A primal-dual Laplacian gradient flow dynamics for distributed resource allocation problems

Size: px
Start display at page:

Download "A primal-dual Laplacian gradient flow dynamics for distributed resource allocation problems"

Transcription

1 2018 Annual American Conrol Conference (ACC) June 27 29, Wisconsin Cener, Milwaukee, USA A primal-dual Laplacian gradien flow dynamics for disribued resource allocaion problems Dongsheng Ding and Mihailo R. Jovanović Absrac We employ he proximal augmened Lagrangian mehod o solve a class of convex resource allocaion problems over a conneced undireced nework of n agens. The agens are coupled by a linear resource equaliy consrain and heir saes are confined o a nonnegaive orhan. By inroducing he indicaor funcion associaed wih a nonnegaive orhan, we bring he problem ino a composie form wih a nonsmooh objecive and linear equaliy consrains. A primaldual Laplacian gradien flow dynamics based on he proximal augmened Lagrangian is proposed o solve he problem in a disribued way. These dynamics conserve he sum of he agen saes and he corresponding equilibrium poins are he Karush- Kuhn-Tucker poins of he original problem. We combine a Lyapunov-based argumen wih LaSalle s invariance principle o esablish global asympoic sabiliy and use an economic dispach case sudy o demonsrae he effeciveness of he proposed algorihm. I. INTRODUCTION The resource allocaion problem considers how o allocae he resource o individuals by minimizing he cos of he allocaion. I arises in many areas, including communicaion neworks, power grids, and producion managemen in economics [1]. Due o he emergence of he nex-generaion communicaion neworks and he cyber-physical sysems, he resource allocaion problem seeks o allocae he given resource in a disribued way, using local informaion exchange in a nework. Paricular insances of resource allocaion are given by he nework uiliy maximizaion [2] [5] and he economic dispach problems [6] [10]. A class of resource allocaion problems can be formulaed as follows x f(x) x i Ω i, i = 1,..., n where f(x) = n i=1 f i(x i ) is a separable convex objecive funcion, x = [ x 1 x n ] T R n is he opimizaion variable, 1 is he vecor of all-ones, b R is he given resource, and Ω i is a convex se. All x i s are coupled by he resource equaliy consrain and each x i has a se consrain. Disribued resource allocaion problems have araced significan aenion in recen years. In his seup, each node i in a nework has a local objecive funcion f i and a decision variable x i. The nodes can exchange informaion Financial suppor from he Naional Science Foundaion under Award ECCS is graefully acknowledged. D. Ding and M. R. Jovanović are wih he Ming Hsieh Deparmen of Elecrical Engineering, Universiy of Souhern California, Los Angeles, CA s: dongshed@usc.edu, mihailo@usc.edu. (1) wih heir neighbors wih he goal of compuing soluion of he resource allocaion problem in a disribued manner. Many disribued algorihms have been proposed for (1) or have been applied o (1) as a case sudy. Two main classes include he consensus-based algorihms and he primal-dual dynamics. Some recen works in he firs class include [7] [11], where consensus algorihms were incorporaed ino he cenralized gradien descen o enable disribued compuaion. In [8], a robus disribued algorihm was presened for he problem wih box consrains. In [9], wo disribued projeced algorihms were proposed. While [8] and [9] do no require any special iniializaion, algorihms in [7], [10], [11] do need paricular iniializaion. For (1) wihou he se consrain, he exponenial convergence of consensus-based algorihms was esablished in [9] [11]. The second class is based on primal-dual dynamics [12] [15], where he se consrain can be inequaliy consrains. In [12], [13], for a non-smooh augmened Lagrangian, i was shown ha is primal-dual dynamics can converge exponenially for (1) wih eiher he resource equaliy consrain or he se consrain. In [14], [15], global asympoic sabiliy of he projeced primal-dual dynamics was esablished by uilizing local convexiy-concaviy of he saddle funcion. The asympoic convergence of he primal-dual dynamics was furher discussed in [16] [18]. Alhough he primal-dual dynamics are asympoically sable, he se consrain may cause he difficuly in esablishing he exponenial convergence and he resource allocaion consrain can impede he disribued implemenaion. Moivaed by he above resuls, we are ineresed in developing a disribued primal-dual algorihm o solve (1) wih non-negaive consrains. To deal wih such consrains, we uilize he proximal augmened Lagrangian mehod [19] which yields a non-smooh composie opimizaion problem wih linear equaliy consrains. For his class of problems, he gradien flow dynamics associaed wih he proximal Lagrangian are no convenien for disribued implemenaion. Insead, we propose a primal-dual Laplacian gradien flow dynamics o compue he soluion via in-nework opimizaion and prove ha hese dynamics converge o he opima globally and asympoically. Finally, we use he economic dispach problem o illusrae he performance of he proposed algorihm. Our presenaion is organized as follows. In Secion II, we formulae he problem and provide background on he proximal augmened Lagrangian mehod. In Secion III, we propose primal-dual Laplacian gradien flow dynamics for /$ AACC 5316

2 in-nework resource allocaion and combine a Lyapunovbased argumen wih LaSalle s invariance principle o esablish global asympoic sabiliy. In Secion IV, we use compuaional experimens on an economic dispach problem o demonsrae meris and effeciveness of our approach. We conclude he paper and highligh fuure direcions in Secion V. II. PROBLEM FORMULATION AND BACKGROUND In his secion, we formulae he resource allocaion problem as a non-smooh composie opimizaion problem and inroduce he proximal augmened Lagrangian. We consider he resource allocaion problem, x f(x) x 0 where f is a convex objecive funcion, x R n is he opimizaion variable, and b is he given resource. For b > 0, opimizaion problem (2) is feasible. By inroducing an auxiliary variable z := [ z 1 z n ] T, (2) can be cas as a nonsmooh composie opimizaion wih equaliy consrains, x, z f(x) + g(z) x z = 0 where g(z) = n i = 1 g i(z i ) and each g i is an indicaor funcion of he non-negaive orhan, 0, zi 0 g i (z i ) := (4), oherwise. Since he objecive funcion in (3) is non-differeniable, he exising resuls (e.g., [13]) canno be employed direcly o esablish exponenial convergence. The proximal augmened Lagrangian mehod was recenly proposed in [19] o deal wih non-smooh composie opimizaion problems. The proposed approach explois he proximal operaor associaed wih he non-smooh par of he objecive funcion o eliminae he auxiliary variable z from he augmened Lagrangian. The proximal operaor associaed wih he funcion g is he r of he following opimizaion problem prox µg (v) := argmin z g(z) + 1 2µ z v 2. Here, v is a given vecor, µ is a posiive parameer, and he associaed value funcion specifies he Moreau envelope, M µg (v) = g(prox µg (v)) + 1 2µ prox µg(v) v 2. Noably, he Moreau envelope is coninuously differeniable, even when g is no, and is gradien is deermined by M µg (v) = 1 ( v proxµg (v) ). µ For he indicaor funcion of he non-negaive orhan (4), we (2) (3) have prox µg (v i ) = [v i ] + = M µg (v i ) = 1 2µ [v i] 2, M µg (v i ) = 1 µ [v i]. vi, v i 0 0, v i < 0 The Lagrangian associaed wih (3) is given by L(x, z; λ, y) = f(x) + g(z) + λ(1 T x b) + y T (x z) (5) where λ R and y R n are he Lagrange mulipliers. The augmened Lagrangian conains addiional erms ha inroduce quadraic penalies on he violaion of linear consrains, L µ (x, z; λ, y) = L(x, z; λ, y) + 1 2µ ((1T x b) 2 + x z 2 ) where µ is a posiive parameer. Wihou loss of generaliy, we have seleced he same penaly on he violaion of boh linear consrains in (3). The compleion of squares can be used o rewrie L µ as L µ (x, z; λ, y) = f(x) + λ(1 T x b) + 1 2µ (1T x b) 2 + g(z) + 1 2µ z (x + µy) 2 µ 2 y 2 (7) and he explici r of (7) wih respec o z is given by z µ(x, y) = prox µg (x + µy). As demonsraed in [19], he proximal augmened Lagrangian is obained by resricing he augmened Lagrangian on he manifold ha resuls from explici minimizaion of L µ over z. This eliminaes he non-smooh erm from L µ and cass i in erms of he Moreau envelope, L µ (x; λ, y) = f(x) + λ(1 T x b) + 1 2µ (1T x b) 2 + M µg (x + µy) µ 2 y 2 (8) The Arrow-Hurwicz-Uzawa gradien flow dynamics based on proximal augmened Lagrangian (8) is given by or, equivalenly, ẋ = L µ (x; λ, y) λ = + L µ (x; λ, y) ẏ = + L µ (x; λ, y) ẋ = ( f(x) + M µg (x + µy) + 1λ + 1 µ 1(1T x b)) λ = 1 T x b ẏ = µ ( M µg (x + µy) y). (9) In [19], global asympoic sabiliy of he gradien flow dynamics for problems of he form (3), bu wihou he resource allocaion consrain, was esablished. The above gradien flow dynamics accouns for his consrain; however, due o he appearance of 1 T x, he updae of x i in (9) requires global informaion which impedes disribued implemenaion. In wha follows, we modify (9) in order o solve he resource allocaion problem (2) over a conneced undireced nework (6) 5317

3 and esablish global asympoic sabiliy of he resuling primal-dual Laplacian gradien flow dynamics. III. PRIMAL-DUAL LAPLACIAN GRADIENT FLOW DYNAMICS In his secion, we propose a modificaion of he Arrow- Hurwicz-Uzawa gradien flow dynamics ha saisfies he resource allocaion consrain for all imes. We show ha he equilibrium poins of he resuling Laplacian gradien flow dynamics correspond o he KKT poins in (3) and prove global asympoic sabiliy. I is easy o show ha he following primal-dual Laplacian gradien flow dynamics ẋ = L ( f(x) + M µg (x + µy)) ẏ = µ ( M µg (x + µy) y) (10) saisfies 1 T x() b = 0 for all 0 if 1 T x(0) b = 0. This is because 1 T ẋ = 0 implies ha 1 T x() is a conserved quaniy of (10). Furhermore, since f(x) and M µg (x+µy) are sums of n individual funcions, primal-dual Laplacian gradien flow dynamics (10) is a disribued algorihm. The se of KKT poins for (3) is given by } Ω f(x) + 1λ + y = 0, x z = 0 = (x, z, λ, y) y g(z), 1 T x b = 0 where g(z) is he sub-gradien se of g a z. On he oher hand, if 1 T x(0) b = 0, he equilibrium poins ( x, ȳ) of (10) are given by } L ( f(x) + M µg (x + µy)) = 0 Ω = (x, y) M µg (x + µy) y = 0, 1 T x b = 0 The following lemma esablishes correspondence beween he ses Ω and Ω. Lemma 1: For any ( x, ȳ) Ω, here exiss z and λ such ha ( x, z, λ, ȳ) Ω. Similarly, for any (x, z, λ, y ) Ω, (x, λ ) Ω. Proof: For any ( x, ȳ) Ω, f( x) + M µg ( x + µȳ) is in he null space of he Laplacian L and M µg ( x+µȳ) = ȳ. Thus, here is a non-zero λ such ha f( x) + ȳ + λ1 = 0. Furhermore, M µg ( x + µȳ) = 1 µ( x + µȳ proxµg ( x + µȳ) ) = ȳ implies x = prox µg ( x + µȳ) = z and, consequenly, ȳ = ( f( x) + λ1) g( x). Finally, since he equaliy consrain 1 T x = b is, by consrucion, saisfied for all imes, by selecing x = z = x, λ = λ, and y = ȳ, we conclude ha (x, z, λ, y ) Ω. Conversely, for any (x, z, λ, y ) Ω, we have L( f(x ) + y ) = 0, 1 T x = b, and y g(x ). Furhemore, since x = z = prox µg (x + µy ), from he definiion of he gradien of he Moreau envelope, we have M µg (x + µy ) = y. By selecing x = x, ȳ = y, we conclude ha ( x, ȳ) Ω. In wha follows, we conduc sabiliy analysis of primaldual Laplacian gradien flow dynamics (10) wih he iniial condiion 1 T x(0) = b. This is done for objecive funcions f in (2) ha are srongly convex and have Lipschiz coninuous gradiens. Assumpion 1: The objecive funcion f in (2) is srongly convex and is gradien is Lipschiz coninuous. A. Transformed primal-dual Laplacian gradien flow dynamics The eigenvalue decomposiion of he graph Laplacian L of a conneced undireced nework is given by L = V ΛV T = [ [ ] [ ] U 1 n 1] Λ 0 0 U T = UΛ U T 1 n 1T where Λ 0 is a diagonal marix of non-zero eigenvalues of L and he marix U R n (n 1) and saisfies he following properies: (i) U T U = I; (ii) UU T = I 1 n 11T ; (iii) 1 T U = 0 and U T 1 = 0. We inroduce an affine coordinae ransformaion x = Uξ + 1θ (11) where ξ R n 1 and θ = 1 n 1T x(0). From he above properies of U, we have ξ = U T x, and (10) can be ransformed ino he following form ξ = Λ 0 U T ( f(x) + M µg (x + µy) ) ẏ = µ ( M µg (x + µy) y) (12) where ξ(0) R n 1 and y(0) R n are arbirary iniial condiions. B. Global asympoic sabiliy We firs sae a lemma abou proximal operaors which is used in he proof of Theorem 3. Lemma 2: Le g : R n R be a proper, lower semiconinuous, convex funcion and le prox µg : R n R n be he corresponding proximal operaor. Then, for any a, b R n, we can wrie prox µg (a) prox µg (b) = D(a b) (13) where D is a symmeric marix saisfying 0 D I. Proof: See [19, Lemma 2]. We nex esablish global asympoic sabiliy of ransformed primal-dual Laplacian gradien flow dynamics (12). Theorem 3: Le Assumpion 1 hold. Then, (12) is globally asympoically sable. Proof: Le ( ξ, ỹ) denoe perurbaions around he equilibrium poins ( ξ, ȳ) of (12). Then, ξ = Λ 0 U T ( f(x) f( x) + 1 µ m ) ỹ = m µỹ (14) 5318

4 where m = µ( M µg (x + µy) M µg ( x + µȳ)), x = Uξ + 1θ, and x = U ξ + 1θ. Derivaive of he Lyapunov funcion candidae V ( ξ, ỹ) = 1 2 ξ T Λ 1 0 ξ ỹ 2 along he soluions of (14) is deermined by V = ξ T Λ 1 0 ξ + ỹ T ỹ = ξ T U T ( f(uξ + 1θ) f(u ξ + 1θ)) 1 µ ξ T U T m + ỹ T m µ ỹ 2 From Lemma 2, we have m = (I D)(U ξ + µỹ), where 0 D I, and V = ξ T U T ( f(uξ + 1θ) f(u ξ + 1θ)) 1 µ ξ T U T (I D)U ξ µỹ T Dỹ. (15) Srong convexiy of f implies V 0. Furhermore, V = 0 is equivalen o U ξ = 0 and Dỹ = 0. We nex show ha he larges invarian se of V = 0 is conained in Ω. Firs, from x = U ξ, we have x = x. Second, from he propery of he marix D in Lemma 2, we have prox µg ( x + µy) = prox µg ( x + µȳ). Third, x = prox µg ( x+µȳ) implies 1 µ ( x+µy prox µg( x+µy)) y = M µg ( x + µy) y = 0. Thus, (x, y) saisfies he second and hird condiions in Ω. Furhermore, since L( f( x) + M µg ( x + µy)) = L( M µg ( x + µy) M µg ( x + µȳ)) = L(I D)ỹ = Lỹ, subsiuion of U ξ = 0, Dỹ = 0, and x = 0 ino (14) implies ha he saionary poin for ỹ is ỹ = c1 where c is a non-zero scalar. Therefore, we have L( f( x) + M µg ( x + µy)) = 0 and he firs condiion in Ω holds for (x, y). By combing he above expressions, we conclude ha (x, y) = ( x, y) is in Ω. Thus, since V is radially unbounded LaSalle s invariance principle implies ha (12) is globally asympoically sable. IV. AN EXAMPLE: ECONOMIC DISPATCH PROBLEM In his secion, we provide an example of he economic dispach problem o illusrae he performance of he proposed primal-dual Laplacian gradien flow dynamics. An IEEE 118-bus benchmark problem [20] has 54 generaors and each generaor has a quadraic cos funcion, f i (x i ) = a i + b i x i + c i x 2 i, where x i is he power injecion a bus i, a i [6.78, 74.33], b i [8.3391, ], and c i [0.0024, ]. The load is b = The communicaion nework has a ring opology wih several addiional edges (1, 11), (11, 21), (21, 31), (31, 41), and (41, 51). To show he robusness o he flucuaions in elecriciy price, a = 2000 we increase he parameers in he objecive funcion by 20%. We use ODE45 in MATLAB o simulae primal-dual Laplacian gradien flow dynamics (10), wih n = 54 and µ = In all simulaions, he relaive and absolue error olerances are se o and 10 15, respecively, he iniial condiion for each generaor is se o x i (0) = 4200/54, and y(0) = 0. The simulaion resuls are shown in Figs. 1 and 2. Figures 1a and 2a demonsrae ha all power injecions x i say in he feasible region and converge o he opimal soluion compued by CVX [21]. The following relaive error x x 2 + y ȳ 2 x(0) x 2 + y(0) ȳ 2 is used o show he exponenial convergence in righ plos in Figs. 1 and 2, where verical axis is shown in he logarihmic scale. In boh cases, logarihmic relaive errors decrease linearly excep for he ime insan a which he objecive funcion has been perurbed. V. CONCLUDING REMARKS In his paper, we employ he proximal augmened Lagrangian mehod o solve a class of convex resource allocaion problems over a conneced undireced nework wih n agens. We show ha he resource allocaion problem can be formulaed as a composie opimizaion problem. To perform in-nework compuaions, we propose a primal-dual Laplacian gradien flow dynamics based on he proximal augmened Lagrangian. We demonsrae ha he equilibrium poins correspond o KKT poins of he original problem. A Lyapunov-based argumen is used o esablish global asympoic sabiliy and demonsrae ha he proposed gradien flow dynamics globally converge o he opimal soluion. Finally, we apply he proposed algorihm o an economic dispach problem o illusraes is effeciveness. Several fuure direcions are of ineres. Firs, as indicaed in our simulaions, he proposed gradien flow dynamics appear o be exponenially converging. The exponenial convergence of he primal-dual Laplacian gradien flow dynamics is an open problem ha requires furher invesigaion. Second, he robusness of he proposed algorihm o various uncerainy sources is worh exploring. Third, o deal wih a broader class of problems, i is of ineres o exend he proposed gradien flow dynamics o more general consrains. ACKNOWLEDGMENTS We would like o hank Sepideh Hassan-Moghaddam for useful discussion. REFERENCES [1] P. Michael, A survey on he coninuous nonlinear resource allocaion problem, Eur. J. Oper. Res., vol. 185, no. 1, pp. 1 46, [2] E. Wei, A. Ozdaglar, and A. Jadbabaie, A disribued Newon mehod for nework uiliy maximizaion, in Proceedings of he 49h IEEE Conference on Decision and Conrol, 2010, pp [3], A disribued Newon mehod for nework uiliy maximizaion I: Algorihm, IEEE Trans. Auom. Conrol, vol. 58, no. 9, pp , [4], A disribued Newon mehod for nework uiliy maximizaion Par II: Convergence, IEEE Trans. Auom. Conrol, vol. 58, no. 9, pp , [5] A. Beck, A. Nedic, A. Ozdaglar, and M. Teboulle, An (1/k) gradien mehod for nework resource allocaion problems, IEEE Trans. Conrol New. Sys., vol. 1, no. 1, pp , [6] A. Cherukuri and J. Corés, Disribued generaor coordinaion for iniializaion and anyime opimizaion in economic dispach, IEEE Trans. Conrol New. Sys., vol. 2, no. 3, pp ,

5 (a) x i (), i = 1,, 54 (b) Relaive error Fig. 1: (a) Power injecions from 54 generaors; and (b) relaive error o he opima. (a) x i (), i = 1,, 54 (b) Relaive error Fig. 2: (a) Power injecions from 54 generaors; and (b) relaive error o he opima. [7] S. Kia, Disribued opimal resource allocaion over neworked sysems and use of an ɛ-exac penaly funcion, 14h IFAC Symposium on Large Scale Complex Sysems, vol. 49, no. 4, pp , [8] A. Cherukuri and J. Corés, Iniializaion-free disribued coordinaion for economic dispach under varying loads and generaor commimen, Auomaica, vol. 74, pp , [9] Y. H. P. Yi and F. Liu, Iniializaion-free disribued algorihms for opimal resource allocaion wih feasibiliy consrains and applicaion o economic dispach of power sysems, Auomaica, vol. 74, pp , [10] S. Kia, An augmened Lagrangian disribued algorihm for an innework opimal resource allocaion problem, in Proceedings of he 2017 American Conrol Conference, 2017, pp [11], Disribued opimal in-nework resource allocaion algorihm design via a conrol heoreic approach, Sys. Conrol Le., vol. 107, pp , [12] F. A. S. Niederländer and J. Corés, Exponenially fas disribued coordinaion for nonsmooh convex opimizaion, in Proceedings of he 55h IEEE Conference on Decision and Conrol, 2016, pp [13] S. Niederländer and J. Corés, Disribued coordinaion for nonsmooh convex opimizaion via saddle-poin dynamics, 2016, arxiv: [14] S. L. A. Cherukuri, E. Mallada and J. Corés, The role of srong convexiy-concaviy in he convergence and robusness of he saddlepoin dynamics, in Proceedings of he 54h Annual Alleron Conference on Communicaion, Conrol, and Compuing, 2016, pp [15] A. Cherukuri, E. Mallada, S. Low, and J. Cores, The role of convexiy on saddle-poin dynamics: Lyapunov funcion and robusness, 2016, arxiv: [16] T. Holding and I. Lesas, On he convergence o saddle poins of concave-convex funcions, he gradien mehod and emergence of oscillaions, in Proceedings of he 53rd IEEE Conference on Decision and Conrol, 2014, pp [17] A. Cherukuri and J. Corés, Asympoic sabiliy of saddle poins under he saddle-poin dynamics, in Proceedings of he 2015 American Conrol Conference, 2015, pp [18] B. G. A. Cherukuri and J. Corés, Saddle-poin dynamics: condiions for asympoic sabiliy of saddle poins, SIAM J. Conrol Opim., vol. 55, no. 1, pp , [19] N. K. Dhingra, S. Z. Khong, and M. R. Jovanović, The proximal augmened Lagrangian mehod for nonsmooh composie opimizaion, IEEE Trans. Auoma. Conrol, 2016, submied; also arxiv: [20] Available online a hp://moor.ece.ii.edu/daa/jeas IEEE118.doc. [21] M. Gran and S. Boyd, CVX: Malab sofware for disciplined convex programming, version 2.1, hp://cvxr.com/cvx, Mar

A Primal-Dual Type Algorithm with the O(1/t) Convergence Rate for Large Scale Constrained Convex Programs

A Primal-Dual Type Algorithm with the O(1/t) Convergence Rate for Large Scale Constrained Convex Programs PROC. IEEE CONFERENCE ON DECISION AND CONTROL, 06 A Primal-Dual Type Algorihm wih he O(/) Convergence Rae for Large Scale Consrained Convex Programs Hao Yu and Michael J. Neely Absrac This paper considers

More information

An Augmented Lagrangian Distributed Algorithm for an In-network Optimal Resource Allocation Problem

An Augmented Lagrangian Distributed Algorithm for an In-network Optimal Resource Allocation Problem An Augmened Lagrangian Disribued Algorihm for an In-nework Opimal Resource Allocaion Problem Solmaz S. Kia (Appeared in he proceeding of ACC 217) Absrac This paper sudies disribued soluions for an opimal

More information

A Decentralized Second-Order Method with Exact Linear Convergence Rate for Consensus Optimization

A Decentralized Second-Order Method with Exact Linear Convergence Rate for Consensus Optimization 1 A Decenralized Second-Order Mehod wih Exac Linear Convergence Rae for Consensus Opimizaion Aryan Mokhari, Wei Shi, Qing Ling, and Alejandro Ribeiro Absrac This paper considers decenralized consensus

More information

Chapter 2. First Order Scalar Equations

Chapter 2. First Order Scalar Equations Chaper. Firs Order Scalar Equaions We sar our sudy of differenial equaions in he same way he pioneers in his field did. We show paricular echniques o solve paricular ypes of firs order differenial equaions.

More information

Course Notes for EE227C (Spring 2018): Convex Optimization and Approximation

Course Notes for EE227C (Spring 2018): Convex Optimization and Approximation Course Noes for EE7C Spring 018: Convex Opimizaion and Approximaion Insrucor: Moriz Hard Email: hard+ee7c@berkeley.edu Graduae Insrucor: Max Simchowiz Email: msimchow+ee7c@berkeley.edu Ocober 15, 018 3

More information

Lecture 9: September 25

Lecture 9: September 25 0-725: Opimizaion Fall 202 Lecure 9: Sepember 25 Lecurer: Geoff Gordon/Ryan Tibshirani Scribes: Xuezhi Wang, Subhodeep Moira, Abhimanu Kumar Noe: LaTeX emplae couresy of UC Berkeley EECS dep. Disclaimer:

More information

Aryan Mokhtari, Wei Shi, Qing Ling, and Alejandro Ribeiro. cost function n

Aryan Mokhtari, Wei Shi, Qing Ling, and Alejandro Ribeiro. cost function n IEEE TRANSACTIONS ON SIGNAL AND INFORMATION PROCESSING OVER NETWORKS, VOL. 2, NO. 4, DECEMBER 2016 507 A Decenralized Second-Order Mehod wih Exac Linear Convergence Rae for Consensus Opimizaion Aryan Mokhari,

More information

arxiv: v1 [math.gm] 4 Nov 2018

arxiv: v1 [math.gm] 4 Nov 2018 Unpredicable Soluions of Linear Differenial Equaions Mara Akhme 1,, Mehme Onur Fen 2, Madina Tleubergenova 3,4, Akylbek Zhamanshin 3,4 1 Deparmen of Mahemaics, Middle Eas Technical Universiy, 06800, Ankara,

More information

MATH 5720: Gradient Methods Hung Phan, UMass Lowell October 4, 2018

MATH 5720: Gradient Methods Hung Phan, UMass Lowell October 4, 2018 MATH 5720: Gradien Mehods Hung Phan, UMass Lowell Ocober 4, 208 Descen Direcion Mehods Consider he problem min { f(x) x R n}. The general descen direcions mehod is x k+ = x k + k d k where x k is he curren

More information

Online Appendix to Solution Methods for Models with Rare Disasters

Online Appendix to Solution Methods for Models with Rare Disasters Online Appendix o Soluion Mehods for Models wih Rare Disasers Jesús Fernández-Villaverde and Oren Levinal In his Online Appendix, we presen he Euler condiions of he model, we develop he pricing Calvo block,

More information

A DELAY-DEPENDENT STABILITY CRITERIA FOR T-S FUZZY SYSTEM WITH TIME-DELAYS

A DELAY-DEPENDENT STABILITY CRITERIA FOR T-S FUZZY SYSTEM WITH TIME-DELAYS A DELAY-DEPENDENT STABILITY CRITERIA FOR T-S FUZZY SYSTEM WITH TIME-DELAYS Xinping Guan ;1 Fenglei Li Cailian Chen Insiue of Elecrical Engineering, Yanshan Universiy, Qinhuangdao, 066004, China. Deparmen

More information

Network Newton Distributed Optimization Methods

Network Newton Distributed Optimization Methods Nework Newon Disribued Opimizaion Mehods Aryan Mokhari, Qing Ling, and Alejandro Ribeiro Absrac We sudy he problem of minimizing a sum of convex objecive funcions where he componens of he objecive are

More information

Stability and Bifurcation in a Neural Network Model with Two Delays

Stability and Bifurcation in a Neural Network Model with Two Delays Inernaional Mahemaical Forum, Vol. 6, 11, no. 35, 175-1731 Sabiliy and Bifurcaion in a Neural Nework Model wih Two Delays GuangPing Hu and XiaoLing Li School of Mahemaics and Physics, Nanjing Universiy

More information

Optimality Conditions for Unconstrained Problems

Optimality Conditions for Unconstrained Problems 62 CHAPTER 6 Opimaliy Condiions for Unconsrained Problems 1 Unconsrained Opimizaion 11 Exisence Consider he problem of minimizing he funcion f : R n R where f is coninuous on all of R n : P min f(x) x

More information

Ordinary Differential Equations

Ordinary Differential Equations Ordinary Differenial Equaions 5. Examples of linear differenial equaions and heir applicaions We consider some examples of sysems of linear differenial equaions wih consan coefficiens y = a y +... + a

More information

An Introduction to Backward Stochastic Differential Equations (BSDEs) PIMS Summer School 2016 in Mathematical Finance.

An Introduction to Backward Stochastic Differential Equations (BSDEs) PIMS Summer School 2016 in Mathematical Finance. 1 An Inroducion o Backward Sochasic Differenial Equaions (BSDEs) PIMS Summer School 2016 in Mahemaical Finance June 25, 2016 Chrisoph Frei cfrei@ualbera.ca This inroducion is based on Touzi [14], Bouchard

More information

Boundedness and Exponential Asymptotic Stability in Dynamical Systems with Applications to Nonlinear Differential Equations with Unbounded Terms

Boundedness and Exponential Asymptotic Stability in Dynamical Systems with Applications to Nonlinear Differential Equations with Unbounded Terms Advances in Dynamical Sysems and Applicaions. ISSN 0973-531 Volume Number 1 007, pp. 107 11 Research India Publicaions hp://www.ripublicaion.com/adsa.hm Boundedness and Exponenial Asympoic Sabiliy in Dynamical

More information

Appendix to Online l 1 -Dictionary Learning with Application to Novel Document Detection

Appendix to Online l 1 -Dictionary Learning with Application to Novel Document Detection Appendix o Online l -Dicionary Learning wih Applicaion o Novel Documen Deecion Shiva Prasad Kasiviswanahan Huahua Wang Arindam Banerjee Prem Melville A Background abou ADMM In his secion, we give a brief

More information

An introduction to the theory of SDDP algorithm

An introduction to the theory of SDDP algorithm An inroducion o he heory of SDDP algorihm V. Leclère (ENPC) Augus 1, 2014 V. Leclère Inroducion o SDDP Augus 1, 2014 1 / 21 Inroducion Large scale sochasic problem are hard o solve. Two ways of aacking

More information

The Asymptotic Behavior of Nonoscillatory Solutions of Some Nonlinear Dynamic Equations on Time Scales

The Asymptotic Behavior of Nonoscillatory Solutions of Some Nonlinear Dynamic Equations on Time Scales Advances in Dynamical Sysems and Applicaions. ISSN 0973-5321 Volume 1 Number 1 (2006, pp. 103 112 c Research India Publicaions hp://www.ripublicaion.com/adsa.hm The Asympoic Behavior of Nonoscillaory Soluions

More information

On Oscillation of a Generalized Logistic Equation with Several Delays

On Oscillation of a Generalized Logistic Equation with Several Delays Journal of Mahemaical Analysis and Applicaions 253, 389 45 (21) doi:1.16/jmaa.2.714, available online a hp://www.idealibrary.com on On Oscillaion of a Generalized Logisic Equaion wih Several Delays Leonid

More information

Hamilton- J acobi Equation: Weak S olution We continue the study of the Hamilton-Jacobi equation:

Hamilton- J acobi Equation: Weak S olution We continue the study of the Hamilton-Jacobi equation: M ah 5 7 Fall 9 L ecure O c. 4, 9 ) Hamilon- J acobi Equaion: Weak S oluion We coninue he sudy of he Hamilon-Jacobi equaion: We have shown ha u + H D u) = R n, ) ; u = g R n { = }. ). In general we canno

More information

Mean-square Stability Control for Networked Systems with Stochastic Time Delay

Mean-square Stability Control for Networked Systems with Stochastic Time Delay JOURNAL OF SIMULAION VOL. 5 NO. May 7 Mean-square Sabiliy Conrol for Newored Sysems wih Sochasic ime Delay YAO Hejun YUAN Fushun School of Mahemaics and Saisics Anyang Normal Universiy Anyang Henan. 455

More information

POSITIVE SOLUTIONS OF NEUTRAL DELAY DIFFERENTIAL EQUATION

POSITIVE SOLUTIONS OF NEUTRAL DELAY DIFFERENTIAL EQUATION Novi Sad J. Mah. Vol. 32, No. 2, 2002, 95-108 95 POSITIVE SOLUTIONS OF NEUTRAL DELAY DIFFERENTIAL EQUATION Hajnalka Péics 1, János Karsai 2 Absrac. We consider he scalar nonauonomous neural delay differenial

More information

Sliding Mode Extremum Seeking Control for Linear Quadratic Dynamic Game

Sliding Mode Extremum Seeking Control for Linear Quadratic Dynamic Game Sliding Mode Exremum Seeking Conrol for Linear Quadraic Dynamic Game Yaodong Pan and Ümi Özgüner ITS Research Group, AIST Tsukuba Eas Namiki --, Tsukuba-shi,Ibaraki-ken 5-856, Japan e-mail: pan.yaodong@ais.go.jp

More information

Inventory Analysis and Management. Multi-Period Stochastic Models: Optimality of (s, S) Policy for K-Convex Objective Functions

Inventory Analysis and Management. Multi-Period Stochastic Models: Optimality of (s, S) Policy for K-Convex Objective Functions Muli-Period Sochasic Models: Opimali of (s, S) Polic for -Convex Objecive Funcions Consider a seing similar o he N-sage newsvendor problem excep ha now here is a fixed re-ordering cos (> 0) for each (re-)order.

More information

Lecture 20: Riccati Equations and Least Squares Feedback Control

Lecture 20: Riccati Equations and Least Squares Feedback Control 34-5 LINEAR SYSTEMS Lecure : Riccai Equaions and Leas Squares Feedback Conrol 5.6.4 Sae Feedback via Riccai Equaions A recursive approach in generaing he marix-valued funcion W ( ) equaion for i for he

More information

The L p -Version of the Generalized Bohl Perron Principle for Vector Equations with Infinite Delay

The L p -Version of the Generalized Bohl Perron Principle for Vector Equations with Infinite Delay Advances in Dynamical Sysems and Applicaions ISSN 973-5321, Volume 6, Number 2, pp. 177 184 (211) hp://campus.ms.edu/adsa The L p -Version of he Generalized Bohl Perron Principle for Vecor Equaions wih

More information

Hamilton Jacobi equations

Hamilton Jacobi equations Hamilon Jacobi equaions Inoducion o PDE The rigorous suff from Evans, mosly. We discuss firs u + H( u = 0, (1 where H(p is convex, and superlinear a infiniy, H(p lim p p = + This by comes by inegraion

More information

Robust estimation based on the first- and third-moment restrictions of the power transformation model

Robust estimation based on the first- and third-moment restrictions of the power transformation model h Inernaional Congress on Modelling and Simulaion, Adelaide, Ausralia, 6 December 3 www.mssanz.org.au/modsim3 Robus esimaion based on he firs- and hird-momen resricions of he power ransformaion Nawaa,

More information

Adaptation and Synchronization over a Network: stabilization without a reference model

Adaptation and Synchronization over a Network: stabilization without a reference model Adapaion and Synchronizaion over a Nework: sabilizaion wihou a reference model Travis E. Gibson (gibson@mi.edu) Harvard Medical School Deparmen of Pahology, Brigham and Women s Hospial 55 h Conference

More information

Math Final Exam Solutions

Math Final Exam Solutions Mah 246 - Final Exam Soluions Friday, July h, 204 () Find explici soluions and give he inerval of definiion o he following iniial value problems (a) ( + 2 )y + 2y = e, y(0) = 0 Soluion: In normal form,

More information

The motions of the celt on a horizontal plane with viscous friction

The motions of the celt on a horizontal plane with viscous friction The h Join Inernaional Conference on Mulibody Sysem Dynamics June 8, 18, Lisboa, Porugal The moions of he cel on a horizonal plane wih viscous fricion Maria A. Munisyna 1 1 Moscow Insiue of Physics and

More information

On Robust Stability of Uncertain Neutral Systems with Discrete and Distributed Delays

On Robust Stability of Uncertain Neutral Systems with Discrete and Distributed Delays 009 American Conrol Conference Hya Regency Riverfron S. Louis MO USA June 0-009 FrC09.5 On Robus Sabiliy of Uncerain Neural Sysems wih Discree and Disribued Delays Jian Sun Jie Chen G.P. Liu Senior Member

More information

RANDOM LAGRANGE MULTIPLIERS AND TRANSVERSALITY

RANDOM LAGRANGE MULTIPLIERS AND TRANSVERSALITY ECO 504 Spring 2006 Chris Sims RANDOM LAGRANGE MULTIPLIERS AND TRANSVERSALITY 1. INTRODUCTION Lagrange muliplier mehods are sandard fare in elemenary calculus courses, and hey play a cenral role in economic

More information

Physics 235 Chapter 2. Chapter 2 Newtonian Mechanics Single Particle

Physics 235 Chapter 2. Chapter 2 Newtonian Mechanics Single Particle Chaper 2 Newonian Mechanics Single Paricle In his Chaper we will review wha Newon s laws of mechanics ell us abou he moion of a single paricle. Newon s laws are only valid in suiable reference frames,

More information

Technical Report Doc ID: TR March-2013 (Last revision: 23-February-2016) On formulating quadratic functions in optimization models.

Technical Report Doc ID: TR March-2013 (Last revision: 23-February-2016) On formulating quadratic functions in optimization models. Technical Repor Doc ID: TR--203 06-March-203 (Las revision: 23-Februar-206) On formulaing quadraic funcions in opimizaion models. Auhor: Erling D. Andersen Convex quadraic consrains quie frequenl appear

More information

Class Meeting # 10: Introduction to the Wave Equation

Class Meeting # 10: Introduction to the Wave Equation MATH 8.5 COURSE NOTES - CLASS MEETING # 0 8.5 Inroducion o PDEs, Fall 0 Professor: Jared Speck Class Meeing # 0: Inroducion o he Wave Equaion. Wha is he wave equaion? The sandard wave equaion for a funcion

More information

d 1 = c 1 b 2 - b 1 c 2 d 2 = c 1 b 3 - b 1 c 3

d 1 = c 1 b 2 - b 1 c 2 d 2 = c 1 b 3 - b 1 c 3 and d = c b - b c c d = c b - b c c This process is coninued unil he nh row has been compleed. The complee array of coefficiens is riangular. Noe ha in developing he array an enire row may be divided or

More information

STATE-SPACE MODELLING. A mass balance across the tank gives:

STATE-SPACE MODELLING. A mass balance across the tank gives: B. Lennox and N.F. Thornhill, 9, Sae Space Modelling, IChemE Process Managemen and Conrol Subjec Group Newsleer STE-SPACE MODELLING Inroducion: Over he pas decade or so here has been an ever increasing

More information

PENALIZED LEAST SQUARES AND PENALIZED LIKELIHOOD

PENALIZED LEAST SQUARES AND PENALIZED LIKELIHOOD PENALIZED LEAST SQUARES AND PENALIZED LIKELIHOOD HAN XIAO 1. Penalized Leas Squares Lasso solves he following opimizaion problem, ˆβ lasso = arg max β R p+1 1 N y i β 0 N x ij β j β j (1.1) for some 0.

More information

Energy-Efficient Link Adaptation on Parallel Channels

Energy-Efficient Link Adaptation on Parallel Channels Energy-Efficien Link Adapaion on Parallel Channels Chrisian Isheden and Gerhard P. Feweis Technische Universiä Dresden Vodafone Chair Mobile Communicaions Sysems D-0069 Dresden Germany Email: (see hp://www.vodafone-chair.com/

More information

Particle Swarm Optimization Combining Diversification and Intensification for Nonlinear Integer Programming Problems

Particle Swarm Optimization Combining Diversification and Intensification for Nonlinear Integer Programming Problems Paricle Swarm Opimizaion Combining Diversificaion and Inensificaion for Nonlinear Ineger Programming Problems Takeshi Masui, Masaoshi Sakawa, Kosuke Kao and Koichi Masumoo Hiroshima Universiy 1-4-1, Kagamiyama,

More information

10. State Space Methods

10. State Space Methods . Sae Space Mehods. Inroducion Sae space modelling was briefly inroduced in chaper. Here more coverage is provided of sae space mehods before some of heir uses in conrol sysem design are covered in he

More information

Primal-Dual Splitting: Recent Improvements and Variants

Primal-Dual Splitting: Recent Improvements and Variants Primal-Dual Spliing: Recen Improvemens and Varians 1 Thomas Pock and 2 Anonin Chambolle 1 Insiue for Compuer Graphics and Vision, TU Graz, Ausria 2 CMAP & CNRS École Polyechnique, France The proximal poin

More information

4. Advanced Stability Theory

4. Advanced Stability Theory Applied Nonlinear Conrol Nguyen an ien - 4 4 Advanced Sabiliy heory he objecive of his chaper is o presen sabiliy analysis for non-auonomous sysems 41 Conceps of Sabiliy for Non-Auonomous Sysems Equilibrium

More information

3.1.3 INTRODUCTION TO DYNAMIC OPTIMIZATION: DISCRETE TIME PROBLEMS. A. The Hamiltonian and First-Order Conditions in a Finite Time Horizon

3.1.3 INTRODUCTION TO DYNAMIC OPTIMIZATION: DISCRETE TIME PROBLEMS. A. The Hamiltonian and First-Order Conditions in a Finite Time Horizon 3..3 INRODUCION O DYNAMIC OPIMIZAION: DISCREE IME PROBLEMS A. he Hamilonian and Firs-Order Condiions in a Finie ime Horizon Define a new funcion, he Hamilonian funcion, H. H he change in he oal value of

More information

An Introduction to Malliavin calculus and its applications

An Introduction to Malliavin calculus and its applications An Inroducion o Malliavin calculus and is applicaions Lecure 5: Smoohness of he densiy and Hörmander s heorem David Nualar Deparmen of Mahemaics Kansas Universiy Universiy of Wyoming Summer School 214

More information

On-line Adaptive Optimal Timing Control of Switched Systems

On-line Adaptive Optimal Timing Control of Switched Systems On-line Adapive Opimal Timing Conrol of Swiched Sysems X.C. Ding, Y. Wardi and M. Egersed Absrac In his paper we consider he problem of opimizing over he swiching imes for a muli-modal dynamic sysem when

More information

Matrix Versions of Some Refinements of the Arithmetic-Geometric Mean Inequality

Matrix Versions of Some Refinements of the Arithmetic-Geometric Mean Inequality Marix Versions of Some Refinemens of he Arihmeic-Geomeric Mean Inequaliy Bao Qi Feng and Andrew Tonge Absrac. We esablish marix versions of refinemens due o Alzer ], Carwrigh and Field 4], and Mercer 5]

More information

EXERCISES FOR SECTION 1.5

EXERCISES FOR SECTION 1.5 1.5 Exisence and Uniqueness of Soluions 43 20. 1 v c 21. 1 v c 1 2 4 6 8 10 1 2 2 4 6 8 10 Graph of approximae soluion obained using Euler s mehod wih = 0.1. Graph of approximae soluion obained using Euler

More information

Constrained Flow Control in Storage Networks: Capacity Maximization and Balancing

Constrained Flow Control in Storage Networks: Capacity Maximization and Balancing Consrained Flow Conrol in Sorage Neworks: Capaciy Maximizaion and Balancing Claus Danielson a, Francesco Borrelli a, Douglas Oliver b, Dyche Anderson b, Tony Phillips b a Universiy of California, Berkeley,

More information

Haar Wavelet Operational Matrix Method for Solving Fractional Partial Differential Equations

Haar Wavelet Operational Matrix Method for Solving Fractional Partial Differential Equations Copyrigh 22 Tech Science Press CMES, vol.88, no.3, pp.229-243, 22 Haar Wavele Operaional Mari Mehod for Solving Fracional Parial Differenial Equaions Mingu Yi and Yiming Chen Absrac: In his paper, Haar

More information

arxiv: v1 [math.ca] 15 Nov 2016

arxiv: v1 [math.ca] 15 Nov 2016 arxiv:6.599v [mah.ca] 5 Nov 26 Counerexamples on Jumarie s hree basic fracional calculus formulae for non-differeniable coninuous funcions Cheng-shi Liu Deparmen of Mahemaics Norheas Peroleum Universiy

More information

di Bernardo, M. (1995). A purely adaptive controller to synchronize and control chaotic systems.

di Bernardo, M. (1995). A purely adaptive controller to synchronize and control chaotic systems. di ernardo, M. (995). A purely adapive conroller o synchronize and conrol chaoic sysems. hps://doi.org/.6/375-96(96)8-x Early version, also known as pre-prin Link o published version (if available):.6/375-96(96)8-x

More information

Optimal Server Assignment in Multi-Server

Optimal Server Assignment in Multi-Server Opimal Server Assignmen in Muli-Server 1 Queueing Sysems wih Random Conneciviies Hassan Halabian, Suden Member, IEEE, Ioannis Lambadaris, Member, IEEE, arxiv:1112.1178v2 [mah.oc] 21 Jun 2013 Yannis Viniois,

More information

A Necessary and Sufficient Condition for the Solutions of a Functional Differential Equation to Be Oscillatory or Tend to Zero

A Necessary and Sufficient Condition for the Solutions of a Functional Differential Equation to Be Oscillatory or Tend to Zero JOURNAL OF MAEMAICAL ANALYSIS AND APPLICAIONS 24, 7887 1997 ARICLE NO. AY965143 A Necessary and Sufficien Condiion for he Soluions of a Funcional Differenial Equaion o Be Oscillaory or end o Zero Piambar

More information

Random Walk with Anti-Correlated Steps

Random Walk with Anti-Correlated Steps Random Walk wih Ani-Correlaed Seps John Noga Dirk Wagner 2 Absrac We conjecure he expeced value of random walks wih ani-correlaed seps o be exacly. We suppor his conjecure wih 2 plausibiliy argumens and

More information

Math-Net.Ru All Russian mathematical portal

Math-Net.Ru All Russian mathematical portal Mah-Ne.Ru All Russian mahemaical poral Aleksei S. Rodin, On he srucure of singular se of a piecewise smooh minimax soluion of Hamilon-Jacobi-Bellman equaion, Ural Mah. J., 2016, Volume 2, Issue 1, 58 68

More information

Heat kernel and Harnack inequality on Riemannian manifolds

Heat kernel and Harnack inequality on Riemannian manifolds Hea kernel and Harnack inequaliy on Riemannian manifolds Alexander Grigor yan UHK 11/02/2014 onens 1 Laplace operaor and hea kernel 1 2 Uniform Faber-Krahn inequaliy 3 3 Gaussian upper bounds 4 4 ean-value

More information

On the Convergence Time of Dual Subgradient Methods for Strongly Convex Programs

On the Convergence Time of Dual Subgradient Methods for Strongly Convex Programs IEEE TRANSACTIONS ON AUTOMATIC CONTROL, 63(4), PP. 05, APRIL, 08 On he Convergence Time of Dual Subgradien Mehods for Srongly Convex Programs Hao Yu and Michael J. Neely Universiy of Souhern California

More information

IMPLICIT AND INVERSE FUNCTION THEOREMS PAUL SCHRIMPF 1 OCTOBER 25, 2013

IMPLICIT AND INVERSE FUNCTION THEOREMS PAUL SCHRIMPF 1 OCTOBER 25, 2013 IMPLICI AND INVERSE FUNCION HEOREMS PAUL SCHRIMPF 1 OCOBER 25, 213 UNIVERSIY OF BRIISH COLUMBIA ECONOMICS 526 We have exensively sudied how o solve sysems of linear equaions. We know how o check wheher

More information

Let us start with a two dimensional case. We consider a vector ( x,

Let us start with a two dimensional case. We consider a vector ( x, Roaion marices We consider now roaion marices in wo and hree dimensions. We sar wih wo dimensions since wo dimensions are easier han hree o undersand, and one dimension is a lile oo simple. However, our

More information

STABILITY OF NONLINEAR NEUTRAL DELAY DIFFERENTIAL EQUATIONS WITH VARIABLE DELAYS

STABILITY OF NONLINEAR NEUTRAL DELAY DIFFERENTIAL EQUATIONS WITH VARIABLE DELAYS Elecronic Journal of Differenial Equaions, Vol. 217 217, No. 118, pp. 1 14. ISSN: 172-6691. URL: hp://ejde.mah.xsae.edu or hp://ejde.mah.un.edu STABILITY OF NONLINEAR NEUTRAL DELAY DIFFERENTIAL EQUATIONS

More information

Lecture 2 October ε-approximation of 2-player zero-sum games

Lecture 2 October ε-approximation of 2-player zero-sum games Opimizaion II Winer 009/10 Lecurer: Khaled Elbassioni Lecure Ocober 19 1 ε-approximaion of -player zero-sum games In his lecure we give a randomized ficiious play algorihm for obaining an approximae soluion

More information

Feedback design using nonsmooth control Lyapunov functions: A numerical case study for the nonholonomic integrator

Feedback design using nonsmooth control Lyapunov functions: A numerical case study for the nonholonomic integrator Feedback design using nonsmooh conrol Lyapunov funcions: A numerical case sudy for he nonholonomic inegraor Philipp Braun,2, Lars Grüne, Chrisopher M. Kelle 2 Absrac Theoreical resuls for he exisence of

More information

arxiv: v1 [math.fa] 9 Dec 2018

arxiv: v1 [math.fa] 9 Dec 2018 AN INVERSE FUNCTION THEOREM CONVERSE arxiv:1812.03561v1 [mah.fa] 9 Dec 2018 JIMMIE LAWSON Absrac. We esablish he following converse of he well-known inverse funcion heorem. Le g : U V and f : V U be inverse

More information

SUFFICIENT CONDITIONS FOR EXISTENCE SOLUTION OF LINEAR TWO-POINT BOUNDARY PROBLEM IN MINIMIZATION OF QUADRATIC FUNCTIONAL

SUFFICIENT CONDITIONS FOR EXISTENCE SOLUTION OF LINEAR TWO-POINT BOUNDARY PROBLEM IN MINIMIZATION OF QUADRATIC FUNCTIONAL HE PUBLISHING HOUSE PROCEEDINGS OF HE ROMANIAN ACADEMY, Series A, OF HE ROMANIAN ACADEMY Volume, Number 4/200, pp 287 293 SUFFICIEN CONDIIONS FOR EXISENCE SOLUION OF LINEAR WO-POIN BOUNDARY PROBLEM IN

More information

BU Macro BU Macro Fall 2008, Lecture 4

BU Macro BU Macro Fall 2008, Lecture 4 Dynamic Programming BU Macro 2008 Lecure 4 1 Ouline 1. Cerainy opimizaion problem used o illusrae: a. Resricions on exogenous variables b. Value funcion c. Policy funcion d. The Bellman equaion and an

More information

Lecture 4: November 13

Lecture 4: November 13 Compuaional Learning Theory Fall Semeser, 2017/18 Lecure 4: November 13 Lecurer: Yishay Mansour Scribe: Guy Dolinsky, Yogev Bar-On, Yuval Lewi 4.1 Fenchel-Conjugae 4.1.1 Moivaion Unil his lecure we saw

More information

Bifurcation Analysis of a Stage-Structured Prey-Predator System with Discrete and Continuous Delays

Bifurcation Analysis of a Stage-Structured Prey-Predator System with Discrete and Continuous Delays Applied Mahemaics 4 59-64 hp://dx.doi.org/.46/am..4744 Published Online July (hp://www.scirp.org/ournal/am) Bifurcaion Analysis of a Sage-Srucured Prey-Predaor Sysem wih Discree and Coninuous Delays Shunyi

More information

Differential Harnack Estimates for Parabolic Equations

Differential Harnack Estimates for Parabolic Equations Differenial Harnack Esimaes for Parabolic Equaions Xiaodong Cao and Zhou Zhang Absrac Le M,g be a soluion o he Ricci flow on a closed Riemannian manifold In his paper, we prove differenial Harnack inequaliies

More information

Decentralized Stochastic Control with Partial History Sharing: A Common Information Approach

Decentralized Stochastic Control with Partial History Sharing: A Common Information Approach 1 Decenralized Sochasic Conrol wih Parial Hisory Sharing: A Common Informaion Approach Ashuosh Nayyar, Adiya Mahajan and Demoshenis Tenekezis arxiv:1209.1695v1 [cs.sy] 8 Sep 2012 Absrac A general model

More information

SZG Macro 2011 Lecture 3: Dynamic Programming. SZG macro 2011 lecture 3 1

SZG Macro 2011 Lecture 3: Dynamic Programming. SZG macro 2011 lecture 3 1 SZG Macro 2011 Lecure 3: Dynamic Programming SZG macro 2011 lecure 3 1 Background Our previous discussion of opimal consumpion over ime and of opimal capial accumulaion sugges sudying he general decision

More information

Chapter 3 Boundary Value Problem

Chapter 3 Boundary Value Problem Chaper 3 Boundary Value Problem A boundary value problem (BVP) is a problem, ypically an ODE or a PDE, which has values assigned on he physical boundary of he domain in which he problem is specified. Le

More information

dy dx = xey (a) y(0) = 2 (b) y(1) = 2.5 SOLUTION: See next page

dy dx = xey (a) y(0) = 2 (b) y(1) = 2.5 SOLUTION: See next page Assignmen 1 MATH 2270 SOLUTION Please wrie ou complee soluions for each of he following 6 problems (one more will sill be added). You may, of course, consul wih your classmaes, he exbook or oher resources,

More information

A Forward-Backward Splitting Method with Component-wise Lazy Evaluation for Online Structured Convex Optimization

A Forward-Backward Splitting Method with Component-wise Lazy Evaluation for Online Structured Convex Optimization A Forward-Backward Spliing Mehod wih Componen-wise Lazy Evaluaion for Online Srucured Convex Opimizaion Yukihiro Togari and Nobuo Yamashia March 28, 2016 Absrac: We consider large-scale opimizaion problems

More information

Optimality and complexity for constrained optimization problems with nonconvex regularization

Optimality and complexity for constrained optimization problems with nonconvex regularization Opimaliy and complexiy for consrained opimizaion problems wih nonconvex regularizaion Wei Bian Deparmen of Mahemaics, Harbin Insiue of Technology, Harbin, China, bianweilvse520@163.com Xiaojun Chen Deparmen

More information

Global Synchronization of Directed Networks with Fast Switching Topologies

Global Synchronization of Directed Networks with Fast Switching Topologies Commun. Theor. Phys. (Beijing, China) 52 (2009) pp. 1019 1924 c Chinese Physical Sociey and IOP Publishing Ld Vol. 52, No. 6, December 15, 2009 Global Synchronizaion of Direced Neworks wih Fas Swiching

More information

Section 3.5 Nonhomogeneous Equations; Method of Undetermined Coefficients

Section 3.5 Nonhomogeneous Equations; Method of Undetermined Coefficients Secion 3.5 Nonhomogeneous Equaions; Mehod of Undeermined Coefficiens Key Terms/Ideas: Linear Differenial operaor Nonlinear operaor Second order homogeneous DE Second order nonhomogeneous DE Soluion o homogeneous

More information

Sliding Mode Controller for Unstable Systems

Sliding Mode Controller for Unstable Systems S. SIVARAMAKRISHNAN e al., Sliding Mode Conroller for Unsable Sysems, Chem. Biochem. Eng. Q. 22 (1) 41 47 (28) 41 Sliding Mode Conroller for Unsable Sysems S. Sivaramakrishnan, A. K. Tangirala, and M.

More information

T L. t=1. Proof of Lemma 1. Using the marginal cost accounting in Equation(4) and standard arguments. t )+Π RB. t )+K 1(Q RB

T L. t=1. Proof of Lemma 1. Using the marginal cost accounting in Equation(4) and standard arguments. t )+Π RB. t )+K 1(Q RB Elecronic Companion EC.1. Proofs of Technical Lemmas and Theorems LEMMA 1. Le C(RB) be he oal cos incurred by he RB policy. Then we have, T L E[C(RB)] 3 E[Z RB ]. (EC.1) Proof of Lemma 1. Using he marginal

More information

Distributed Linear Supervisory Control

Distributed Linear Supervisory Control 3rd IEEE Conference on Decision and Conrol December -7, Los Angeles, California, USA Disribued Linear Supervisory Conrol Ali Khanafer, Tamer Başar, and Daniel Liberzon Absrac In his work, we propose a

More information

Finish reading Chapter 2 of Spivak, rereading earlier sections as necessary. handout and fill in some missing details!

Finish reading Chapter 2 of Spivak, rereading earlier sections as necessary. handout and fill in some missing details! MAT 257, Handou 6: Ocober 7-2, 20. I. Assignmen. Finish reading Chaper 2 of Spiva, rereading earlier secions as necessary. handou and fill in some missing deails! II. Higher derivaives. Also, read his

More information

APPLICATION OF CHEBYSHEV APPROXIMATION IN THE PROCESS OF VARIATIONAL ITERATION METHOD FOR SOLVING DIFFERENTIAL- ALGEBRAIC EQUATIONS

APPLICATION OF CHEBYSHEV APPROXIMATION IN THE PROCESS OF VARIATIONAL ITERATION METHOD FOR SOLVING DIFFERENTIAL- ALGEBRAIC EQUATIONS Mahemaical and Compuaional Applicaions, Vol., No. 4, pp. 99-978,. Associaion for Scienific Research APPLICATION OF CHEBYSHEV APPROXIMATION IN THE PROCESS OF VARIATIONAL ITERATION METHOD FOR SOLVING DIFFERENTIAL-

More information

Accelerated Distributed Nesterov Gradient Descent for Convex and Smooth Functions

Accelerated Distributed Nesterov Gradient Descent for Convex and Smooth Functions 07 IEEE 56h Annual Conference on Decision and Conrol (CDC) December -5, 07, Melbourne, Ausralia Acceleraed Disribued Neserov Gradien Descen for Convex and Smooh Funcions Guannan Qu, Na Li Absrac This paper

More information

Optimal rate of convergence of an ODE associated to the Fast Gradient Descent schemes for b>0

Optimal rate of convergence of an ODE associated to the Fast Gradient Descent schemes for b>0 Opimal rae of convergence of an ODE associaed o he Fas Gradien Descen schemes for b>0 Jf Aujol, Ch Dossal To cie his version: Jf Aujol, Ch Dossal. Opimal rae of convergence of an ODE associaed o he Fas

More information

18 Biological models with discrete time

18 Biological models with discrete time 8 Biological models wih discree ime The mos imporan applicaions, however, may be pedagogical. The elegan body of mahemaical heory peraining o linear sysems (Fourier analysis, orhogonal funcions, and so

More information

Notes on Kalman Filtering

Notes on Kalman Filtering Noes on Kalman Filering Brian Borchers and Rick Aser November 7, Inroducion Daa Assimilaion is he problem of merging model predicions wih acual measuremens of a sysem o produce an opimal esimae of he curren

More information

Approximation Algorithms for Unique Games via Orthogonal Separators

Approximation Algorithms for Unique Games via Orthogonal Separators Approximaion Algorihms for Unique Games via Orhogonal Separaors Lecure noes by Konsanin Makarychev. Lecure noes are based on he papers [CMM06a, CMM06b, LM4]. Unique Games In hese lecure noes, we define

More information

Guest Lectures for Dr. MacFarlane s EE3350 Part Deux

Guest Lectures for Dr. MacFarlane s EE3350 Part Deux Gues Lecures for Dr. MacFarlane s EE3350 Par Deux Michael Plane Mon., 08-30-2010 Wrie name in corner. Poin ou his is a review, so I will go faser. Remind hem o go lisen o online lecure abou geing an A

More information

Differential Equations

Differential Equations Mah 21 (Fall 29) Differenial Equaions Soluion #3 1. Find he paricular soluion of he following differenial equaion by variaion of parameer (a) y + y = csc (b) 2 y + y y = ln, > Soluion: (a) The corresponding

More information

Modal identification of structures from roving input data by means of maximum likelihood estimation of the state space model

Modal identification of structures from roving input data by means of maximum likelihood estimation of the state space model Modal idenificaion of srucures from roving inpu daa by means of maximum likelihood esimaion of he sae space model J. Cara, J. Juan, E. Alarcón Absrac The usual way o perform a forced vibraion es is o fix

More information

Novel robust stability criteria of neutral-type bidirectional associative memory neural networks

Novel robust stability criteria of neutral-type bidirectional associative memory neural networks www.ijcsi.org 0 Novel robus sabiliy crieria of neural-ype bidirecional associaive memory neural neworks Shu-Lian Zhang and Yu-Li Zhang School of Science Dalian Jiaoong Universiy Dalian 608 P. R. China

More information

Vanishing Viscosity Method. There are another instructive and perhaps more natural discontinuous solutions of the conservation law

Vanishing Viscosity Method. There are another instructive and perhaps more natural discontinuous solutions of the conservation law Vanishing Viscosiy Mehod. There are anoher insrucive and perhaps more naural disconinuous soluions of he conservaion law (1 u +(q(u x 0, he so called vanishing viscosiy mehod. This mehod consiss in viewing

More information

Západočeská Univerzita v Plzni, Czech Republic and Groupe ESIEE Paris, France

Západočeská Univerzita v Plzni, Czech Republic and Groupe ESIEE Paris, France ADAPTIVE SIGNAL PROCESSING USING MAXIMUM ENTROPY ON THE MEAN METHOD AND MONTE CARLO ANALYSIS Pavla Holejšovsá, Ing. *), Z. Peroua, Ing. **), J.-F. Bercher, Prof. Assis. ***) Západočesá Univerzia v Plzni,

More information

SPECTRAL EVOLUTION OF A ONE PARAMETER EXTENSION OF A REAL SYMMETRIC TOEPLITZ MATRIX* William F. Trench. SIAM J. Matrix Anal. Appl. 11 (1990),

SPECTRAL EVOLUTION OF A ONE PARAMETER EXTENSION OF A REAL SYMMETRIC TOEPLITZ MATRIX* William F. Trench. SIAM J. Matrix Anal. Appl. 11 (1990), SPECTRAL EVOLUTION OF A ONE PARAMETER EXTENSION OF A REAL SYMMETRIC TOEPLITZ MATRIX* William F Trench SIAM J Marix Anal Appl 11 (1990), 601-611 Absrac Le T n = ( i j ) n i,j=1 (n 3) be a real symmeric

More information

2. Nonlinear Conservation Law Equations

2. Nonlinear Conservation Law Equations . Nonlinear Conservaion Law Equaions One of he clear lessons learned over recen years in sudying nonlinear parial differenial equaions is ha i is generally no wise o ry o aack a general class of nonlinear

More information

Average consensus by gossip algorithms with quantized communication

Average consensus by gossip algorithms with quantized communication Average consensus by gossip algorihms wih quanized communicaion Paolo Frasca, Ruggero Carli, Fabio Fagnani, and Sandro Zampieri Absrac This work sudies how he randomized gossip algorihm can solve he average

More information

Two Coupled Oscillators / Normal Modes

Two Coupled Oscillators / Normal Modes Lecure 3 Phys 3750 Two Coupled Oscillaors / Normal Modes Overview and Moivaion: Today we ake a small, bu significan, sep owards wave moion. We will no ye observe waves, bu his sep is imporan in is own

More information