Two-Way and Multple-Access Energy Harvestng Systems wth Energy Cooperaton Berk Gurakan, Omur Ozel, Jng Yang 2, and Sennur Ulukus Department of Electrcal and Computer Engneerng, Unversty of Maryland, College

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

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

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

More information

Linear Approximation with Regularization and Moving Least Squares

Linear Approximation with Regularization and Moving Least Squares Lnear Approxmaton wth Regularzaton and Movng Least Squares Igor Grešovn May 007 Revson 4.6 (Revson : March 004). 5 4 3 0.5 3 3.5 4 Contents: Lnear Fttng...4. Weghted Least Squares n Functon Approxmaton...

More information

A SEMIDEFINITE RELAXATION BASED CONSERVATIVE APPROACH TO ROBUST TRANSMIT BEAMFORMING WITH PROBABILISTIC SINR CONSTRAINTS

A SEMIDEFINITE RELAXATION BASED CONSERVATIVE APPROACH TO ROBUST TRANSMIT BEAMFORMING WITH PROBABILISTIC SINR CONSTRAINTS th European Signal Processing Conference EUSIPCO-1 Aalborg, Denmark, August 3-7, 1 A SEMIDEFINITE RELAXATION BASED CONSERVATIVE APPROACH TO ROBUST TRANSMIT BEAMFORMING WITH PROBABILISTIC SINR CONSTRAINTS

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

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

Composite Hypotheses testing

Composite Hypotheses testing Composte ypotheses testng In many hypothess testng problems there are many possble dstrbutons that can occur under each of the hypotheses. The output of the source s a set of parameters (ponts n a parameter

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

Explicit and Implicit Temperature Constraints in Energy Harvesting Communications

Explicit and Implicit Temperature Constraints in Energy Harvesting Communications Explct and Implct Temperature Constrants n Energy Harvestng Communcatons Abdulrahman Baknna, Omur Ozel 2, and Sennur Ulukus Department of Electrcal and Computer Engneerng, Unversty of Maryland, College

More information

x = x 1 + :::+ x K and the nput covarance matrces are of the form ± = E[x x y ]. 3.2 Dualty Next, we ntroduce the concept of dualty wth the followng t

x = x 1 + :::+ x K and the nput covarance matrces are of the form ± = E[x x y ]. 3.2 Dualty Next, we ntroduce the concept of dualty wth the followng t Sum Power Iteratve Water-fllng for Mult-Antenna Gaussan Broadcast Channels N. Jndal, S. Jafar, S. Vshwanath and A. Goldsmth Dept. of Electrcal Engg. Stanford Unversty, CA, 94305 emal: njndal,syed,srram,andrea@wsl.stanford.edu

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

EEE 241: Linear Systems

EEE 241: Linear Systems EEE : Lnear Systems Summary #: Backpropagaton BACKPROPAGATION The perceptron rule as well as the Wdrow Hoff learnng were desgned to tran sngle layer networks. They suffer from the same dsadvantage: they

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

ISSN: ISO 9001:2008 Certified International Journal of Engineering and Innovative Technology (IJEIT) Volume 3, Issue 1, July 2013

ISSN: ISO 9001:2008 Certified International Journal of Engineering and Innovative Technology (IJEIT) Volume 3, Issue 1, July 2013 ISSN: 2277-375 Constructon of Trend Free Run Orders for Orthogonal rrays Usng Codes bstract: Sometmes when the expermental runs are carred out n a tme order sequence, the response can depend on the run

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

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

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

Chapter 8 Indicator Variables

Chapter 8 Indicator Variables Chapter 8 Indcator Varables In general, e explanatory varables n any regresson analyss are assumed to be quanttatve n nature. For example, e varables lke temperature, dstance, age etc. are quanttatve n

More information

Control of Uncertain Bilinear Systems using Linear Controllers: Stability Region Estimation and Controller Design

Control of Uncertain Bilinear Systems using Linear Controllers: Stability Region Estimation and Controller Design Control of Uncertan Blnear Systems usng Lnear Controllers: Stablty Regon Estmaton Controller Desgn Shoudong Huang Department of Engneerng Australan Natonal Unversty Canberra, ACT 2, Australa shoudong.huang@anu.edu.au

More information

Estimation: Part 2. Chapter GREG estimation

Estimation: Part 2. Chapter GREG estimation Chapter 9 Estmaton: Part 2 9. GREG estmaton In Chapter 8, we have seen that the regresson estmator s an effcent estmator when there s a lnear relatonshp between y and x. In ths chapter, we generalzed the

More information

Appendix B. The Finite Difference Scheme

Appendix B. The Finite Difference Scheme 140 APPENDIXES Appendx B. The Fnte Dfference Scheme In ths appendx we present numercal technques whch are used to approxmate solutons of system 3.1 3.3. A comprehensve treatment of theoretcal and mplementaton

More information

Systems of Equations (SUR, GMM, and 3SLS)

Systems of Equations (SUR, GMM, and 3SLS) Lecture otes on Advanced Econometrcs Takash Yamano Fall Semester 4 Lecture 4: Sstems of Equatons (SUR, MM, and 3SLS) Seemngl Unrelated Regresson (SUR) Model Consder a set of lnear equatons: $ + ɛ $ + ɛ

More information

The Concept of Beamforming

The Concept of Beamforming ELG513 Smart Antennas S.Loyka he Concept of Beamformng Generc representaton of the array output sgnal, 1 where w y N 1 * = 1 = w x = w x (4.1) complex weghts, control the array pattern; y and x - narrowband

More information

On the Global Linear Convergence of the ADMM with Multi-Block Variables

On the Global Linear Convergence of the ADMM with Multi-Block Variables On the Global Lnear Convergence of the ADMM wth Mult-Block Varables Tany Ln Shqan Ma Shuzhong Zhang May 31, 01 Abstract The alternatng drecton method of multplers ADMM has been wdely used for solvng structured

More information

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

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

More information

Natural Language Processing and Information Retrieval

Natural Language Processing and Information Retrieval Natural Language Processng and Informaton Retreval Support Vector Machnes Alessandro Moschtt Department of nformaton and communcaton technology Unversty of Trento Emal: moschtt@ds.untn.t Summary Support

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

Eigenvalues of Random Graphs

Eigenvalues of Random Graphs Spectral Graph Theory Lecture 2 Egenvalues of Random Graphs Danel A. Spelman November 4, 202 2. Introducton In ths lecture, we consder a random graph on n vertces n whch each edge s chosen to be n the

More information

DUE: WEDS FEB 21ST 2018

DUE: WEDS FEB 21ST 2018 HOMEWORK # 1: FINITE DIFFERENCES IN ONE DIMENSION DUE: WEDS FEB 21ST 2018 1. Theory Beam bendng s a classcal engneerng analyss. The tradtonal soluton technque makes smplfyng assumptons such as a constant

More information

SELECTED SOLUTIONS, SECTION (Weak duality) Prove that the primal and dual values p and d defined by equations (4.3.2) and (4.3.3) satisfy p d.

SELECTED SOLUTIONS, SECTION (Weak duality) Prove that the primal and dual values p and d defined by equations (4.3.2) and (4.3.3) satisfy p d. SELECTED SOLUTIONS, SECTION 4.3 1. Weak dualty Prove that the prmal and dual values p and d defned by equatons 4.3. and 4.3.3 satsfy p d. We consder an optmzaton problem of the form The Lagrangan for ths

More information

Limited Dependent Variables

Limited Dependent Variables Lmted Dependent Varables. What f the left-hand sde varable s not a contnuous thng spread from mnus nfnty to plus nfnty? That s, gven a model = f (, β, ε, where a. s bounded below at zero, such as wages

More information

I + HH H N 0 M T H = UΣV H = [U 1 U 2 ] 0 0 E S. X if X 0 0 if X < 0 (X) + = = M T 1 + N 0. r p + 1

I + HH H N 0 M T H = UΣV H = [U 1 U 2 ] 0 0 E S. X if X 0 0 if X < 0 (X) + = = M T 1 + N 0. r p + 1 Homework 4 Problem Capacty wth CSI only at Recever: C = log det I + E )) s HH H N M T R SS = I) SVD of the Channel Matrx: H = UΣV H = [U 1 U ] [ Σr ] [V 1 V ] H Capacty wth CSI at both transmtter and

More information

4 Analysis of Variance (ANOVA) 5 ANOVA. 5.1 Introduction. 5.2 Fixed Effects ANOVA

4 Analysis of Variance (ANOVA) 5 ANOVA. 5.1 Introduction. 5.2 Fixed Effects ANOVA 4 Analyss of Varance (ANOVA) 5 ANOVA 51 Introducton ANOVA ANOVA s a way to estmate and test the means of multple populatons We wll start wth one-way ANOVA If the populatons ncluded n the study are selected

More information

On Finite Rank Perturbation of Diagonalizable Operators

On Finite Rank Perturbation of Diagonalizable Operators Functonal Analyss, Approxmaton and Computaton 6 (1) (2014), 49 53 Publshed by Faculty of Scences and Mathematcs, Unversty of Nš, Serba Avalable at: http://wwwpmfnacrs/faac On Fnte Rank Perturbaton of Dagonalzable

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

SIO 224. m(r) =(ρ(r),k s (r),µ(r))

SIO 224. m(r) =(ρ(r),k s (r),µ(r)) SIO 224 1. A bref look at resoluton analyss Here s some background for the Masters and Gubbns resoluton paper. Global Earth models are usually found teratvely by assumng a startng model and fndng small

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

On a direct solver for linear least squares problems

On a direct solver for linear least squares problems ISSN 2066-6594 Ann. Acad. Rom. Sc. Ser. Math. Appl. Vol. 8, No. 2/2016 On a drect solver for lnear least squares problems Constantn Popa Abstract The Null Space (NS) algorthm s a drect solver for lnear

More information

Chapter 13: Multiple Regression

Chapter 13: Multiple Regression Chapter 13: Multple Regresson 13.1 Developng the multple-regresson Model The general model can be descrbed as: It smplfes for two ndependent varables: The sample ft parameter b 0, b 1, and b are used to

More information

Lecture 14: Bandits with Budget Constraints

Lecture 14: Bandits with Budget Constraints IEOR 8100-001: Learnng and Optmzaton for Sequental Decson Makng 03/07/16 Lecture 14: andts wth udget Constrants Instructor: Shpra Agrawal Scrbed by: Zhpeng Lu 1 Problem defnton In the regular Mult-armed

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

PROBLEM SET 7 GENERAL EQUILIBRIUM

PROBLEM SET 7 GENERAL EQUILIBRIUM PROBLEM SET 7 GENERAL EQUILIBRIUM Queston a Defnton: An Arrow-Debreu Compettve Equlbrum s a vector of prces {p t } and allocatons {c t, c 2 t } whch satsfes ( Gven {p t }, c t maxmzes βt ln c t subject

More information

UNIVERSITY OF TORONTO Faculty of Arts and Science. December 2005 Examinations STA437H1F/STA1005HF. Duration - 3 hours

UNIVERSITY OF TORONTO Faculty of Arts and Science. December 2005 Examinations STA437H1F/STA1005HF. Duration - 3 hours UNIVERSITY OF TORONTO Faculty of Arts and Scence December 005 Examnatons STA47HF/STA005HF Duraton - hours AIDS ALLOWED: (to be suppled by the student) Non-programmable calculator One handwrtten 8.5'' x

More information

x = , so that calculated

x = , so that calculated Stat 4, secton Sngle Factor ANOVA notes by Tm Plachowsk n chapter 8 we conducted hypothess tests n whch we compared a sngle sample s mean or proporton to some hypotheszed value Chapter 9 expanded ths to

More information

NON-CENTRAL 7-POINT FORMULA IN THE METHOD OF LINES FOR PARABOLIC AND BURGERS' EQUATIONS

NON-CENTRAL 7-POINT FORMULA IN THE METHOD OF LINES FOR PARABOLIC AND BURGERS' EQUATIONS IJRRAS 8 (3 September 011 www.arpapress.com/volumes/vol8issue3/ijrras_8_3_08.pdf NON-CENTRAL 7-POINT FORMULA IN THE METHOD OF LINES FOR PARABOLIC AND BURGERS' EQUATIONS H.O. Bakodah Dept. of Mathematc

More information

Psychology 282 Lecture #24 Outline Regression Diagnostics: Outliers

Psychology 282 Lecture #24 Outline Regression Diagnostics: Outliers Psychology 282 Lecture #24 Outlne Regresson Dagnostcs: Outlers In an earler lecture we studed the statstcal assumptons underlyng the regresson model, ncludng the followng ponts: Formal statement of assumptons.

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

Power Allocation for Distributed BLUE Estimation with Full and Limited Feedback of CSI

Power Allocation for Distributed BLUE Estimation with Full and Limited Feedback of CSI Power Allocaton for Dstrbuted BLUE Estmaton wth Full and Lmted Feedback of CSI Mohammad Fanae, Matthew C. Valent, and Natala A. Schmd Lane Department of Computer Scence and Electrcal Engneerng West Vrgna

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

Linear precoding via conic optimization for fixed MIMO receivers

Linear precoding via conic optimization for fixed MIMO receivers Lnear precodng va conc optmzaton for fxed MIMO recevers Am Wesel, Yonna C Eldar, and Shlomo Shama (Shtz) Department of Electrcal Engneerng Technon - Israel Insttute of Technology June 8, 004 Abstract We

More information

Chapter 11: Simple Linear Regression and Correlation

Chapter 11: Simple Linear Regression and Correlation Chapter 11: Smple Lnear Regresson and Correlaton 11-1 Emprcal Models 11-2 Smple Lnear Regresson 11-3 Propertes of the Least Squares Estmators 11-4 Hypothess Test n Smple Lnear Regresson 11-4.1 Use of t-tests

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

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

A Comparative Study for Estimation Parameters in Panel Data Model

A Comparative Study for Estimation Parameters in Panel Data Model A Comparatve Study for Estmaton Parameters n Panel Data Model Ahmed H. Youssef and Mohamed R. Abonazel hs paper examnes the panel data models when the regresson coeffcents are fxed random and mxed and

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

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

x i1 =1 for all i (the constant ).

x i1 =1 for all i (the constant ). Chapter 5 The Multple Regresson Model Consder an economc model where the dependent varable s a functon of K explanatory varables. The economc model has the form: y = f ( x,x,..., ) xk Approxmate ths by

More information

A generalization of a trace inequality for positive definite matrices

A generalization of a trace inequality for positive definite matrices A generalzaton of a trace nequalty for postve defnte matrces Elena Veronca Belmega, Marc Jungers, Samson Lasaulce To cte ths verson: Elena Veronca Belmega, Marc Jungers, Samson Lasaulce. A generalzaton

More information

Basic Statistical Analysis and Yield Calculations

Basic Statistical Analysis and Yield Calculations October 17, 007 Basc Statstcal Analyss and Yeld Calculatons Dr. José Ernesto Rayas Sánchez 1 Outlne Sources of desgn-performance uncertanty Desgn and development processes Desgn for manufacturablty A general

More information

Welfare Properties of General Equilibrium. What can be said about optimality properties of resource allocation implied by general equilibrium?

Welfare Properties of General Equilibrium. What can be said about optimality properties of resource allocation implied by general equilibrium? APPLIED WELFARE ECONOMICS AND POLICY ANALYSIS Welfare Propertes of General Equlbrum What can be sad about optmalty propertes of resource allocaton mpled by general equlbrum? Any crteron used to compare

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