Semidefinite Facial Reduction for Euclidean Distance Matrix Completion
|
|
- Ruby Armstrong
- 5 years ago
- Views:
Transcription
1 Semidefinite Facial Reduction for Euclidean Distance Matrix Completion Nathan Krislock, Henry Wolkowicz Department of Combinatorics & Optimization University of Waterloo First Alpen-Adria Workshop on Optimization University of Klagenfurt, Austria June 3-6, 2010 Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
2 Sensor Network Localization n = 100, m = 9, R = 2 Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
3 Molecular Conformation Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
4 Previous Work I Abdo Y. Alfakih, Amir Khandani, and Henry Wolkowicz. Solving Euclidean distance matrix completion problems via semidefinite programming. Comput. Optim. Appl., 12(1-3):13 30, L. Doherty, K. S. J. Pister, and L. El Ghaoui. Convex position estimation in wireless sensor networks. In INFOCOM Twentieth Annual Joint Conference of the IEEE Computer and Communications Societies. Proceedings. IEEE, volume 3, pages vol.3, Pratik Biswas and Yinyu Ye. Semidefinite programming for ad hoc wireless sensor network localization. In IPSN 04: Proceedings of the 3rd international symposium on Information processing in sensor networks, pages 46 54, New York, NY, USA, ACM. Pratik Biswas, Tzu-Chen Liang, Kim-Chuan Toh, Ta-Chung Wang, and Yinyu Ye. Semidefinite programming approaches for sensor network localization with noisy distance measurements. IEEE Transactions on Automation Science and Engineering, 3: , Pratik Biswas, Kim-Chuan Toh, and Yinyu Ye. A distributed SDP approach for large-scale noisy anchor-free graph realization with applications to molecular conformation. SIAM Journal on Scientific Computing, 30(3): , Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
5 Previous Work II Zizhuo Wang, Song Zheng, Yinyu Ye, and Stephen Boyd. Further relaxations of the semidefinite programming approach to sensor network localization. SIAM Journal on Optimization, 19(2): , Sunyoung Kim, Masakazu Kojima, and Hayato Waki. Exploiting sparsity in SDP relaxation for sensor network localization. SIAM Journal on Optimization, 20(1): , Ting Pong and Paul Tseng. (Robust) Edge-based semidefinite programming relaxation of sensor network localization. Mathematical Programming, Published online. Nathan Krislock and Henry Wolkowicz. Explicit sensor network localization using semidefinite representations and facial reductions. To appear in SIAM Journal on Optimization, Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
6 Summary We developed a theory of semidefinite facial reduction for the EDM completion problem Using this theory, we can transform the SDP relaxations into much smaller equivalent problems We developed a highly efficient algorithm for EDM completion: SDP solver not required can solve problems with up to 100,000 sensors in a few minutes on a laptop computer obtained very high accuracy for noiseless problems our running times are highly competitive with other SDP-based codes Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
7 Outline 1 Euclidean Distance Matrices Introduction Euclidean Distance Matrices and Semidefinite Matrices Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
8 Outline 1 Euclidean Distance Matrices Introduction Euclidean Distance Matrices and Semidefinite Matrices 2 Facial Reduction Semidefinite Facial Reduction Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
9 Outline 1 Euclidean Distance Matrices Introduction Euclidean Distance Matrices and Semidefinite Matrices 2 Facial Reduction Semidefinite Facial Reduction 3 Results Facial Reduction Theory for EDM Completion Numerical Results Noisy Problems Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
10 Outline 1 Euclidean Distance Matrices Introduction Euclidean Distance Matrices and Semidefinite Matrices 2 Facial Reduction Semidefinite Facial Reduction 3 Results Facial Reduction Theory for EDM Completion Numerical Results Noisy Problems Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
11 Introduction Euclidean Distance Matrices An n n matrix D is an EDM if p 1,..., p n R r D ij = p i p j 2, for all i, j = 1,..., n E n := set of all n n EDMs Embedding Dimension of D E n { } embdim(d) := min r : p 1,..., p n R r s.t. D ij = p i p j 2, for all i, j Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
12 Introduction Partial Euclidean Distance Matrices D is a partial EDM in R r if every entry D ij is either specified or unspecified, diag(d) = 0 for all α {1,..., n}, if the principal submatrix D[α] is fully specified, then D[α] is an EDM with embdim(d[α]) r Problem: determine if there is a completion ˆD E n with embdim( ˆD) = r Graph of a Partial EDM G = (N, E, ω) weighted graph with nodes N := {1,..., n} edges E := { ij : i j, and D ij is specified } weights ω R E + with ω ij := D ij Problem: determine if there is a realization p : N R r of G Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
13 Introduction n = 100, m = 9, R = 2 Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
14 Outline 1 Euclidean Distance Matrices Introduction Euclidean Distance Matrices and Semidefinite Matrices 2 Facial Reduction Semidefinite Facial Reduction 3 Results Facial Reduction Theory for EDM Completion Numerical Results Noisy Problems Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
15 EDMs and Semidefinite Matrices Linear Transformation K Let D E n be given by the points p 1,..., p n R r (or P R n r ) Let Y := PP T = (pi T p j ) S+ n D ij = p i p j 2 = pi T p i + pj T p j 2pi T p j = Y ii + Y jj 2Y ij Thus D = K(Y ), where: K(Y ) := diag(y )e T + ediag(y ) T 2Y K(S+) n = E n (but not one-to-one) Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
16 EDMs and Semidefinite Matrices Moore-Penrose Pseudoinverse of K K (D) = 1 2 J [offdiag(d)] J where J := I 1 n eet Theorem: (Schoenberg, 1935) A matrix D with diag(d) = 0 is a Euclidean distance matrix if and only if K (D) is positive semidefinite. Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
17 EDMs and Semidefinite Matrices Properties of K and K S n C := {Y Sn : Ye = 0} and S n H := {D Sn : diag(d) = 0} K (SC n ) = Sn H and K (SH n ) = Sn C K ( S+ n SC n ) = E n and K (E n ) = S+ n SC n ( ) embdim(d) = rank K (D), for all D E n Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
18 EDMs and Semidefinite Matrices Vector Formulation Find p 1,..., p n R r such that p i p j 2 = D ij, ij E Matrix Formulation using K Find P R n r such that H K(Y ) = H D, Y = PP T Semidefinite Programming (SDP) Relaxation Find Y S+ n SC n such that H K(Y ) = H D Vector/Matrix Formulation is non-convex and NP-hard SDP Relaxation is tractable, but only problems of limited size can be directly handled by an SDP solver To solve this SDP, we use facial reduction to obtain a much smaller equivalent problem Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
19 Outline 1 Euclidean Distance Matrices Introduction Euclidean Distance Matrices and Semidefinite Matrices 2 Facial Reduction Semidefinite Facial Reduction 3 Results Facial Reduction Theory for EDM Completion Numerical Results Noisy Problems Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
20 Semidefinite Facial Reduction An LP Example minimize 2x 1 + 6x 2 x 3 2x 4 + 7x 5 subject to x 1 + x 2 + x 3 + x 4 = 1 x 1 x 2 x 3 + x 5 = -1 x 1, x 2, x 3, x 4, x 5 0 Summing the constraints: 2x 1 + x 4 + x 5 = 0 x 1 = x 4 = x 5 = 0 Restrict LP to the face { x R 5 + : x 1 = x 4 = x 5 = 0 } R 5 + minimize 6x 2 x 3 subject to x 2 + x 3 = 1 x 2, x 3 0 Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
21 Semidefinite Facial Reduction The Minimal Face If K E is a convex cone and S K, then face(s) := F Proposition: S F K Let K E be a convex cone and let S F K. If S and convex, then: face(s) = F if and only if S relint(f). Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
22 Semidefinite Facial Reduction Representing Faces of S n + If F S n + and X relint(f) with rank(x) = t, then F = US t +U T and relint(f) = US t ++U T, where U R n t has full column rank and range(u) = range(x). Semidefinite Facial Reduction If: Then: face ({ X S n + : A i, X = b i, i }) = US t +U T minimize C, X subject to A i, X = b i, i X S+ n = minimize C, UZU T subject to A i, UZU T = b i, i Z S+ t Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
23 Outline 1 Euclidean Distance Matrices Introduction Euclidean Distance Matrices and Semidefinite Matrices 2 Facial Reduction Semidefinite Facial Reduction 3 Results Facial Reduction Theory for EDM Completion Numerical Results Noisy Problems Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
24 Facial Reduction Theory for EDM Completion Theorem: Clique Facial Reduction Let: Then: D E k t := embdim( D) F := { Y S k + : K(Y ) = D } where Ū := [ U C e ] face( F) = ŪSt+1 + ŪT U C R k t full column rank and range(u C ) = range(k ( D)) Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
25 Facial Reduction Theory for EDM Completion Theorem: Extending the Facial Reduction Let D be a partial EDM, α {1,..., n}, and: If: Then: F := { Y S+ n SC n : H K(Y ) = H D} F α := { Y S+ n SC n : H[α] K(Y [α]) = H[α] D[α]} { } F α := Y S α + : H[α] K(Y ) = H[α] D[α] where U := face(f α ) face( F α ) ŪSt+1 + ŪT [Ū ] 0 R 0 I n (n α +t+1) ( US n α +t+1 + U T ) S n C Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
26 Facial Reduction Theory for EDM Completion Theorem: Distance Constraint Reduction Let D be a partial EDM, α {1,..., n}, and: F α := { Y S+ n SC n : H[α] K(Y [α]) = H[α] D[α]} { } F α := Y S α + : H[α] K(Y ) = H[α] D[α] face( F α ) ŪSt+1 + ŪT If Ȳ F α and β α is a clique with embdim(d[β]) = t, then: F α = where U := { Y ( US n α +t+1 + U T ) S n C : K(Y [β]) = D[β] } [Ū ] 0 R 0 I n (n α +t+1) Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
27 Facial Reduction Theory for EDM Completion Theorem: Disjoint Subproblems Let F be as above. Let {α i } l i=1 be disjoint subsets, α := α i, and : F i := { Y S+ n SC n : H[α i] K(Y [α i ]) = H[α i ] D[α i ] } { } F i := Y S α i + : H[α i] K(Y ) = H[α i ] D[α i ] If face( F i ) ŪiS t i +1 + ŪT i, then face(f) ( US n α +t+1 + U T ) S n C Ū where U := Ū l 0 Rn (n α +t+1) 0 0 I Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
28 Facial Reduction Theory for EDM Completion Facial Reduction Algorithm For each node i = 1,..., n, find a clique C i containing i Let C n+1 be the clique of anchors ( ) Let F i := U i S n C i +t i +1 + Ui T SC n be the corresponding faces Compute U R n (t+1) full column rank such that Then: range(u) = n+1 i=1 range(u i ) face ({ Y S n + S n C : H K(Y ) = H D}) ( US t+1 + UT ) S n C Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
29 Facial Reduction Theory for EDM Completion Rigid Intersection C i C j Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
30 Facial Reduction Theory for EDM Completion Lemma: Rigid Face Intersection Suppose: and U Then: 1, U Satisfies: U 1 0 I 0 U 1 = U 1 0 and U 2 = 0 U 2 0 I 0 U 2 2 Rk (t+1) full column rank with range(u 1 ) = range(u U 1 U := U 1 or U := U 2 (U 2 ) U 1 U 1 (U 1 ) U 2 U 2 U 2 range(u) = range(u 1 ) range(u 2 ) 2 ) Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
31 Facial Reduction Theory for EDM Completion Clique Union Node Absorption C i Rigid C i C j j C i Non-rigid Ci C j j Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
32 Facial Reduction Theory for EDM Completion Theorem: Euclidean Distance Matrix Completion Let D be a partial EDM and : If Ȳ F := { Y S+ n SC n : H K(Y ) = H D} ( ) face(f) US+ t+1 UT SC n = (UV )St +(UV ) T F and β is a clique with embdim(d[β]) = t, then: Ȳ = (UV ) Z (UV ) T, where Z is the unique solution of (JU[β, :]V )Z (JU[β, :]V ) T = K (D[β]) (1) F = { Ȳ } D := K( P P T ) E n is the unique completion of D, where P := UV Z 1/2 R n t Note: In this case, an SDP solver is not required. Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
33 Outline 1 Euclidean Distance Matrices Introduction Euclidean Distance Matrices and Semidefinite Matrices 2 Facial Reduction Semidefinite Facial Reduction 3 Results Facial Reduction Theory for EDM Completion Numerical Results Noisy Problems Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
34 Numerical Results Random noiseless problems Dimension r = 2 Square region: [0, 1] [0, 1] m = 4 anchors R = radio range Using only Rigid Clique Union and Rigid Node Absorption Error measure: Root Mean Square Deviation ( 1 RMSD := # positioned i positioned ) 1 2 p i pi true 2. Results averaged over 10 instances Used MATLAB on a 2.16 GHz Intel Core 2 Duo with 2 GB of RAM Source code SNLSDPclique available on author s website, released under a GNU General Public License Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
35 Numerical Results Face Representation Approach # # Sensors CPU sensors R Positioned Time RMSD s 2e s 3e s 2e s 4e s 8e s 7e s 1e s 3e s 7e s 1e s 2e s 1e s 2e m 53 s 7e m 46 s 9e-11 Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
36 Numerical Results Point Representation Approach # # Sensors CPU sensors R Positioned Time RMSD s 5e s 6e s 7e s 7e s 5e s 5e s 8e s 6e s 9e s 7e s 6e s 1e s 8e m 27 s 9e m 55 s 1e-15 Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
37 Outline 1 Euclidean Distance Matrices Introduction Euclidean Distance Matrices and Semidefinite Matrices 2 Facial Reduction Semidefinite Facial Reduction 3 Results Facial Reduction Theory for EDM Completion Numerical Results Noisy Problems Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
38 Noisy Problems Multiplicative Noise Model d 2 ij = p i p j 2 (1 + σε ij ) 2, for all ij E ε ij is normally distributed with mean 0 and standard deviation 1 σ 0 is the noise factor Least Squares Problem minimize ij E v 2 ij subject to p i p j 2 (1 + v ij ) 2 = d 2 n i=1 p i = 0 p 1,..., p n R r ij, for all ij E Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
39 Noisy Problems Point Representation Approach # CPU σ sensors R Time RMSD s 5e-16 1e s 1e-06 1e s 1e-04 1e s 7e s 5e-16 1e s 1e-06 1e s 1e-04 1e s 2e s 1e-15 1e s 1e-06 1e s 1e-04 1e s 2e-01 Note: No refinement technique is used in our numerical tests Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
40 Noisy Problems 1LFB nf = 0.01%, RMSD = LFB nf = 0.1%, RMSD = Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
41 Noisy Problems 1LFB nf = 1%, RMSD = LFB nf = 10%, RMSD = Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
42 Summary We developed a theory of semidefinite facial reduction for the EDM completion problem Using this theory, we can transform the SDP relaxations into much smaller equivalent problems We developed a highly efficient algorithm for EDM completion: SDP solver not required can solve problems with up to 100,000 sensors in a few minutes on a laptop computer obtained very high accuracy for noiseless problems our running times are highly competitive with other SDP-based codes Nathan Krislock (University of Waterloo) SDP Facial Reduction for EDM Completion WO / 41
Facial reduction for Euclidean distance matrix problems
Facial reduction for Euclidean distance matrix problems Nathan Krislock Department of Mathematical Sciences Northern Illinois University, USA Distance Geometry: Theory and Applications DIMACS Center, Rutgers
More informationExplicit Sensor Network Localization using Semidefinite Programming and Clique Reductions
Explicit Sensor Network Localization using Semidefinite Programming and Clique Reductions Nathan Krislock and Henry Wolkowicz Dept. of Combinatorics and Optimization University of Waterloo Haifa Matrix
More informationFacial Reduction for Cone Optimization
Facial Reduction for Cone Optimization Henry Wolkowicz Dept. Combinatorics and Optimization University of Waterloo Tuesday Apr. 21, 2015 (with: Drusvyatskiy, Krislock, (Cheung) Voronin) Motivation: Loss
More informationExplicit Sensor Network Localization using Semidefinite Programming and Facial Reduction
Explicit Sensor Network Localization using Semidefinite Programming and Facial Reduction Nathan Krislock and Henry Wolkowicz Dept. of Combinatorics and Optimization University of Waterloo INFORMS, San
More informationFurther Relaxations of the SDP Approach to Sensor Network Localization
Further Relaxations of the SDP Approach to Sensor Network Localization Zizhuo Wang, Song Zheng, Stephen Boyd, and Yinyu Ye September 27, 2006; Revised May 2, 2007 Abstract Recently, a semi-definite programming
More informationThe Graph Realization Problem
The Graph Realization Problem via Semi-Definite Programming A. Y. Alfakih alfakih@uwindsor.ca Mathematics and Statistics University of Windsor The Graph Realization Problem p.1/21 The Graph Realization
More informationFurther Relaxations of the SDP Approach to Sensor Network Localization
Further Relaxations of the SDP Approach to Sensor Network Localization Zizhuo Wang, Song Zheng, Stephen Boyd, and inyu e September 27, 2006 Abstract Recently, a semidefinite programming (SDP) relaxation
More informationc 2017 Society for Industrial and Applied Mathematics
SIAM J. OPTIM. Vol. 27, No. 4, pp. 2301 2331 c 2017 Society for Industrial and Applied Mathematics NOISY EUCLIDEAN DISTANCE REALIZATION: ROBUST FACIAL REDUCTION AND THE PARETO FRONTIER D. DRUSVYATSKIY,
More informationTheory and Applications of Degeneracy in Cone Optimization
Theory and Applications of Degeneracy in Cone Optimization Henry Wolkowicz Dept. of Combinatorics and Optimization University of Waterloo Colloquium at: Dept. of Math. & Stats, Sept. 24 Outline Part I:
More informationA Unified Theorem on SDP Rank Reduction. yyye
SDP Rank Reduction Yinyu Ye, EURO XXII 1 A Unified Theorem on SDP Rank Reduction Yinyu Ye Department of Management Science and Engineering and Institute of Computational and Mathematical Engineering Stanford
More informationTheory of semidefinite programming for Sensor Network Localization
Math. Program. Ser. B DOI 10.1007/s10107-006-0040-1 FULL LENGTH PAPER Theory of semidefinite programming for Sensor Network Localization Anthony Man-Cho So Yinyu Ye Received: 22 November 2004 / Accepted:
More informationSOCP Relaxation of Sensor Network Localization
SOCP Relaxation of Sensor Network Localization Paul Tseng Mathematics, University of Washington Seattle IWCSN, Simon Fraser University August 19, 2006 Abstract This is a talk given at Int. Workshop on
More informationA Semidefinite Programming Approach to Tensegrity Theory and Realizability of Graphs
A Semidefinite Programming Approach to Tensegrity Theory and Realizability of Graphs Anthony Man Cho So, Yinyu Ye July 6, 2005 Abstract Recently, Connelly and Sloughter [12] have introduced the notion
More informationOn SDP and ESDP Relaxation of Sensor Network Localization
On SDP and ESDP Relaxation of Sensor Network Localization Paul Tseng Mathematics, University of Washington Seattle IASI, Roma September 16, 2008 (joint work with Ting Kei Pong) ON SDP/ESDP RELAXATION OF
More informationEuclidean Distance Matrices, Semidefinite Programming, and Sensor Network Localization
Euclidean Distance Matrices, Semidefinite Programming, and Sensor Network Localization Abdo Y. Alfakih Miguel F. Anjos Veronica Piccialli Henry Wolkowicz November 2, 2009 University of Waterloo Department
More informationOn SDP and ESDP Relaxation of Sensor Network Localization
On SDP and ESDP Relaxation of Sensor Network Localization Paul Tseng Mathematics, University of Washington Seattle Southern California Optimization Day, UCSD March 19, 2009 (joint work with Ting Kei Pong)
More informationA Gradient Search Method to Round the Semidefinite Programming Relaxation Solution for Ad Hoc Wireless Sensor Network Localization
A Gradient Search Method to Round the Semidefinite Programming Relaxation Solution for Ad Hoc Wireless Sensor Network Localization Tzu-Chen Liang, Ta-Chung Wang, and Yinyu Ye August 28, 2004; Updated October
More informationA Semidefinite Programming Approach to Tensegrity Theory and Realizability of Graphs
A Semidefinite Programming Approach to Tensegrity Theory and Realizability of Graphs Anthony Man Cho So, Yinyu Ye Abstract Recently, Connelly and Sloughter [14] have introduced the notion of d realizability
More informationRobust Principal Component Analysis
ELE 538B: Mathematics of High-Dimensional Data Robust Principal Component Analysis Yuxin Chen Princeton University, Fall 2018 Disentangling sparse and low-rank matrices Suppose we are given a matrix M
More informationNetwork Localization via Schatten Quasi-Norm Minimization
Network Localization via Schatten Quasi-Norm Minimization Anthony Man-Cho So Department of Systems Engineering & Engineering Management The Chinese University of Hong Kong (Joint Work with Senshan Ji Kam-Fung
More informationUniversal Rigidity and Edge Sparsification for Sensor Network Localization
Universal Rigidity and Edge Sparsification for Sensor Network Localization Zhisu Zhu Anthony Man Cho So Yinyu Ye September 23, 2009 Abstract Owing to their high accuracy and ease of formulation, there
More informationFacial Reduction in Cone Optimization with Applications to Matrix Completions
1 Facial Reduction in Cone Optimization with Applications to Matrix Completions Henry Wolkowicz Dept. Combinatorics and Optimization, University of Waterloo, Canada Wed. July 27, 2016, 2-3:20 PM at: DIMACS
More informationBeyond Convex Relaxation: A Polynomial Time Non Convex Optimization Approach to Network Localization
Beyond Convex Relaxation: A Polynomial Time Non Convex Optimization Approach to Network Localization Senshan Ji, Kam Fung Sze, Zirui Zhou, Anthony Man Cho So and Yinyu Ye Department of Systems Engineering
More informationSensor Network Localization, Euclidean Distance Matrix Completions, and Graph Realization
Sensor Network Localization, Euclidean Distance Matrix Completions, and Graph Realization Yichuan Ding Nathan Krislock Jiawei Qian Henry Wolkowicz October 27, 2008 University of Waterloo Department of
More informationSpherical Euclidean Distance Embedding of a Graph
Spherical Euclidean Distance Embedding of a Graph Hou-Duo Qi University of Southampton Presented at Isaac Newton Institute Polynomial Optimization August 9, 2013 Spherical Embedding Problem The Problem:
More informationIntroducing PENLAB a MATLAB code for NLP-SDP
Introducing PENLAB a MATLAB code for NLP-SDP Michal Kočvara School of Mathematics, The University of Birmingham jointly with Jan Fiala Numerical Algorithms Group Michael Stingl University of Erlangen-Nürnberg
More informationarxiv: v1 [math.oc] 26 Sep 2015
arxiv:1509.08021v1 [math.oc] 26 Sep 2015 Degeneracy in Maximal Clique Decomposition for Semidefinite Programs Arvind U. Raghunathan and Andrew V. Knyazev Mitsubishi Electric Research Laboratories 201 Broadway,
More informationCooperative Non-Line-of-Sight Localization Using Low-rank + Sparse Matrix Decomposition
Cooperative Non-Line-of-Sight Localization Using Low-rank + Sparse Matrix Decomposition Venkatesan Ekambaram Kannan Ramchandran Electrical Engineering and Computer Sciences University of California at
More informationSOCP Relaxation of Sensor Network Localization
SOCP Relaxation of Sensor Network Localization Paul Tseng Mathematics, University of Washington Seattle University of Vienna/Wien June 19, 2006 Abstract This is a talk given at Univ. Vienna, 2006. SOCP
More informationLocalization from Incomplete Noisy Distance Measurements
Localization from Incomplete Noisy Distance Measurements Adel Javanmard Department of Electrical Engineering Stanford University Andrea Montanari Department of Electrical Engineering and Department of
More informationComputing the Nearest Euclidean Distance Matrix with Low Embedding Dimensions
Computing the Nearest Euclidean Distance Matrix with Low Embedding Dimensions Hou-Duo Qi and Xiaoming Yuan September 12, 2012 and Revised May 12, 2013 Abstract Euclidean distance embedding appears in many
More informationECE 8201: Low-dimensional Signal Models for High-dimensional Data Analysis
ECE 8201: Low-dimensional Signal Models for High-dimensional Data Analysis Lecture 7: Matrix completion Yuejie Chi The Ohio State University Page 1 Reference Guaranteed Minimum-Rank Solutions of Linear
More informationSecond Order Cone Programming Relaxation of Positive Semidefinite Constraint
Research Reports on Mathematical and Computing Sciences Series B : Operations Research Department of Mathematical and Computing Sciences Tokyo Institute of Technology 2-12-1 Oh-Okayama, Meguro-ku, Tokyo
More informationOn Equivalence of Semidefinite Relaxations for Quadratic Matrix Programming
MAHEMAICS OF OPERAIONS RESEARCH Vol. 36, No., February 20, pp. 88 04 issn0364-765x eissn526-547 360 0088 informs doi 0.287/moor.00.0473 20 INFORMS On Equivalence of Semidefinite Relaxations for Quadratic
More informationLow-Rank Matrix Completion using Nuclear Norm with Facial Reduction
1 2 3 4 Low-Rank Matrix Completion using Nuclear Norm with Facial Reduction Shimeng Huang Henry Wolkowicz Sunday 14 th August, 2016, 17:43 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
More informationFast ADMM for Sum of Squares Programs Using Partial Orthogonality
Fast ADMM for Sum of Squares Programs Using Partial Orthogonality Antonis Papachristodoulou Department of Engineering Science University of Oxford www.eng.ox.ac.uk/control/sysos antonis@eng.ox.ac.uk with
More informationResearch Reports on Mathematical and Computing Sciences
ISSN 1342-284 Research Reports on Mathematical and Computing Sciences Exploiting Sparsity in Linear and Nonlinear Matrix Inequalities via Positive Semidefinite Matrix Completion Sunyoung Kim, Masakazu
More informationDegeneracy in Maximal Clique Decomposition for Semidefinite Programs
MITSUBISHI ELECTRIC RESEARCH LABORATORIES http://www.merl.com Degeneracy in Maximal Clique Decomposition for Semidefinite Programs Raghunathan, A.U.; Knyazev, A. TR2016-040 July 2016 Abstract Exploiting
More informationOptimal Value Function Methods in Numerical Optimization Level Set Methods
Optimal Value Function Methods in Numerical Optimization Level Set Methods James V Burke Mathematics, University of Washington, (jvburke@uw.edu) Joint work with Aravkin (UW), Drusvyatskiy (UW), Friedlander
More informationHandout 6: Some Applications of Conic Linear Programming
ENGG 550: Foundations of Optimization 08 9 First Term Handout 6: Some Applications of Conic Linear Programming Instructor: Anthony Man Cho So November, 08 Introduction Conic linear programming CLP, and
More informationModeling with semidefinite and copositive matrices
Modeling with semidefinite and copositive matrices Franz Rendl http://www.math.uni-klu.ac.at Alpen-Adria-Universität Klagenfurt Austria F. Rendl, Singapore workshop 2006 p.1/24 Overview Node and Edge relaxations
More informationMathematical Optimization Models and Applications
Mathematical Optimization Models and Applications Yinyu Ye Department of Management Science and Engineering Stanford University Stanford, CA 94305, U.S.A. http://www.stanford.edu/ yyye Chapters 1, 2.1-2,
More informationHybrid System Identification via Sparse Polynomial Optimization
2010 American Control Conference Marriott Waterfront, Baltimore, MD, USA June 30-July 02, 2010 WeA046 Hybrid System Identification via Sparse Polynomial Optimization Chao Feng, Constantino M Lagoa and
More informationSemidefinite Relaxations for Non-Convex Quadratic Mixed-Integer Programming
Semidefinite Relaxations for Non-Convex Quadratic Mixed-Integer Programming Christoph Buchheim 1 and Angelika Wiegele 2 1 Fakultät für Mathematik, Technische Universität Dortmund christoph.buchheim@tu-dortmund.de
More informationSparse Covariance Selection using Semidefinite Programming
Sparse Covariance Selection using Semidefinite Programming A. d Aspremont ORFE, Princeton University Joint work with O. Banerjee, L. El Ghaoui & G. Natsoulis, U.C. Berkeley & Iconix Pharmaceuticals Support
More informationDimension reduction for semidefinite programming
1 / 22 Dimension reduction for semidefinite programming Pablo A. Parrilo Laboratory for Information and Decision Systems Electrical Engineering and Computer Science Massachusetts Institute of Technology
More informationSolving Euclidean Distance Matrix Completion Problems Via Semidefinite Programming*
Computational Optimization and Applications 12, 13 30 1999) c 1999 Kluwer Academic Publishers. Manufactured in The Netherlands. Solving Euclidean Distance Matrix Completion Problems Via Semidefinite Programming*
More informationA Bregman alternating direction method of multipliers for sparse probabilistic Boolean network problem
A Bregman alternating direction method of multipliers for sparse probabilistic Boolean network problem Kangkang Deng, Zheng Peng Abstract: The main task of genetic regulatory networks is to construct a
More informationRobust Convex Approximation Methods for TDOA-Based Localization under NLOS Conditions
Robust Convex Approximation Methods for TDOA-Based Localization under NLOS Conditions Anthony Man-Cho So Department of Systems Engineering & Engineering Management The Chinese University of Hong Kong (CUHK)
More informationMATHEMATICAL ENGINEERING TECHNICAL REPORTS. Exact SDP Relaxations with Truncated Moment Matrix for Binary Polynomial Optimization Problems
MATHEMATICAL ENGINEERING TECHNICAL REPORTS Exact SDP Relaxations with Truncated Moment Matrix for Binary Polynomial Optimization Problems Shinsaku SAKAUE, Akiko TAKEDA, Sunyoung KIM, and Naoki ITO METR
More informationA SEMIDEFINITE PROGRAMMING APPROACH TO DISTANCE-BASED COOPERATIVE LOCALIZATION IN WIRELESS SENSOR NETWORKS
A SEMIDEFINITE PROGRAMMING APPROACH TO DISTANCE-BASED COOPERATIVE LOCALIZATION IN WIRELESS SENSOR NETWORKS By NING WANG A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL
More informationResearch Reports on Mathematical and Computing Sciences
ISSN 1342-2804 Research Reports on Mathematical and Computing Sciences Correlative sparsity in primal-dual interior-point methods for LP, SDP and SOCP Kazuhiro Kobayashi, Sunyoung Kim and Masakazu Kojima
More informationSensor Network Localization from Local Connectivity : Performance Analysis for the MDS-MAP Algorithm
Sensor Network Localization from Local Connectivity : Performance Analysis for the MDS-MAP Algorithm Sewoong Oh, Amin Karbasi, and Andrea Montanari August 27, 2009 Abstract Sensor localization from only
More informationTractable Upper Bounds on the Restricted Isometry Constant
Tractable Upper Bounds on the Restricted Isometry Constant Alex d Aspremont, Francis Bach, Laurent El Ghaoui Princeton University, École Normale Supérieure, U.C. Berkeley. Support from NSF, DHS and Google.
More informationResearch Reports on Mathematical and Computing Sciences
ISSN 1342-2804 Research Reports on Mathematical and Computing Sciences Sums of Squares and Semidefinite Programming Relaxations for Polynomial Optimization Problems with Structured Sparsity Hayato Waki,
More informationDistance Geometry-Matrix Completion
Christos Konaxis Algs in Struct BioInfo 2010 Outline Outline Tertiary structure Incomplete data Tertiary structure Measure diffrence of matched sets Def. Root Mean Square Deviation RMSD = 1 n x i y i 2,
More informationResearch Reports on Mathematical and Computing Sciences
ISSN 1342-2804 Research Reports on Mathematical and Computing Sciences Doubly Nonnegative Relaxations for Quadratic and Polynomial Optimization Problems with Binary and Box Constraints Sunyoung Kim, Masakazu
More informationGraph Laplacian Regularization for Large-Scale Semidefinite Programming
Graph Laplacian Regularization for Large-Scale Semidefinite Programming Kilian Q. Weinberger Dept of Computer and Information Science U of Pennsylvania, Philadelphia, PA 19104 kilianw@seas.upenn.edu Qihui
More informationSparsity in Sums of Squares of Polynomials
Research Reports on Mathematical and Computing Sciences Series B : Operations Research Department of Mathematical and Computing Sciences Tokyo Institute of Technology -1-1 Oh-Okayama, Meguro-ku, Tokyo
More informationRank-one Generated Spectral Cones Defined by Two Homogeneous Linear Matrix Inequalities
Rank-one Generated Spectral Cones Defined by Two Homogeneous Linear Matrix Inequalities C.J. Argue Joint work with Fatma Kılınç-Karzan October 22, 2017 INFORMS Annual Meeting Houston, Texas C. Argue, F.
More informationHandout 8: Dealing with Data Uncertainty
MFE 5100: Optimization 2015 16 First Term Handout 8: Dealing with Data Uncertainty Instructor: Anthony Man Cho So December 1, 2015 1 Introduction Conic linear programming CLP, and in particular, semidefinite
More informationEasy and not so easy multifacility location problems... (In 20 minutes.)
Easy and not so easy multifacility location problems... (In 20 minutes.) MINLP 2014 Pittsburgh, June 2014 Justo Puerto Institute of Mathematics (IMUS) Universidad de Sevilla Outline 1 Introduction (In
More informationSparse Matrix Theory and Semidefinite Optimization
Sparse Matrix Theory and Semidefinite Optimization Lieven Vandenberghe Department of Electrical Engineering University of California, Los Angeles Joint work with Martin S. Andersen and Yifan Sun Third
More informationSolving Global Optimization Problems with Sparse Polynomials and Unbounded Semialgebraic Feasible Sets
Solving Global Optimization Problems with Sparse Polynomials and Unbounded Semialgebraic Feasible Sets V. Jeyakumar, S. Kim, G. M. Lee and G. Li June 6, 2014 Abstract We propose a hierarchy of semidefinite
More informationProtein Structure by Semidefinite Facial Reduction
Protein Structure by Semidefinite Facial Reduction Babak Alipanahi 1, Nathan Krislock 2, Ali Ghodsi 3, Henry Wolkowicz 4, Logan Donaldson 5, and Ming Li 1 1 David R. Cheriton School of Computer Science,
More informationVerification Methods for Ill-Conditioned Relaxation Frameworks of Quadratic Assignment Problems.
Methods for Ill-Conditioned Relaxation Frameworks of Quadratic Assignment Problems. Hamburg University of Technology October 16, 2012 Outline Outline Outline Outline Introduction Applications Complexity
More informationOn the eigenvalues of Euclidean distance matrices
Volume 27, N. 3, pp. 237 250, 2008 Copyright 2008 SBMAC ISSN 00-8205 www.scielo.br/cam On the eigenvalues of Euclidean distance matrices A.Y. ALFAKIH Department of Mathematics and Statistics University
More informationBounds on the radius of nonsingularity
Towards bounds on the radius of nonsingularity, Milan Hladík Department of Applied Mathematics, Charles University, Prague and Institute of Computer Science of the Academy of Czech Republic SCAN 2014,
More informationUsing the Johnson-Lindenstrauss lemma in linear and integer programming
Using the Johnson-Lindenstrauss lemma in linear and integer programming Vu Khac Ky 1, Pierre-Louis Poirion, Leo Liberti LIX, École Polytechnique, F-91128 Palaiseau, France Email:{vu,poirion,liberti}@lix.polytechnique.fr
More informationw Kluwer Academic Publishers Boston/Dordrecht/London HANDBOOK OF SEMIDEFINITE PROGRAMMING Theory, Algorithms, and Applications
HANDBOOK OF SEMIDEFINITE PROGRAMMING Theory, Algorithms, and Applications Edited by Henry Wolkowicz Department of Combinatorics and Optimization Faculty of Mathematics University of Waterloo Waterloo,
More informationCOORDINATE SHADOWS OF SEMI-DEFINITE AND EUCLIDEAN DISTANCE MATRICES
3 4 5 6 7 8 9 0 COORDINATE SHADOWS OF SEMI-DEFINITE AND EUCLIDEAN DISTANCE MATRICES DMITRIY DRUSVYATSKIY, GÁBOR PATAKI, AND HENRY WOLKOWICZ Abstract. We consider the projected semi-definite and Euclidean
More information1. Introduction. To motivate the discussion, consider an undirected graph G
COORDINATE SHADOWS OF SEMI-DEFINITE AND EUCLIDEAN DISTANCE MATRICES DMITRIY DRUSVYATSKIY, GÁBOR PATAKI, AND HENRY WOLKOWICZ Abstract. We consider the projected semi-definite and Euclidean distance cones
More informationMatrix Completion: Fundamental Limits and Efficient Algorithms
Matrix Completion: Fundamental Limits and Efficient Algorithms Sewoong Oh PhD Defense Stanford University July 23, 2010 1 / 33 Matrix completion Find the missing entries in a huge data matrix 2 / 33 Example
More informationCopositive Programming and Combinatorial Optimization
Copositive Programming and Combinatorial Optimization Franz Rendl http://www.math.uni-klu.ac.at Alpen-Adria-Universität Klagenfurt Austria joint work with I.M. Bomze (Wien) and F. Jarre (Düsseldorf) IMA
More informationAn SDP-based divide-and-conquer algorithm for large scale noisy anchor-free graph realization
An SDP-based divide-and-conquer algorithm for large scale noisy anchor-free graph realization Ngai-Hang Z. Leung, and Kim-Chuan Toh August 19, 2008 Abstract We propose the DISCO algorithm for graph realization
More informationOn Equivalence of Semidefinite Relaxations for Quadratic Matrix Programming
On Equivalence of Semidefinite Relaxations for Quadratic Matrix Programming Yichuan Ding Dongdong Ge Henry Wolkowicz April 28, 2010 Contents 1 Introduction 2 1.1 Preliminaries........................................
More informationA direct formulation for sparse PCA using semidefinite programming
A direct formulation for sparse PCA using semidefinite programming A. d Aspremont, L. El Ghaoui, M. Jordan, G. Lanckriet ORFE, Princeton University & EECS, U.C. Berkeley Available online at www.princeton.edu/~aspremon
More informationConvex Euclidean distance embedding for collaborative position localization with NLOS mitigation
Noname manuscript No. (will be inserted by the editor) Convex Euclidean distance embedding for collaborative position localization with NLOS mitigation Chao Ding and Hou-Duo Qi the date of receipt and
More informationFast and Robust Phase Retrieval
Fast and Robust Phase Retrieval Aditya Viswanathan aditya@math.msu.edu CCAM Lunch Seminar Purdue University April 18 2014 0 / 27 Joint work with Yang Wang Mark Iwen Research supported in part by National
More informationPositive semidefinite rank
1/15 Positive semidefinite rank Hamza Fawzi (MIT, LIDS) Joint work with João Gouveia (Coimbra), Pablo Parrilo (MIT), Richard Robinson (Microsoft), James Saunderson (Monash), Rekha Thomas (UW) DIMACS Workshop
More informationSparse PCA with applications in finance
Sparse PCA with applications in finance A. d Aspremont, L. El Ghaoui, M. Jordan, G. Lanckriet ORFE, Princeton University & EECS, U.C. Berkeley Available online at www.princeton.edu/~aspremon 1 Introduction
More informationMaximum determinant positive definite Toeplitz completions
Maximum determinant positive definite Toeplitz completions Stefan Sremac, Hugo J. Woerdeman and Henry Wolkowicz Dedicated to our friend Rien Kaashoek in celebration of his eightieth birthday. Abstract.
More informationOn Representations of Graphs as Two-Distance Sets
arxiv:808.0595v [math.mg] 7 Aug 208 On Representations of Graphs as Two-Distance Sets A. Y. Alfakih Department of Mathematics and Statistics University of Windsor Windsor, Ontario N9B 3P4 Canada August
More informationA PARALLEL INTERIOR POINT DECOMPOSITION ALGORITHM FOR BLOCK-ANGULAR SEMIDEFINITE PROGRAMS
A PARALLEL INTERIOR POINT DECOMPOSITION ALGORITHM FOR BLOCK-ANGULAR SEMIDEFINITE PROGRAMS Kartik K. Sivaramakrishnan Department of Mathematics North Carolina State University kksivara@ncsu.edu http://www4.ncsu.edu/
More informationUniqueness of the Solutions of Some Completion Problems
Uniqueness of the Solutions of Some Completion Problems Chi-Kwong Li and Tom Milligan Abstract We determine the conditions for uniqueness of the solutions of several completion problems including the positive
More informationRELAX AND UNFOLD: MICROPHONE LOCALIZATION WITH EUCLIDEAN DISTANCE MATRICES
015 3rd European Signal Processing Conference (EUSIPCO) RELAX AND UNOLD: MICROPHONE LOCALIZATION WITH EUCLIDEAN DISTANCE MATRICES Ivan Dokmanić, Juri Ranieri, and Martin Vetterli School of Computer and
More informationA new graph parameter related to bounded rank positive semidefinite matrix completions
Mathematical Programming manuscript No. (will be inserted by the editor) Monique Laurent Antonios Varvitsiotis A new graph parameter related to bounded rank positive semidefinite matrix completions the
More informationCSCI 1951-G Optimization Methods in Finance Part 10: Conic Optimization
CSCI 1951-G Optimization Methods in Finance Part 10: Conic Optimization April 6, 2018 1 / 34 This material is covered in the textbook, Chapters 9 and 10. Some of the materials are taken from it. Some of
More informationProblem 1 (Exercise 2.2, Monograph)
MS&E 314/CME 336 Assignment 2 Conic Linear Programming January 3, 215 Prof. Yinyu Ye 6 Pages ASSIGNMENT 2 SOLUTIONS Problem 1 (Exercise 2.2, Monograph) We prove the part ii) of Theorem 2.1 (Farkas Lemma
More informationAcyclic Semidefinite Approximations of Quadratically Constrained Quadratic Programs
2015 American Control Conference Palmer House Hilton July 1-3, 2015. Chicago, IL, USA Acyclic Semidefinite Approximations of Quadratically Constrained Quadratic Programs Raphael Louca and Eilyan Bitar
More informationLecture 9: Laplacian Eigenmaps
Lecture 9: Radu Balan Department of Mathematics, AMSC, CSCAMM and NWC University of Maryland, College Park, MD April 18, 2017 Optimization Criteria Assume G = (V, W ) is a undirected weighted graph with
More informationarxiv: v1 [math.oc] 23 Nov 2012
arxiv:1211.5406v1 [math.oc] 23 Nov 2012 The equivalence between doubly nonnegative relaxation and semidefinite relaxation for binary quadratic programming problems Abstract Chuan-Hao Guo a,, Yan-Qin Bai
More informationSelf-Calibration and Biconvex Compressive Sensing
Self-Calibration and Biconvex Compressive Sensing Shuyang Ling Department of Mathematics, UC Davis July 12, 2017 Shuyang Ling (UC Davis) SIAM Annual Meeting, 2017, Pittsburgh July 12, 2017 1 / 22 Acknowledgements
More informationSemidefinite Programming Based Preconditioning for More Robust Near-Separable Nonnegative Matrix Factorization
Semidefinite Programming Based Preconditioning for More Robust Near-Separable Nonnegative Matrix Factorization Nicolas Gillis nicolas.gillis@umons.ac.be https://sites.google.com/site/nicolasgillis/ Department
More informationHow to generate weakly infeasible semidefinite programs via Lasserre s relaxations for polynomial optimization
CS-11-01 How to generate weakly infeasible semidefinite programs via Lasserre s relaxations for polynomial optimization Hayato Waki Department of Computer Science, The University of Electro-Communications
More informationA CONIC DANTZIG-WOLFE DECOMPOSITION APPROACH FOR LARGE SCALE SEMIDEFINITE PROGRAMMING
A CONIC DANTZIG-WOLFE DECOMPOSITION APPROACH FOR LARGE SCALE SEMIDEFINITE PROGRAMMING Kartik Krishnan Advanced Optimization Laboratory McMaster University Joint work with Gema Plaza Martinez and Tamás
More informationConvex Optimization of Graph Laplacian Eigenvalues
Convex Optimization of Graph Laplacian Eigenvalues Stephen Boyd Stanford University (Joint work with Persi Diaconis, Arpita Ghosh, Seung-Jean Kim, Sanjay Lall, Pablo Parrilo, Amin Saberi, Jun Sun, Lin
More informationProject Discussions: SNL/ADMM, MDP/Randomization, Quadratic Regularization, and Online Linear Programming
Project Discussions: SNL/ADMM, MDP/Randomization, Quadratic Regularization, and Online Linear Programming Yinyu Ye Department of Management Science and Engineering Stanford University Stanford, CA 94305,
More informationGraph and Controller Design for Disturbance Attenuation in Consensus Networks
203 3th International Conference on Control, Automation and Systems (ICCAS 203) Oct. 20-23, 203 in Kimdaejung Convention Center, Gwangju, Korea Graph and Controller Design for Disturbance Attenuation in
More informationA solution approach for linear optimization with completely positive matrices
A solution approach for linear optimization with completely positive matrices Franz Rendl http://www.math.uni-klu.ac.at Alpen-Adria-Universität Klagenfurt Austria joint work with M. Bomze (Wien) and F.
More information