Distributed Fractional Packing and Maximum Weighted b-matching via Tail-Recursive Duality

Size: px
Start display at page:

Download "Distributed Fractional Packing and Maximum Weighted b-matching via Tail-Recursive Duality"

Transcription

1 Ditributed Fractional Packing and Maximum Weighted b-matching via Tail-Recurive Duality Chrito Koufogiannaki, Neal E. Young Department of Computer Science, Univerity of California, Riveride {ckou, Abtract. We preent efficient ditributed δ-approximation algorithm for fractional packing and maximum weighted b-matching in hypergraph, where δ i the maximum number of packing contraint in which a variable appear (for maximum weighted b-matching δ i the maximum edge degree for graph δ = 2). (a) For δ = 2 the algorithm run in O(log m) round in expectation and with high probability. (b) For general δ, the algorithm run in O(log 2 m) round in expectation and with high probability. 1 Background and reult Given a weight vector w IR m +, a coefficient matrix A IR n m + and a vector b IR n +, the fractional packing problem i to compute a vector x IR m + to maximize m j=1 w jx j and at the ame time meet all the contraint m j=1 A ijx j b i ( i = 1... n). We ue δ to denote the maximum number of packing contraint in which a variable appear, that i, δ = max j {i A ij 0}. In the centralized etting, fractional packing can be olved optimally in polynomial time uing linear programming. Alternatively, one can ue a fater approximation algorithm (i.e. [11]). maximum weighted b-matching on a (hyper)graph i the variant where each A ij {0, 1} and the olution x mut take integer value (without lo of generality each vertex capacity i alo integer). An intance i defined by a given hypergraph H(V, E) and b Z V + ; a olution i given by a vector x Z E + maximizing e E w ex e and meeting all the vertex capacity contraint e E(u) x e b u ( u V ), where E(u) i the et of edge incident to vertex u. For thi problem, n = V, m = E and δ i the maximum (hyper)edge degree (for graph δ = 2). maximum weighted b-matching i a cornertone optimization problem in graph theory and Computer Science. A a pecial cae it include the ordinary maximum weighted matching problem (b u = 1 for all u V ). In the centralized etting, maximum weighted b-matching on graph belong to the well-olved cla of integer linear program in the ene that it can be olved in polynomial time [5, 6, 19]. Moreover, getting a 2-approximate 1 olution for maximum weighted matching i relatively eay, ince the obviou greedy algorithm, which elect the heaviet edge that i not conflicting with already elected edge, give a 2-approximation. For hypergraph the problem i NP-hard, ince it generalize et packing, one of Karp 21 NP-complete problem [10]. Our reult. In thi work we preent efficient ditributed δ-approximation algorithm for the above problem. If the input i a maximum weighted b-matching intance, the algorithm produce integral olution. The method we ue i of particular interet in the ditributed etting, where it i the firt primal-dual extenion of a non-tandard local-ratio technique [13, 2]. For fractional packing where each variable appear in at mot two contraint (δ = 2), we how a ditributed 2-approximation algorithm running in O(log m) round in expectation and with high probability. Thi i the firt 2-approximation algorithm requiring only O(log m) round. Thi improve the approximation ratio over the previouly bet known algorithm [14]. (For a ummary of known reult ee Figure 1.) Partially upported by NSF award CNS , CCF Since it i a maximization problem it i alo referred to a a 1/2-approximation

2 problem approx. ratio running time where when O( ) O(1) [22] O(log 2 n) [23] O(m) [7] 2004 O(1)(> 2) O(log n) [14] 2006 max weighted matching on graph (4 + ε) O(ε 1 log ε 1 log n) [18] 2007 (2 + ε) O(log ε 1 log n) [17] 2008 (1 + ε) O(ε 4 log 2 n) [17] 2008 (1 + ε) O(ε 2 + ε 1 log(ε 1 n) log n) [20] O(log 2 n) [17, 20] (ε = 1) O(log n) here 2009 fractional packing with δ = 2 max weighted matching on hypergraph fractional packing with general δ O(1)(> 2) O(log m) [14] O(log m) here 2009 O(δ) > δ O(log m) [14] 2006 δ O(log 2 m) here 2009 O(1) > 12 O(log m) [14] 2006 δ O(log 2 m) here 2009 Fig. 1. Ditributed algorithm for fractional packing and maximum weighted matching. For fractional packing where each variable appear in at mot δ contraint, we give a ditributed δ-approximation algorithm running in O(log 2 m) round in expectation and with high probability, where m i the number of variable. For mall δ, thi improve over the bet previouly known contant factor approximation [14], but the running time i lower by a logarithmic-factor. For maximum weighted b-matching on graph we give a ditributed 2-approximation algorithm running in O(log n) round in expectation and with high probability. maximum weighted b-matching generalize the well tudied maximum weighted matching problem. For a 2-approximation, our algorithm i fater by at leat a logarithmic factor than any previou algorithm. Specifically, in O(log n) round, our algorithm give the bet known approximation ratio. The bet previouly known algorithm compute a (1 + ε)- approximation in O(ε 4 log 2 n) round [17] or in O(ε 2 + ε 1 log(ε 1 n) log n) round [20]. For a 2- approximation both thee algorithm need O(log 2 n) round. For maximum weighted b-matching on hypergraph with maximum hyperedge degree δ we give a ditributed δ-approximation algorithm running in O(log 2 m) round in expectation and with high probability, where m i the number of hyperedge. Our reult improve over the bet previouly known O(δ)-approximation ratio by [14], but it i lower by a logarithmic factor. Related work for Maximum Weighted Matching. There are everal work conidering ditributed maximum weighted matching on edge-weighted graph. Uehara and Chen preent a contant time O( )- approximation algorithm [22], where i the maximum vertex degree. Wattenhofer and Wattenhofer improve thi reult, howing a randomized 5-approximation algorithm taking O(log 2 n) round [23]. Hoepman how a determinitic 2-approximation algorithm taking O(m) round [7]. Lotker, Patt-Shamir and Roén give a randomized (4+ε)-approximation algorithm running in O(ε 1 log ε 1 log n) round [18]. Lotker, Patt-Shamir and Pettie improve thi reult to a randomized (2 + ε)-approximation algorithm taking O(log ε 1 log n) round [17]. Their algorithm ue a a black box any ditributed contant-factor approximation algorithm for maximum weighted matching which take O(log n) round (i.e. [18]). Moreover, they mention (without detail) that there i a ditributed (1+ε)-approximation algorithm taking O(ε 4 log 2 n) round, baed on the parallel algorithm by Hougardy and Vinkemeier [8]. Nieberg preent a (1 + ε)-approximation algorithm in O(ε 2 +ε 1 log(ε 1 n) log n) round [20]. The latter two reult give randomized 2-approximation algorithm for maximum weighted matching in O(log 2 n) round. Related work for Fractional Packing. Kuhn, Mocibroda and Wattenhofer how efficient ditributed approximation algorithm for fractional packing [14]. They firt how a (1+ε)-approximation algorithm for fractional packing with logarithmic meage ize, but the running time depend on the input coefficient. For unbounded meage ize they how a contant-factor approximation algorithm for fractional packing which take O(log m) round. If an integer olution i deired, then ditributed randomized rounding ([15])

3 can be ued. Thi give an O(δ)-approximation for maximum weighted b-matching on (hyper)graph with high probability in O(log m) round, where δ i the maximum hyperedge degree (for graph δ = 2). (The hidden contant factor in the big-o notation of the approximation ratio can be relative large compared to a mall δ, ay δ = 2). Lower bound. The bet lower bound known for ditributed packing and matching are given by Kuhn, Mocibroda and Wattenhofer [14]. They prove that to achieve a contant or even a poly-logarithmic approximation ratio for fractional maximum matching, any algorithm require at leat Ω( log n/ log log n) round and Ω(log / log log ), where i the maximum vertex degree. Other related work. For unweighted maximum matching on graph, Iraeli and Itai give a randomized ditributed 2-approximation algorithm running in O(log n) round [9]. Lotker, Patt-Shamir and Pettie improve thi reult giving a randomized (1 + ε)-approximation algorithm taking O(ε 3 log n) round [17]. Czygrinow, Hańćkowiak, and Szymańka how a determinitic 3/2-approximation algorithm which take O(log 4 n) round [4]. A (1 + ε)-approximation for maximum weighted matching on graph i in NC [8]. The ret of the paper i organized a follow. In Section 2 we decribe a non-tandard primal-dual technique to get a δ-approximation algorithm for fractional packing and maximum weighted b-matching. In Section 3 we preent the ditributed implementation for δ = 2. Then in Section 4 we how the ditributed δ-approximation algorithm for general δ. We conclude in Section 5. 2 Covering and packing Koufogiannaki and Young how equential and ditributed δ-approximation algorithm for general covering problem [13, 12], where δ i the maximum number of covering variable on which a covering contraint depend. A a pecial cae their algorithm compute δ-approximate olution for fractional covering problem of the form min{ n i=1 b iy i : n i=1 A ijy i w j ( j = 1..m), y IR n +}. The linear programming dual of uch a problem i the following fractional packing problem: max{ m j=1 w jx j : m j=1 A ijx j b i ( i = 1... n), x IR m + }. For packing, δ i the maximum number of packing contraint in which a packing variable appear, δ = max j {i A ij 0}. Here we extend the ditributed approximation algorithm for fractional covering by [12] to compute δ-approximate olution for fractional packing uing a non-tandard primal-dual approach. Notation. Let C j denote the j-th covering contraint ( n i=1 A ijy i w j ) and P i denote the i-th packing contraint ( m j=1 A ijx j b i ). Let Var(S) denote the et of (covering or packing) variable indexe that appear in (covering or packing) contraint S. Let Con(z) denote the et of (covering or packing) contraint indexe in which (covering or packing) variable z appear. Let N(x ) denote the et of packing variable that appear in the packing contraint in which x appear, that i, N(x ) = {x j j Var(P i ) for ome i Con(x )} = Var(Con(x )). Fractional Covering. Firt we give a brief decription of the δ-approximation algorithm for fractional covering by [13, 12] 2. The algorithm perform tep to cover non-yet-atified covering contraint. Let y t be the olution after the firt t tep have been performed. (Initially y 0 = 0.) Given y t, let wj t = w j n i=1 A ijyi t be the lack of C j after the firt t tep. (Initially w 0 = w.) The algorithm i given by Alg. 1. There may be covering contraint for which the algorithm never perform a tep becaue they are covered by tep done for other contraint with which they hare variable. Alo note that increaing y i for all i Var(C ), decreae the lack of all contraint which depend on y i. Our general approach. [13] how that the above algorithm i a δ-approximation for covering, but they don t how any reult for matching or other packing problem. Our general approach i to recat their analyi a a primal-dual analyi, howing that the algorithm (Alg. 1) implicitly compute a olution to the dual packing problem of interet here. To do thi we ue the tail-recurive approach implicit in previou local-ratio analye [3]. 2 The algorithm i equivalent to local-ratio when A {0, 1} n m and y {0, 1} n [1, 2]. See [13] for a more general algorithm and a dicuion on the relation between thi algorithm and local ratio.

4 greedy δ-approximation algorithm for fractional covering [13, 12] alg Initialize y 0 0, w 0 w, t While there exit an unatified covering contraint C do a tep for C : 3. Set t = t Let β w t 1 min i Var(C) b i/a i.... OPT cot to atify C given the current olution 5. For each i Var(C ): 6. Set yi t = y t 1 i + β /b i.... increae y i inverely proportional to it cot 7. For each j Con(y i) update wj t = w t 1 j A ijβ /b i.... new lack 8. Return y = y t. After the t-th tep of the algorithm, define the reidual covering problem to be min{ n i=1 b iy i : n i=1 A ijy i wj t ( j = 1..m), y IR n +} and the reidual packing problem to be it dual, max{ m j=1 wt j x j : m j=1 A ijx j b i ( i = 1... n), x IR m + }. The algorithm will compute δ-approximate primal and dual pair (x t, y T t ) for the reidual problem for each t. A hown in what follow, the algorithm increment the covering olution x in a forward way, and the packing olution y in a tail-recurive manner. Standard Primal-Dual approach doe not work. For even imple intance, generating a δ-approximate primal-dual pair for the above greedy algorithm require a non-tandard approach. For example, conider min{y 1 +y 2 +y 3 : y 1 +y 2 1, y 1 +y 3 5, y 1, y 2 0}. If the greedy algorithm (Alg. 1) doe the contraint in either order and chooe β maximally, it give a olution of cot 10. In the dual max{x x 13 : x 12 + x 13 1, x 12, x 13 0}, the only way to generate a olution of cot 5 i to et x 13 = 1 and x 12 = 0. A tandard primaldual approach would raie the dual variable for each covering contraint when that contraint i proceed (eentially allowing a dual olution to be generated in an online fahion, contraint by contraint). That can t work here. For example, if the contraint y 1 + y 2 1 i covered firt by etting y 1 = y 2 = 1, then the dual variable x 12 would be increaed, thu preventing x 13 from reaching 1. Intead, auming the tep to cover y 1 + y 2 1 i done firt, the algorithm hould not increae any packing variable until a olution to the reidual dual problem i computed. After thi tep the reidual primal problem i min{y 1 + y 2 + y 3 : y 1 + y 2 1, y 1 + y 3 4, y 1, y 2 0}, and the reidual dual problem i max{ x x 13 : x 12 + x 13 1, x 12, x 13 0}. Once a olution x to the reidual dual problem i computed (either recurively or a hown later in thi ection) then the dual variable x 12 for the current covering contraint hould be raied maximally, giving the dual olution x for the current problem. In detail, the reidual dual olution x i x 12 = 0 and x 13 = 1 and the cot of the reidual dual olution i 4. Then the variable x 12 i raied maximally to give x 12. However, ince x 13 = 1, x 12 cannot be increaed, thu x = x. Although neither dual coordinate i increaed at thi tep, the dual cot i increaed from 4 to 5, becaue the weight of x 13 i increaed from w 13 = 4 to w 13 = 5. (See Figure 2 in the appendix.) In what follow we preent thi formally. Fractional Packing. We how that the greedy algorithm for covering create an ordering of the covering contraint for which it perform tep, which we can then ue to raie the correponding packing variable. Let t j denote the time 3 at which a tep to cover C j wa performed. Let t j = 0 if no tep wa performed for C j. We define the relation C j C j on two covering contraint C j and C j which hare a variable and for which the algorithm performed tep to indicate that contraint C j wa done firt by the algorithm. Definition 1. Let C j C j if Var(C j ) Var(C j ) and 0 < t j < t j. Note that the relation i not defined for covering contraint for which a tep wa never performed by the algorithm. Then let D be the partially ordered et (poet) of all covering contraint for which the algorithm performed a tep, ordered according to. D i partially ordered becaue i not defined for covering contraint that do not hare a variable. In addition, ince for each covering contraint C j we have a correponding dual packing variable x j, abuing notation we write x j x j if C j C j. Therefore, D i alo a poet of packing variable. 3 In general by time we mean ome reaonable way to ditinguih in which order tep were performed to atify covering contraint. For now, the time at which a tep wa performed can be thought a the tep number (line 3 at Alg. 1). It will be lightly different in the ditributed etting.

5 Definition 2. A revere order of poet D i an order C j1, C j2,..., C jk (or equivalently x j1, x j2,..., x jk ) uch that for l > i either we have C jl C ji or the relation i not defined for contraint C ji and C jl (becaue they do not hare a variable). Then the following figure (Alg. 2) how the equential δ-approximation algorithm for fractional packing. greedy δ-approximation algorithm for fractional packing alg Run Alg. 1, recording the poet D. 2. Let T be the number of tep performed by Alg Initialize x T 0, t T.... note that t will be decreaing from T to 0 4. Let Π be ome revere order of D.... any revere order of D work, ee Lemma 1 5. For each variable x D in the order given by Π do: 6. Set x t 1 = x t. 7. Raie x t 1 until a packing contraint that depend on x t 1 m j=1 Aijxt 1 j ). 8. Set t = t Return x = x 0. i tight, that i, et x t 1 = max i Con(xj )(b i The algorithm imply conider the packing variable correponding to covering contraint that Alg. 1 did tep for, and raie each variable maximally without violating the packing contraint. The order in which the variable are conidered matter: the variable hould be conidered in the revere of the order in which tep were done for the correponding contraint, or an order which i equivalent (ee Lemma 1). (Thi flexibility i neceary for the ditributed etting.) The olution x i feaible at all time ince a packing variable i increaed only until a packing contraint get tight. Lemma 1. Alg. 2 return the ame olution x uing (at line 4) any revere order of D. Proof. Let Π = x j1, x j2,..., x jk and Π = x j1, x j2,..., x jk be two different revere order of D. Let x Π,1...m be the olution computed o far by Alg. 2 after raiing the firt m packing variable of order Π. We prove that x Π,1...k = x Π,1...k. Aume that Π and Π have the ame order for their firt q variable, that i j i = j i for all i q. Then, x Π,1...q = x Π,1...q. The firt variable in which the two order diagree i the (q + 1)-th one, that i, j q+1 j q+1. Let = j q+1. Then x hould appear in ome poition l in Π uch that q + 1 < l k. The value of x depend only on the value of variable in N(x ) at the time when x i et. We prove that for each x j N(x ) we have x Π,1...q j = x Π,1...l j, thu x Π,1...q each packing variable only once thi implie x Π,1...k = x Π,1...q = x Π,1...l. Moreover ince the algorithm conider = x Π,1...l = x Π,1...k. (a) For each x j N(x ) with x x j, the variable x j hould have already been et in the firt q tep, otherwie Π would not be a valid revere order of D. Moreover each packing variable can be increaed only once, o once it i et it maintain the ame value till the end. Thu, for each x j uch that x x e, we have x Π,1...q j = x Π,1...q j = x Π,1...l j. (b) For each x j N(x ) with x j x, j cannot be in the interval [j q+1,..., j l 1 ) of Π, otherwie Π would not be a valid revere order of D. Thu, for each x j uch that x j x, we have x Π,1...q j x Π,1...l j = 0. So in any cae, for each x j N(x ), we have x Π,1...q j = x Π,1...l j The lemma follow by induction on the number of edge. and thu x Π,1...q = x Π,1...l. = x Π,1...q j = The following lemma and weak duality prove that the olution x returned by Alg. 2 i δ-approximate. Lemma 2. For the olution y and x returned by Alg. 1 and Alg. 2 repectively, m j=1 w jx j 1/δ n i=1 b iy i. Proof. Lemma 1 how that any revere order of D produce the ame olution, o w.l.o.g. here we aume that the revere order Π ued by Alg. 2 i the revere of the order in which tep to atify covering contraint were performed by Alg. 1.

6 When Alg. 1 doe a tep to atify the covering contraint C (by increaing y i by β /b i for all i Var(C )), the cot of the covering olution i b iy i increae by at mot δβ, ince C depend on at mot δ variable ( Var(C ) δ). Thu the final cot of the cover y i at mot D δβ. Define Ψ t = j wt j xt j to be the cot of the packing xt. Recall that x T = 0 o Ψ T = 0, and that the final packing olution i given by vector x 0, o the the cot of the final packing olution i Ψ 0. To prove the theorem we have to how that Ψ 0 D β. We have that Ψ 0 = Ψ 0 Ψ T = T t=1 Ψ t 1 Ψ t o it i enough to how that Ψ t 1 Ψ t β where C i the covering contraint done at the t-th tep of Alg. 1. Then, Ψ t 1 Ψ t i j = w t 1 = w t 1 = β x t 1 β 1 b i = β w t 1 j x t 1 j wjx t t j (1) x t 1 x t j (w t 1 j w t j)x t 1 j (2) i Con(x ) j {Var(P j) } A i max + i Con(x ) b i m j=1 i Con(x ) j {Var(P j) } A ij β b i x t 1 j (3) A ij β b i x t 1 j (4) A ij x t 1 j (for i.t. contraint P i become tight after raiing x ) (5) In equation (2) we ue the fact that x t = 0 and x t 1 j = x t j for all j. For equation (3), we ue the fact that the reidual weight of packing variable in N(x ) are increaed. If x j > 0 for j, then x j wa increaed before x (x x j ) o at the current tep w t 1 j > wj t > 0, and wt 1 j wj t = i Con(x A ) ij β b i. For equation (4), by the definition of β we have w t 1 = β max Ai i Con(x) b i. In inequality (5) we keep only the term that appear in the contraint P i that get tight by raiing x. The lat equality hold becaue P i i tight, that i, m j=1 A ijx j = b i. The following lemma how that Alg. 2 return integral olution if the coefficient A ij are 0/1 and the b i are integer, thu giving a δ-approximation algorithm for maximum weighted b-matching. Lemma 3. If A {0, 1} n m and b Z n + then the returned packing olution x i integral, that i, x Z m +. Proof. Since all non-zero coefficient are 1, the packing contraint are of the form j Var(P x i) j b i ( i). We prove by induction that x Z m +. The bae cae i trivial ince the algorithm tart with a zero olution. Aume that at ome point we have x t Z m +. Let x D, be the next packing variable to be raied by the algorithm. We how that x t 1 Z + and thu the reulting olution remain integral. The algorithm et x t 1 = min i Con(x){b i m j=1 xt 1 j } = min i Con(x){b i m j=1 xt j } 0. By the induction hypothei, each x t j Z +, and ince b Z n +, then x t 1 i alo a non-negative integer. 3 Ditributed Fractional Packing with δ = Ditributed model for δ = 2 We aume the network in which the ditributed computation take place ha vertice for covering variable (packing contraint) and edge for covering contraint (packing variable). So, the network ha a node u i for every covering variable y i. An edge e j connect vertice u i and u i if y i and y i belong to the ame covering contraint C j, that i, there exit a contraint A ij y i + A i jy i w j (δ = 2 o there can be at mot 2 variable in each covering contraint). We aume the tandard ynchronou communication model, where in each round, node can exchange meage with neighbor, and perform ome local computation [21]. We alo aume no retriction on meage ize and local computation. (Note that a ynchronou model algorithm can be tranformed into an aynchronou algorithm with the ame time complexity [21].) (6)

7 3.2 Ditributed algorithm for δ = 2 Koufogiannaki and Young how a ditributed implementation of Alg. 1, for (fractional) covering with δ = 2 that run in O(log n) round in expectation and with high probability [12]. In thi ection we augment their algorithm to ditributively compute 2-approximate olution to the dual fractional packing problem without increaing the time complexity. The high level idea i imilar to that in the previou ection: run the ditributed algorithm for covering to get a partial order of the covering contraint for which tep were performed, then conider the correponding dual packing variable in ome revere order raiing them maximally. The challenge here i that the ditributed algorithm for covering can perform tep for many covering contraint in parallel. Moreover, each covering contraint, ha jut a local view of the ordering, that i, it only know it relative order among the covering contraint with which it hare variable. Ditributed Fractional Covering with δ = 2. Here i a hort decription of the ditributed 2-approximation algorithm for fractional covering (Alg. 5 in appendix from [12]). In each round, the algorithm doe tep on a large ubet of remaining edge (covering contraint), a follow. Each vertex (covering variable) randomly chooe to be a leaf or a root. A not-yet-atified edge e j = (u i, u r ) between a leaf u i and a root u r with b i /A ij b r /A rj i active for the round. Each leaf u i chooe a random tar edge (u i, u r ) from it active edge. Thee tar edge form tar rooted at root. Each root u r then perform tep (of Alg. 1) on it tar edge (in any order) until they are all atified. Note that in a round the algorithm perform tep in parallel for edge not belonging to the ame tar. For edge belonging to the ame tar, their root perform tep for ome of them one by one. There are edge for which the algorithm never perform tep becaue they are covered by tep done for adjacent edge. In the ditributed etting we define the time at which a tep to atify C j i done a a pair (t R j, ts j ), where t R j denote the round in which the tep wa performed and t S j denote that within the tar thi tep i the t S j -th one. Let tr j = 0 if no tep wa performed for C j. Overloading Definition 1, we redefine a follow. Definition 3. Let C j C j (or equivalently x j x j ) if Var(C j ) Var(C j ) (j and j are adjacent edge in the ditributed network) and ([0 < t R j < tr j ] or [tr j = tr j and t S j < ts j ]). The pair (t R j, ts j ) i enough to ditinguih which of two adjacent edge had a tep to atify it covering contraint performed firt. Adjacent edge can have their covering contraint done in the ame round only if they belong to the ame tar (they have a common root), thu they differ in t S j. Otherwie they are done in different round, o they differ in t R j. Thu the pair (tr j, ts j ) and relation define a partially ordered et D of all edge done by the ditributed algorithm for covering. Lemma 4. ([12]) Alg. 5 (for fractional covering with δ = 2) finihe in T = O(log m) round in expectation and with high probability. Simultaneouly, Alg. 5 et (t R j, ts j ) for each edge e j for which it perform a tep (0 < t R j T ), thu defining a poet of edge D, ordered by. Ditributed Fractional Packing with δ = 2. Alg. 3 implement Alg. 2 in a ditributed fahion. Firt, it run Alg. 1 uing the ditributed implementation by [12] (Alg. 5) and recording D. Meanwhile, a it dicover the partial order D, it begin the econd phae of Alg. 2, raiing each packing variable a oon a it can. Specifically it wait to et a given x j D until after it know that (a) x j i in D, (b) for each x j N(x j ) whether x j x j, and (c) each uch x j i et. In other word, (a) a tep ha been done for the covering contraint C j, (b) each adjacent covering contraint C j i atified and (c) for each adjacent C j for which a tep wa done after C j, the variable x j ha been et. Subject to thee contraint it et x j a oon a poible. Note that ome node will be executing the econd phae of the algorithm (packing) while ome other node are till executing the firt phae (covering). Thi i neceary becaue a given node cannot know when ditant node are done with the firt phae. All x j will be determined in 2T round by the following argument. After round T, D i determined. Then by a traightforward induction on t, within T + t round, every contraint C j for which a tep wa done at round T t of the firt phae, will have it variable x j et. Theorem 1. For fractional packing where each variable appear in at mot two contraint there i a ditributed 2-approximation algorithm running in O(log m) round in expectation and with high probability, where m i the number of packing variable.

8 Ditributed 2-approximation Fractional Packing with δ = 2 alg. 3 input: Graph G = (V, E) repreenting a fractional packing problem intance with δ = 2. output: Feaible x, 2-approximately minimizing w x. 1. Each edge e j E initialize x j Each edge e j E initialize done j fale.... thi indicate if x j ha been et to it final value 3. Until each edge e j ha et it variable x j (done j == true), perform a round: 4. Perform a round of Alg covering with δ = 2 augmented to compute (t R j,t S j ) 5. For each node u r that wa a root (in Alg. 5) at any previou round, conider locally at u r all tar S t r that were rooted by u r at any previou round t. For each tar S t r perform IncreaeStar(S t r). IncreaeStar(tar S t r): 6. For each edge e j S t r in decreaing order of t S j : 7. If IncreaePackingVar(e j) == UNDONE then BREAK (top the for loop). IncreaePackingVar(edge e j = (u i, u r)): 8. If e j or any of it adjacent edge ha a non-yet-atified covering contraint return UNDONE. 9. If t R j == 0 then: 10. Set x j = 0 and done j = true. 11. Return DONE. 12. If done j == fale for any edge e j uch that x j x j then return UNDONE. 13. Set x j = min { (b i j A ij x j )/A ij, (b r j A rj x j )/A rj } and donej = true. 14. Return DONE. Proof. By Lemma 4, Alg. 5 compute a covering olution y in T = O(log m) round in expectation and with high probability. At the ame time, the algorithm et (t R j, ts j ) for each edge e j for which it perform a tep to cover C j, and thu defining a poet D of edge. In the ditributed etting the algorithm doe not define a linear order becaue there can be edge with the ame (t R j, ts j ), that i, edge that are covered by tep done in parallel. However, ince thee edge mut be non-adjacent, we can till think that the algorithm give a linear order (a in the equential etting), where tie between edge with the ame (t R j, ts j ) are broken arbitrarily (without changing D). Similarly, we can analyze Alg. 3 a if it conider the packing variable in a revere order of D. Then, by Lemma 1 and Lemma 2 the returned olution x i 2-approximate. We prove that the x can be computed in at mot T extra round after the initial T round to compute y. Firt note that within a tar, even though it edge are ordered according to t S j they can all et their packing variable in a ingle round if none of them wait for ome adjacent edge packing variable that belong to a different tar. So in the ret of the proof we only conider the cae were edge are waiting for adjacent edge that belong to different tar. Note that 1 t R j T for each x j D. Then, at round T, each x j with t R j = T can be et in thi round becaue it doe not have to wait for any other packing variable to be et. At the next round, round T + 1, each x j with t R j = T 1 can be et; they are dependent only on variable x j with t R j = T which have been already et. In general, packing variable with tr j = t can be et once all adjacent x j with t R j t + 1 have been et. Thu by induction on t = 0, 1,... a contraint C j for which a tep wa done at round T t may have to wait until at mot round T + t until it packing variable x j i et. Therefore, the total number of round until olution x i computed i 2T = O(log m) in expectation and with high probability. The following theorem i a direct reult of Lemma 3 and Thm 1 and the fact that for thi problem m = O(n 2 ). Theorem 2. For maximum weighted b-matching on graph there i a ditributed 2-approximation algorithm running in O(log n) round in expectation and with high probability.

9 4 Ditributed Fractional Packing with general δ 4.1 Ditributed model for general δ Here we aume that the ditributed network ha a node v j for each covering contraint C j (packing variable x j ), with edge from v j to each node v j if C j and C j hare a covering variable y i 4. The total number of node in the network i m. Note that in thi model the role of node and edge i revered a compared to the model ued in Section 3. We aume the tandard ynchronou model with unbounded meage ize. 4.2 Ditributed algorithm Koufogiannaki and Young [12] how a ditributed δ-approximation algorithm for (fractional) covering problem with at mot δ variable per covering contraint that run in O(log 2 m) round in expectation and with high probability. Similar to the δ = 2 cae, here we ue thi algorithm to get a poet of packing variable which we then conider in a revere order, raiing them maximally. Ditributed covering with general δ. Here i a brief decription of the ditributed δ-approximation algorithm for (fractional) covering from [12]. To tart each phae, the algorithm find large independent ubet of covering contraint by running one phae of Linial and Sak (LS) decompoition algorithm, with any k uch that k Θ(ln m) 5 [16]. The LS algorithm, for a given k, take O(k) round and produce a random ubet R {v j j = 1... m} of the covering contraint, and for each covering contraint v j R a leader l(v j ) R, with the following propertie: Each v j R i within ditance k of it leader: ( v j R) d(v j, v l(j) ) k. Component do not hare covering variable (edge do not cro component): ( v j, v j R) v l(j) v l(j ) Var(v j ) Var(v j ) =. Each covering contraint node ha a chance to be in R: ( j = 1... m) Pr[v j R] 1/cm 1/k for ome c > 1. Next, each node v j R end it information (the contraint and it variable value) to it leader v l(j). Thi take O(k) round becaue v l(j) i at ditance O(k) from v j. Each leader then contruct (locally) the ubproblem induced by the covering contraint that contacted it and the variable of thoe contraint, with their current value. Uing thi local copy, the leader doe tep until all covering contraint that contacted it are atified. (Ditinct leader ubproblem don t hare covering variable, o they can proceed imultaneouly.) To end the phae, each leader u l return the updated variable information to the contraint that contacted v l. Each covering contraint node in R i atified in the phae. To extend the algorithm to compute a olution to the dual packing problem the idea i imilar to the δ = 2 cae, ubtituting the role of tar by component and the role of root by leader. With each tep done to atify the covering contraint C j, the algorithm record (t R j, ts j ), where tr j i the round and t S j i the within-the-component iteration in which the tep wa performed. Thi define a poet D of covering contraint for which it perform tep. Lemma 5. ([12]) The ditributed δ-approximation algorithm for fractional covering finihe in T = O(log 2 m) round in expectation and with high probability, where m i the number of covering contraint (packing variable). Simultaneouly, it et (t R j, ts j ) for each covering contraint C j for which it perform a tep (0 < t R j T ), thu defining a poet of covering contraint (packing variable) D, ordered by. Ditributed packing with general δ. (ketch) The algorithm (Alg. 4) i very imilar to the cae δ = 2. Firt it run the ditributed algorithm for covering, recording (t R j, ts j ) for each covering contraint C j for which it perform a tep. Meanwhile, a it dicover the partial order D, it begin computing the packing olution, raiing each packing variable a oon a it can. Specifically it wait to et a given x j D until after it know that (a) x j i in D, (b) for each x j N(x j ) whether x j x j, and (c) each uch x j i et. In other word, (a) a tep ha been done for the covering contraint C j, (b) each adjacent covering contraint 4 The computation can eaily be imulated on a network with node for covering variable or node for covering variable and covering contraint. 5 If node don t know a k Θ(ln m), a doubling technique can be ued a a work-around [12].

10 Ditributed δ-approximation Fractional Packing with general δ alg. 4 input: Graph G = (V, E) repreenting a fractional packing problem intance. output: Feaible x, δ-approximately minimizing w x. 1. Initialize x For each j = 1... m initialize done j fale.... thi indicate if x j ha been et to it final value 3. Until each x j ha been et (done j == true) do: 4. Perform a phae of the δ-approximation algorithm for covering by [12], recording (t R j, t S j ). 5. For each node v K that wa a leader at any previou phae, conider locally at v K all component that choe v K a a leader at any previou phae. For each uch component K r perform IncreaeComponent(K r). IncreaeComponent(component K r): 6. For each j K r in decreaing order of t S j : 7. If IncreaePackingVar(j) == UNDONE then BREAK (top the for loop). IncreaePackingVar(j): 8. If C j or any C j that hare covering variable with C j i not yet atified return UNDONE. 9. If t R j == 0 then: 10. Set x j = 0 and done j = true. 11. Return DONE. 12. If done j == fale for any x j uch that x j x j then return UNDONE. 13. Set x j = min i Con(xj ) ( (bi j A ij x j )/A ij) ) and done j = true. 14. Return DONE. C j i atified and (c) for each adjacent C j for which a tep wa done after C j, the variable x j ha been et. Subject to thee contraint it et x j a oon a poible. To do o, the algorithm conider all component that have been done by leader in previou round. For each component, the leader conider the component packing variable x j in order of decreaing t S j. When conidering x j it check if each x j with x j x j i et, and if ye, then x j can be et and the algorithm continue with the next component packing variable (in order of decreaing t S j ). Otherwie the algorithm cannot yet decide about the remaining component packing variable. Theorem 3. For fractional packing where each variable appear in at mot δ contraint there i a ditributed δ-approximation algorithm running in O(log 2 m) round in expectation and with high probability, where m i the number of packing variable. The proof i omitted becaue it i imilar to the proof of Thm 1; the δ-approximation ratio i given by Lemma 1 and Lemma 2, and the running time ue T = O(log 2 m) by Lemma 5. The following theorem i a direct reult of Lemma 3 and Thm 3. Theorem 4. For maximum weighted b-matching on hypergraph, there i a ditributed δ-approximation algorithm running in O(log 2 m) round in expectation and with high probability, where δ i the maximum hyperedge degree and m i the number of hyperedge. 5 Concluion We how a new non-tandard primal-dual method, which extend the (local-ratio related) algorithm for fractional covering by [13, 12] to compute approximate olution to the dual fractional packing problem without increaing the time complexity (even in the ditributed etting). Uing thi new technique, we how a ditributed 2-approximation algorithm for fractional packing where each packing variable appear in at mot 2 contraint and a ditributed 2-approximation algorithm for maximum weighted b-matching on graph, both running in a logarithmic number of round. We alo preent a ditributed δ-approximation algorithm for fractional packing where each variable appear in at mot δ contraint and a ditributed δ-approximation algorithm for maximum weighted b-matching on hypergraph, both running in O(log 2 m) round, where m i the number of packing variable and hyperedge repectively.

11 Reference 1. R. Bar-Yehuda. One for the price of two: A unified approach for approximating covering problem. Algorithmica, 27(2): , R. Bar-Yehuda, K. Bendel, A. Freund, and D. Rawitz. Local ratio: a unified framework for approximation algorithm. ACM Computing Survey, 36(4): , R. Bar-Yehuda and D. Rawitz. On the equivalence between the primal-dual chema and the local-ratio technique. SIAM Journal on Dicrete Mathematic, 19(3): , A. Czygrinow, M. Hańćkowiak, and E. Szymańka. A fat ditributed algorithm for approximating the maximum matching. In the twelfth European Symoium on Algorithm, page , J. Edmond. Path, tree, and flower. Canadian Journal of Mathematic, 17: , J. Edmond and E. L. Johnon. Matching: A well-olved cla of integer linear program. Combinatorial Structure and Their Application, page 89 92, J.H. Hoepman. Simple ditributed weighted matching. Arxiv preprint c.dc/ , S. Hougardy and D.E. Vinkemeier. Approximating weighted matching in parallel. Information Proceing Letter, 99(3): , A. Iraeli and A. Itai. A fat and imple randomized parallel algorithm for maximal matching. Information Proceing Letter, 22:77 80, R. M. Karp. Reducibility among combinatorial problem. Complexity of Computer Computation, R. E. Miller and J. W. Thatcher, Ed., The IBM Reearch Sympoia Serie, New York, NY: Plenum Pre:85 103, C. Koufogiannaki and N.E. Young. Beating implex for fractional packing and covering linear program. In the forty-eighth IEEE ympoium on Foundation of Computer Science, page , C. Koufogiannaki and N.E. Young. Ditributed and parallel algorithm for weighted vertex cover and other covering problem. In the twenty-eighth ACM ympoium on Principle of Ditributed Computing, C. Koufogiannaki and N.E. Young. Greedy -approximation algorithm for covering with arbitrary contraint and ubmodular cot. In the thirty-ixth International Colloquium on Automata, Language and Programming, LNCS 5555: , See alo F. Kuhn, T. Mocibroda, and R. Wattenhofer. The price of being near-ighted. In the eventeenth ACM-SIAM Sympoium On Dicrete Algorithm, page , F. Kuhn and R. Wattenhofer. Contant-time ditributed dominating et approximation. In the twetwenty-econd ACM ympoium on Principle Of Ditributed Computing, page 25 32, N. Linial and M. Sak. Low diameter graph decompoition. Combinatorica, 13(4): , Z. Lotker, B. Patt-Shamir, and S. Pettie. Improved ditributed approximate matching. In the twelfth ACM Sympoium on Parallelim in Algorithm and Architecture, page , Z. Lotker, B. Patt-Shamir, and A. Roén. Ditributed approximate matching. In the twenty-ixth ACM ympoium on Principle Of Ditributed Computing, page , M. Müller-Hannemann and A. Schwartz. Implementing weighted b-matching algorithm: Toward a flexible oftware deign. In the Workhop on Algorithm Engineering and Experimentation (ALENEX), page 18 36, T. Nieberg. Local, ditributed weighted matching on general and wirele topologie. In the fifth ACM Joint Workhop on the Foundation of Mobile Computing, DIALM-POMC, page 87 92, D. Peleg. Ditributed computing: a locality-enitive approach. Society for Indutrial and Applied Mathematic, R. Uehara and Z. Chen. Parallel approximation algorithm for maximum weighted matching in general graph. Information Proceing Letter, 76(1-2):13 17, M. Wattenhofer and R. Wattenhofer. Ditributed weighted matching. In the eighteenth international ympoium on Ditributed Computing, page , Appendix Alg. 5 how the ditributed 2-approximation algorithm for fractional covering with δ = 2.

12 Execution tart here. Follow the arrow Final dual olution: x 0, x 1 dual cot = 5 min chooe y y y y y.t. y y y y tep y 1, y 1 primal cot += 2 y, y, y 0 max x 5x.t. x x 1 x x x, x 0 raie x 12 (it remain 0) maximally x 0, x 1 dual cot += 1 dual cot increae by 1 becaue w increae by 1 and x = min y ' y ' y '.t. y ' y ' y ' y ' y ', y ', y ' 0 max - x ' 4 x '.t. x ' x ' 1 x ', x ' 0 x ' 0, x ' 1 dual cot += 4 chooe y ' y ' tep y ' 4, y ' primal cot += 8 raie x ' maximally 13 (it become 1) min y '' y '' y ''.t. y '' y '' y '' y '' y '', y '', y '' 0 max -5 x '' 4 x ''.t. x '' x '' 1 x '', x '' 0 Bae cae y '' 1 y '' 2 y '' 3 0 x '' 12 x '' 13 0 dual cot = 0 Final primal olution: y 5, y 1, y 4 primal cot = 10 Fig. 2. Example of the execution of our greedy primal-dual algorithm (auming contraint y 1 + y 2 1 i choen firt).

13 Ditributed 2-approximation Fractional Covering with δ = 2 ([12]) alg. 5 input: Graph G = (V, E) repreenting a fractional covering problem intance with δ = 2. output: Feaible y, 2-approximately minimizing b y. 1. Each node u i V initialize y i Each edge e j E initialize t R j 0 and t S j auxiliary variable for Alg Until there i a vertex with unatified incident edge, perform a round: 4. Each node u i, randomly and independently chooe to be a leaf or a root for the round, each with probability 1/2. 5. Each leaf-to-root edge e j = (u i, u r) with unmet covering contraint i active at the tart of the round if u i i a leaf, u r i a root and b i/a ij b r/a rj. Each leaf u i chooe, among it active edge, a random one for the round. Communicate that choice to the neighbor. The choen edge form independent tar rooted tree of depth 1 whoe leave are leaf node and whoe root are root node. 6. For each root node u r, do: (a) Let Sr t contain the tar edge haring variable y r (at thi round t). (b) Until there exit an unatified edge (covering contraint) e j = (u i, u r) Sr, t perform Step(y, e j). Step(y, e j = (u i, u r)): 7. Let β j (w j A ijy i A rjy r) min{b i/a ij, b r/a rj}. 8. Set y i = y i + β j/b i and y r = y r + β j/b r. 9. Set t R j to the number of round performed o far. 10. Set t S j to the number of tep performed by root u r o far at thi round.

Lecture 14: Minimum Spanning Tree I

Lecture 14: Minimum Spanning Tree I COMPSCI 0: Deign and Analyi of Algorithm October 4, 07 Lecture 4: Minimum Spanning Tree I Lecturer: Rong Ge Scribe: Fred Zhang Overview Thi lecture we finih our dicuion of the hortet path problem and introduce

More information

Minimum congestion spanning trees in bipartite and random graphs

Minimum congestion spanning trees in bipartite and random graphs Minimum congetion panning tree in bipartite and random graph M.I. Otrovkii Department of Mathematic and Computer Science St. John Univerity 8000 Utopia Parkway Queen, NY 11439, USA e-mail: otrovm@tjohn.edu

More information

Brief Announcement: Distributed 3/2-Approximation of the Diameter

Brief Announcement: Distributed 3/2-Approximation of the Diameter Brief Announcement: Ditributed /2-Approximation of the Diameter Preliminary verion of a brief announcement to appear at DISC 14 Stephan Holzer MIT holzer@mit.edu David Peleg Weizmann Intitute david.peleg@weizmann.ac.il

More information

Karen L. Collins. Wesleyan University. Middletown, CT and. Mark Hovey MIT. Cambridge, MA Abstract

Karen L. Collins. Wesleyan University. Middletown, CT and. Mark Hovey MIT. Cambridge, MA Abstract Mot Graph are Edge-Cordial Karen L. Collin Dept. of Mathematic Weleyan Univerity Middletown, CT 6457 and Mark Hovey Dept. of Mathematic MIT Cambridge, MA 239 Abtract We extend the definition of edge-cordial

More information

Routing Definition 4.1

Routing Definition 4.1 4 Routing So far, we have only looked at network without dealing with the iue of how to end information in them from one node to another The problem of ending information in a network i known a routing

More information

Universität Augsburg. Institut für Informatik. Approximating Optimal Visual Sensor Placement. E. Hörster, R. Lienhart.

Universität Augsburg. Institut für Informatik. Approximating Optimal Visual Sensor Placement. E. Hörster, R. Lienhart. Univerität Augburg à ÊÇÅÍÆ ËÀǼ Approximating Optimal Viual Senor Placement E. Hörter, R. Lienhart Report 2006-01 Januar 2006 Intitut für Informatik D-86135 Augburg Copyright c E. Hörter, R. Lienhart Intitut

More information

1 The secretary problem

1 The secretary problem Thi i new material: if you ee error, pleae email jtyu at tanford dot edu 1 The ecretary problem We will tart by analyzing the expected runtime of an algorithm, a you will be expected to do on your homework.

More information

xy-monotone path existence queries in a rectilinear environment

xy-monotone path existence queries in a rectilinear environment CCCG 2012, Charlottetown, P.E.I., Augut 8 10, 2012 xy-monotone path exitence querie in a rectilinear environment Gregory Bint Anil Mahehwari Michiel Smid Abtract Given a planar environment coniting of

More information

AN ALGORITHM FOR RESTRICTED NORMAL FORM TO SOLVE DUAL TYPE NON-CANONICAL LINEAR FRACTIONAL PROGRAMMING PROBLEM

AN ALGORITHM FOR RESTRICTED NORMAL FORM TO SOLVE DUAL TYPE NON-CANONICAL LINEAR FRACTIONAL PROGRAMMING PROBLEM RAC Univerity Journal, Vol IV, No, 7, pp 87-9 AN ALGORITHM FOR RESTRICTED NORMAL FORM TO SOLVE DUAL TYPE NON-CANONICAL LINEAR FRACTIONAL PROGRAMMING PROLEM Mozzem Hoain Department of Mathematic Ghior Govt

More information

An Intro to LP and the Simplex Algorithm. Primal Simplex

An Intro to LP and the Simplex Algorithm. Primal Simplex An Intro to LP and the Simplex Algorithm Primal Simplex Linear programming i contrained minimization of a linear objective over a olution pace defined by linear contraint: min cx Ax b l x u A i an m n

More information

Contents. shortest paths. Notation. Shortest path problem. Applications. Algorithms and Networks 2010/2011. In the entire course:

Contents. shortest paths. Notation. Shortest path problem. Applications. Algorithms and Networks 2010/2011. In the entire course: Content Shortet path Algorithm and Network 21/211 The hortet path problem: Statement Verion Application Algorithm (for ingle ource p problem) Reminder: relaxation, Dijktra, Variant of Dijktra, Bellman-Ford,

More information

A note on degenerate and spectrally degenerate graphs

A note on degenerate and spectrally degenerate graphs A note on degenerate and pectrally degenerate graph Noga Alon Abtract A graph G i called pectrally d-degenerate if the larget eigenvalue of each ubgraph of it with maximum degree D i at mot dd. We prove

More information

See chapter 8 in the textbook. Dr Muhammad Al Salamah, Industrial Engineering, KFUPM

See chapter 8 in the textbook. Dr Muhammad Al Salamah, Industrial Engineering, KFUPM Goal programming Objective of the topic: Indentify indutrial baed ituation where two or more objective function are required. Write a multi objective function model dla a goal LP Ue weighting um and preemptive

More information

New Structural Decomposition Techniques for Constraint Satisfaction Problems

New Structural Decomposition Techniques for Constraint Satisfaction Problems New Structural Decompoition Technique for Contraint Satifaction Problem Yaling Zheng and Berthe Y. Choueiry Contraint Sytem Laboratory Univerity of Nebraka-Lincoln Email: yzheng choueiry@ce.unl.edu Abtract.

More information

Delaunay Triangulation: Incremental Construction

Delaunay Triangulation: Incremental Construction Chapter 6 Delaunay Triangulation: Incremental Contruction In the lat lecture, we have learned about the Lawon ip algorithm that compute a Delaunay triangulation of a given n-point et P R 2 with O(n 2 )

More information

A SIMPLE IMPERATIVE LANGUAGE THE STORE FUNCTION NON-TERMINATING COMMANDS

A SIMPLE IMPERATIVE LANGUAGE THE STORE FUNCTION NON-TERMINATING COMMANDS A SIMPLE IMPERATIVE LANGUAGE Eventually we will preent the emantic of a full-blown language, with declaration, type and looping. However, there are many complication, o we will build up lowly. Our firt

More information

Topics. Lecture 37: Global Optimization. Issues. A Simple Example: Copy Propagation X := 3 B > 0 Y := 0 X := 4 Y := Z + W A := 2 * 3X

Topics. Lecture 37: Global Optimization. Issues. A Simple Example: Copy Propagation X := 3 B > 0 Y := 0 X := 4 Y := Z + W A := 2 * 3X Lecture 37: Global Optimization [Adapted from note by R. Bodik and G. Necula] Topic Global optimization refer to program optimization that encompa multiple baic block in a function. (I have ued the term

More information

Shortest Paths Problem. CS 362, Lecture 20. Today s Outline. Negative Weights

Shortest Paths Problem. CS 362, Lecture 20. Today s Outline. Negative Weights Shortet Path Problem CS 6, Lecture Jared Saia Univerity of New Mexico Another intereting problem for graph i that of finding hortet path Aume we are given a weighted directed graph G = (V, E) with two

More information

New Structural Decomposition Techniques for Constraint Satisfaction Problems

New Structural Decomposition Techniques for Constraint Satisfaction Problems 113 New Structural Decompoition Technique for Contraint Satifaction Problem Yaling Zheng and Berthe Y. Choueiry Contraint Sytem Laboratory, Univerity of Nebraka-Lincoln {yzheng,choueiry}@ce.unl.edu Abtract.

More information

DAROS: Distributed User-Server Assignment And Replication For Online Social Networking Applications

DAROS: Distributed User-Server Assignment And Replication For Online Social Networking Applications DAROS: Ditributed Uer-Server Aignment And Replication For Online Social Networking Application Thuan Duong-Ba School of EECS Oregon State Univerity Corvalli, OR 97330, USA Email: duongba@eec.oregontate.edu

More information

Advanced Encryption Standard and Modes of Operation

Advanced Encryption Standard and Modes of Operation Advanced Encryption Standard and Mode of Operation G. Bertoni L. Breveglieri Foundation of Cryptography - AES pp. 1 / 50 AES Advanced Encryption Standard (AES) i a ymmetric cryptographic algorithm AES

More information

Hassan Ghaziri AUB, OSB Beirut, Lebanon Key words Competitive self-organizing maps, Meta-heuristics, Vehicle routing problem,

Hassan Ghaziri AUB, OSB Beirut, Lebanon Key words Competitive self-organizing maps, Meta-heuristics, Vehicle routing problem, COMPETITIVE PROBABIISTIC SEF-ORGANIZING MAPS FOR ROUTING PROBEMS Haan Ghaziri AUB, OSB Beirut, ebanon ghaziri@aub.edu.lb Abtract In thi paper, we have applied the concept of the elf-organizing map (SOM)

More information

Stochastic Search and Graph Techniques for MCM Path Planning Christine D. Piatko, Christopher P. Diehl, Paul McNamee, Cheryl Resch and I-Jeng Wang

Stochastic Search and Graph Techniques for MCM Path Planning Christine D. Piatko, Christopher P. Diehl, Paul McNamee, Cheryl Resch and I-Jeng Wang Stochatic Search and Graph Technique for MCM Path Planning Chritine D. Piatko, Chritopher P. Diehl, Paul McNamee, Cheryl Rech and I-Jeng Wang The John Hopkin Univerity Applied Phyic Laboratory, Laurel,

More information

Today s Outline. CS 561, Lecture 23. Negative Weights. Shortest Paths Problem. The presence of a negative cycle might mean that there is

Today s Outline. CS 561, Lecture 23. Negative Weights. Shortest Paths Problem. The presence of a negative cycle might mean that there is Today Outline CS 56, Lecture Jared Saia Univerity of New Mexico The path that can be trodden i not the enduring and unchanging Path. The name that can be named i not the enduring and unchanging Name. -

More information

Lecture Outline. Global flow analysis. Global Optimization. Global constant propagation. Liveness analysis. Local Optimization. Global Optimization

Lecture Outline. Global flow analysis. Global Optimization. Global constant propagation. Liveness analysis. Local Optimization. Global Optimization Lecture Outline Global flow analyi Global Optimization Global contant propagation Livene analyi Adapted from Lecture by Prof. Alex Aiken and George Necula (UCB) CS781(Praad) L27OP 1 CS781(Praad) L27OP

More information

3D SMAP Algorithm. April 11, 2012

3D SMAP Algorithm. April 11, 2012 3D SMAP Algorithm April 11, 2012 Baed on the original SMAP paper [1]. Thi report extend the tructure of MSRF into 3D. The prior ditribution i modified to atify the MRF property. In addition, an iterative

More information

Optimal Gossip with Direct Addressing

Optimal Gossip with Direct Addressing Optimal Goip with Direct Addreing Bernhard Haeupler Microoft Reearch 1065 La Avenida, Mountain View Mountain View, CA 94043 haeupler@c.cmu.edu Dahlia Malkhi Microoft Reearch 1065 La Avenida, Mountain View

More information

Operational Semantics Class notes for a lecture given by Mooly Sagiv Tel Aviv University 24/5/2007 By Roy Ganor and Uri Juhasz

Operational Semantics Class notes for a lecture given by Mooly Sagiv Tel Aviv University 24/5/2007 By Roy Ganor and Uri Juhasz Operational emantic Page Operational emantic Cla note for a lecture given by Mooly agiv Tel Aviv Univerity 4/5/7 By Roy Ganor and Uri Juhaz Reference emantic with Application, H. Nielon and F. Nielon,

More information

MAT 155: Describing, Exploring, and Comparing Data Page 1 of NotesCh2-3.doc

MAT 155: Describing, Exploring, and Comparing Data Page 1 of NotesCh2-3.doc MAT 155: Decribing, Exploring, and Comparing Data Page 1 of 8 001-oteCh-3.doc ote for Chapter Summarizing and Graphing Data Chapter 3 Decribing, Exploring, and Comparing Data Frequency Ditribution, Graphic

More information

Algorithmic Discrete Mathematics 4. Exercise Sheet

Algorithmic Discrete Mathematics 4. Exercise Sheet Algorithmic Dicrete Mathematic. Exercie Sheet Department of Mathematic SS 0 PD Dr. Ulf Lorenz 0. and. May 0 Dipl.-Math. David Meffert Verion of May, 0 Groupwork Exercie G (Shortet path I) (a) Calculate

More information

arxiv: v1 [cs.ds] 27 Feb 2018

arxiv: v1 [cs.ds] 27 Feb 2018 Incremental Strong Connectivity and 2-Connectivity in Directed Graph Louka Georgiadi 1, Giueppe F. Italiano 2, and Niko Parotidi 2 arxiv:1802.10189v1 [c.ds] 27 Feb 2018 1 Univerity of Ioannina, Greece.

More information

Chapter S:II (continued)

Chapter S:II (continued) Chapter S:II (continued) II. Baic Search Algorithm Sytematic Search Graph Theory Baic State Space Search Depth-Firt Search Backtracking Breadth-Firt Search Uniform-Cot Search AND-OR Graph Baic Depth-Firt

More information

Drawing Lines in 2 Dimensions

Drawing Lines in 2 Dimensions Drawing Line in 2 Dimenion Drawing a traight line (or an arc) between two end point when one i limited to dicrete pixel require a bit of thought. Conider the following line uperimpoed on a 2 dimenional

More information

A Practical Model for Minimizing Waiting Time in a Transit Network

A Practical Model for Minimizing Waiting Time in a Transit Network A Practical Model for Minimizing Waiting Time in a Tranit Network Leila Dianat, MASc, Department of Civil Engineering, Sharif Univerity of Technology, Tehran, Iran Youef Shafahi, Ph.D. Aociate Profeor,

More information

Generic Traverse. CS 362, Lecture 19. DFS and BFS. Today s Outline

Generic Traverse. CS 362, Lecture 19. DFS and BFS. Today s Outline Generic Travere CS 62, Lecture 9 Jared Saia Univerity of New Mexico Travere(){ put (nil,) in bag; while (the bag i not empty){ take ome edge (p,v) from the bag if (v i unmarked) mark v; parent(v) = p;

More information

Cutting Stock by Iterated Matching. Andreas Fritsch, Oliver Vornberger. University of Osnabruck. D Osnabruck.

Cutting Stock by Iterated Matching. Andreas Fritsch, Oliver Vornberger. University of Osnabruck. D Osnabruck. Cutting Stock by Iterated Matching Andrea Fritch, Oliver Vornberger Univerity of Onabruck Dept of Math/Computer Science D-4909 Onabruck andy@informatikuni-onabrueckde Abtract The combinatorial optimization

More information

Markov Random Fields in Image Segmentation

Markov Random Fields in Image Segmentation Preented at SSIP 2011, Szeged, Hungary Markov Random Field in Image Segmentation Zoltan Kato Image Proceing & Computer Graphic Dept. Univerity of Szeged Hungary Zoltan Kato: Markov Random Field in Image

More information

Maneuverable Relays to Improve Energy Efficiency in Sensor Networks

Maneuverable Relays to Improve Energy Efficiency in Sensor Networks Maneuverable Relay to Improve Energy Efficiency in Senor Network Stephan Eidenbenz, Luka Kroc, Jame P. Smith CCS-5, MS M997; Lo Alamo National Laboratory; Lo Alamo, NM 87545. Email: {eidenben, kroc, jpmith}@lanl.gov

More information

Shortest Path Routing in Arbitrary Networks

Shortest Path Routing in Arbitrary Networks Journal of Algorithm, Vol 31(1), 1999 Shortet Path Routing in Arbitrary Network Friedhelm Meyer auf der Heide and Berthold Vöcking Department of Mathematic and Computer Science and Heinz Nixdorf Intitute,

More information

Key Terms - MinMin, MaxMin, Sufferage, Task Scheduling, Standard Deviation, Load Balancing.

Key Terms - MinMin, MaxMin, Sufferage, Task Scheduling, Standard Deviation, Load Balancing. Volume 3, Iue 11, November 2013 ISSN: 2277 128X International Journal of Advanced Reearch in Computer Science and Software Engineering Reearch Paper Available online at: www.ijarce.com Tak Aignment in

More information

CSE 250B Assignment 4 Report

CSE 250B Assignment 4 Report CSE 250B Aignment 4 Report March 24, 2012 Yuncong Chen yuncong@c.ucd.edu Pengfei Chen pec008@ucd.edu Yang Liu yal060@c.ucd.edu Abtract In thi project, we implemented the recurive autoencoder (RAE) a decribed

More information

Edits in Xylia Validity Preserving Editing of XML Documents

Edits in Xylia Validity Preserving Editing of XML Documents dit in Xylia Validity Preerving diting of XML Document Pouria Shaker, Theodore S. Norvell, and Denni K. Peter Faculty of ngineering and Applied Science, Memorial Univerity of Newfoundland, St. John, NFLD,

More information

On successive packing approach to multidimensional (M-D) interleaving

On successive packing approach to multidimensional (M-D) interleaving On ucceive packing approach to multidimenional (M-D) interleaving Xi Min Zhang Yun Q. hi ankar Bau Abtract We propoe an interleaving cheme for multidimenional (M-D) interleaving. To achieved by uing a

More information

Performance of a Robust Filter-based Approach for Contour Detection in Wireless Sensor Networks

Performance of a Robust Filter-based Approach for Contour Detection in Wireless Sensor Networks Performance of a Robut Filter-baed Approach for Contour Detection in Wirele Senor Network Hadi Alati, William A. Armtrong, Jr., and Ai Naipuri Department of Electrical and Computer Engineering The Univerity

More information

Nearly Constant Approximation for Data Aggregation Scheduling in Wireless Sensor Networks

Nearly Constant Approximation for Data Aggregation Scheduling in Wireless Sensor Networks Nearly Contant Approximation for Data Aggregation Scheduling in Wirele Senor Network Scott C.-H. Huang, Peng-Jun Wan, Chinh T. Vu, Yinghu Li and France Yao Computer Science Department, City Univerity of

More information

The norm Package. November 15, Title Analysis of multivariate normal datasets with missing values

The norm Package. November 15, Title Analysis of multivariate normal datasets with missing values The norm Package November 15, 2003 Verion 1.0-9 Date 2002/05/06 Title Analyi of multivariate normal dataet with miing value Author Ported to R by Alvaro A. Novo . Original by Joeph

More information

Distributed Packet Processing Architecture with Reconfigurable Hardware Accelerators for 100Gbps Forwarding Performance on Virtualized Edge Router

Distributed Packet Processing Architecture with Reconfigurable Hardware Accelerators for 100Gbps Forwarding Performance on Virtualized Edge Router Ditributed Packet Proceing Architecture with Reconfigurable Hardware Accelerator for 100Gbp Forwarding Performance on Virtualized Edge Router Satohi Nihiyama, Hitohi Kaneko, and Ichiro Kudo Abtract To

More information

The Data Locality of Work Stealing

The Data Locality of Work Stealing The Data Locality of Work Stealing Umut A. Acar School of Computer Science Carnegie Mellon Univerity umut@c.cmu.edu Guy E. Blelloch School of Computer Science Carnegie Mellon Univerity guyb@c.cmu.edu Robert

More information

A Multi-objective Genetic Algorithm for Reliability Optimization Problem

A Multi-objective Genetic Algorithm for Reliability Optimization Problem International Journal of Performability Engineering, Vol. 5, No. 3, April 2009, pp. 227-234. RAMS Conultant Printed in India A Multi-objective Genetic Algorithm for Reliability Optimization Problem AMAR

More information

Representations and Transformations. Objectives

Representations and Transformations. Objectives Repreentation and Tranformation Objective Derive homogeneou coordinate tranformation matrice Introduce tandard tranformation - Rotation - Tranlation - Scaling - Shear Scalar, Point, Vector Three baic element

More information

Touring a Sequence of Polygons

Touring a Sequence of Polygons Touring a Sequence of Polygon Mohe Dror (1) Alon Efrat (1) Anna Lubiw (2) Joe Mitchell (3) (1) Univerity of Arizona (2) Univerity of Waterloo (3) Stony Brook Univerity Problem: Given a equence of k polygon

More information

Increasing Throughput and Reducing Delay in Wireless Sensor Networks Using Interference Alignment

Increasing Throughput and Reducing Delay in Wireless Sensor Networks Using Interference Alignment Int. J. Communication, Network and Sytem Science, 0, 5, 90-97 http://dx.doi.org/0.436/ijcn.0.50 Publihed Online February 0 (http://www.scirp.org/journal/ijcn) Increaing Throughput and Reducing Delay in

More information

SLA Adaptation for Service Overlay Networks

SLA Adaptation for Service Overlay Networks SLA Adaptation for Service Overlay Network Con Tran 1, Zbigniew Dziong 1, and Michal Pióro 2 1 Department of Electrical Engineering, École de Technologie Supérieure, Univerity of Quebec, Montréal, Canada

More information

A Boyer-Moore Approach for. Two-Dimensional Matching. Jorma Tarhio. University of California. Berkeley, CA Abstract

A Boyer-Moore Approach for. Two-Dimensional Matching. Jorma Tarhio. University of California. Berkeley, CA Abstract A Boyer-Moore Approach for Two-Dimenional Matching Jorma Tarhio Computer Science Diviion Univerity of California Berkeley, CA 94720 Abtract An imple ublinear algorithm i preented for two-dimenional tring

More information

CERIAS Tech Report EFFICIENT PARALLEL ALGORITHMS FOR PLANAR st-graphs. by Mikhail J. Atallah, Danny Z. Chen, and Ovidiu Daescu

CERIAS Tech Report EFFICIENT PARALLEL ALGORITHMS FOR PLANAR st-graphs. by Mikhail J. Atallah, Danny Z. Chen, and Ovidiu Daescu CERIAS Tech Report 2003-15 EFFICIENT PARALLEL ALGORITHMS FOR PLANAR t-graphs by Mikhail J. Atallah, Danny Z. Chen, and Ovidiu Daecu Center for Education and Reearch in Information Aurance and Security,

More information

arxiv: v3 [cs.cg] 1 Oct 2018

arxiv: v3 [cs.cg] 1 Oct 2018 Improved Time-Space Trade-off for Computing Voronoi Diagram Bahareh Banyaady Matia Korman Wolfgang Mulzer André van Renen Marcel Roeloffzen Paul Seiferth Yannik Stein arxiv:1708.00814v3 [c.cg] 1 Oct 2018

More information

Compressed Sensing Image Processing Based on Stagewise Orthogonal Matching Pursuit

Compressed Sensing Image Processing Based on Stagewise Orthogonal Matching Pursuit Senor & randucer, Vol. 8, Iue 0, October 204, pp. 34-40 Senor & randucer 204 by IFSA Publihing, S. L. http://www.enorportal.com Compreed Sening Image Proceing Baed on Stagewie Orthogonal Matching Puruit

More information

Distribution-based Microdata Anonymization

Distribution-based Microdata Anonymization Ditribution-baed Microdata Anonymization Nick Kouda niverity of Toronto kouda@c.toronto.edu Ting Yu North Carolina State niverity yu@cc.ncu.edu Diveh Srivatava AT&T Lab Reearch diveh@reearch.att.com Qing

More information

Planning of scooping position and approach path for loading operation by wheel loader

Planning of scooping position and approach path for loading operation by wheel loader 22 nd International Sympoium on Automation and Robotic in Contruction ISARC 25 - September 11-14, 25, Ferrara (Italy) 1 Planning of cooping poition and approach path for loading operation by wheel loader

More information

This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and

This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and Thi article appeared in a journal publihed by Elevier. The attached copy i furnihed to the author for internal non-commercial reearch and education ue, including for intruction at the author intitution

More information

Shortest Paths with Single-Point Visibility Constraint

Shortest Paths with Single-Point Visibility Constraint Shortet Path with Single-Point Viibility Contraint Ramtin Khoravi Mohammad Ghodi Department of Computer Engineering Sharif Univerity of Technology Abtract Thi paper tudie the problem of finding a hortet

More information

Motion Control (wheeled robots)

Motion Control (wheeled robots) 3 Motion Control (wheeled robot) Requirement for Motion Control Kinematic / dynamic model of the robot Model of the interaction between the wheel and the ground Definition of required motion -> peed control,

More information

Quadrilaterals. Learning Objectives. Pre-Activity

Quadrilaterals. Learning Objectives. Pre-Activity Section 3.4 Pre-Activity Preparation Quadrilateral Intereting geometric hape and pattern are all around u when we tart looking for them. Examine a row of fencing or the tiling deign at the wimming pool.

More information

Modeling of underwater vehicle s dynamics

Modeling of underwater vehicle s dynamics Proceeding of the 11th WEA International Conference on YTEM, Agio Nikolao, Crete Iland, Greece, July 23-25, 2007 44 Modeling of underwater vehicle dynamic ANDRZEJ ZAK Department of Radiolocation and Hydrolocation

More information

Exploring Simple Grid Polygons

Exploring Simple Grid Polygons Exploring Simple Grid Polygon Chritian Icking 1, Tom Kamphan 2, Rolf Klein 2, and Elmar Langetepe 2 1 Univerity of Hagen, Praktiche Informatik VI, 58084 Hagen, Germany. 2 Univerity of Bonn, Computer Science

More information

else end while End References

else end while End References 621-630. [RM89] [SK76] Roenfeld, A. and Melter, R. A., Digital geometry, The Mathematical Intelligencer, vol. 11, No. 3, 1989, pp. 69-72. Sklanky, J. and Kibler, D. F., A theory of nonuniformly digitized

More information

Touring a Sequence of Polygons

Touring a Sequence of Polygons Touring a Sequence of Polygon Mohe Dror (1) Alon Efrat (1) Anna Lubiw (2) Joe Mitchell (3) (1) Univerity of Arizona (2) Univerity of Waterloo (3) Stony Brook Univerity Problem: Given a equence of k polygon

More information

Lemma 1. A 3-connected maximal generalized outerplanar graph is a wheel.

Lemma 1. A 3-connected maximal generalized outerplanar graph is a wheel. 122 (1997) MATHEMATICA BOHEMICA No. 3, 225{230 A LINEAR ALGORITHM TO RECOGNIZE MAXIMAL GENERALIZED OUTERPLANAR GRAPHS Jo C cere, Almer a, Alberto M rquez, Sevilla (Received November 16, 1994, revied May

More information

A New Approach to Pipeline FFT Processor

A New Approach to Pipeline FFT Processor A ew Approach to Pipeline FFT Proceor Shouheng He and Mat Torkelon Department of Applied Electronic, Lund Univerity S- Lund, SWEDE email: he@tde.lth.e; torkel@tde.lth.e Abtract A new VLSI architecture

More information

ES205 Analysis and Design of Engineering Systems: Lab 1: An Introductory Tutorial: Getting Started with SIMULINK

ES205 Analysis and Design of Engineering Systems: Lab 1: An Introductory Tutorial: Getting Started with SIMULINK ES05 Analyi and Deign of Engineering Sytem: Lab : An Introductory Tutorial: Getting Started with SIMULINK What i SIMULINK? SIMULINK i a oftware package for modeling, imulating, and analyzing dynamic ytem.

More information

The Split Domination and Irredundant Number of a Graph

The Split Domination and Irredundant Number of a Graph The Split Domination and Irredundant Number of a Graph S. Delbin Prema 1, C. Jayaekaran 2 1 Department of Mathematic, RVS Technical Campu-Coimbatore, Coimbatore - 641402, Tamil Nadu, India 2 Department

More information

Optimal Multi-Robot Path Planning on Graphs: Complete Algorithms and Effective Heuristics

Optimal Multi-Robot Path Planning on Graphs: Complete Algorithms and Effective Heuristics Optimal Multi-Robot Path Planning on Graph: Complete Algorithm and Effective Heuritic Jingjin Yu Steven M. LaValle Abtract arxiv:507.0390v [c.ro] Jul 05 We tudy the problem of optimal multi-robot path

More information

Service and Network Management Interworking in Future Wireless Systems

Service and Network Management Interworking in Future Wireless Systems Service and Network Management Interworking in Future Wirele Sytem V. Tountopoulo V. Stavroulaki P. Demeticha N. Mitrou and M. Theologou National Technical Univerity of Athen Department of Electrical Engineering

More information

3D MODELLING WITH LINEAR APPROACHES USING GEOMETRIC PRIMITIVES

3D MODELLING WITH LINEAR APPROACHES USING GEOMETRIC PRIMITIVES MAKARA, TEKNOLOGI, VOL. 9, NO., APRIL 5: 3-35 3D MODELLING WITH LINEAR APPROACHES USING GEOMETRIC PRIMITIVES Mochammad Zulianyah Informatic Engineering, Faculty of Engineering, ARS International Univerity,

More information

Size Balanced Tree. Chen Qifeng (Farmer John) Zhongshan Memorial Middle School, Guangdong, China. December 29, 2006.

Size Balanced Tree. Chen Qifeng (Farmer John) Zhongshan Memorial Middle School, Guangdong, China. December 29, 2006. Size Balanced Tree Chen Qifeng (Farmer John) Zhonghan Memorial Middle School, Guangdong, China Email:44687@QQ.com December 9, 006 Abtract Thi paper preent a unique trategy for maintaining balance in dynamically

More information

How to Select Measurement Points in Access Point Localization

How to Select Measurement Points in Access Point Localization Proceeding of the International MultiConference of Engineer and Computer Scientit 205 Vol II, IMECS 205, March 8-20, 205, Hong Kong How to Select Meaurement Point in Acce Point Localization Xiaoling Yang,

More information

Laboratory Exercise 6

Laboratory Exercise 6 Laboratory Exercie 6 Adder, Subtractor, and Multiplier The purpoe of thi exercie i to examine arithmetic circuit that add, ubtract, and multiply number. Each type of circuit will be implemented in two

More information

Shortest-Path Routing in Arbitrary Networks

Shortest-Path Routing in Arbitrary Networks Ž. Journal of Algorithm 31, 105131 1999 Article ID jagm.1998.0980, available online at http:www.idealibrary.com on Shortet-Path Routing in Arbitrary Network Friedhelm Meyer auf der Heide and Berthold Vocking

More information

An Improved Implementation of Elliptic Curve Digital Signature by Using Sparse Elements

An Improved Implementation of Elliptic Curve Digital Signature by Using Sparse Elements The International Arab Journal of Information Technology, Vol. 1, No., July 004 0 An Improved Implementation of Elliptic Curve Digital Signature by Uing Spare Element Eam Al-Daoud Computer Science Department,

More information

A Linear Interpolation-Based Algorithm for Path Planning and Replanning on Girds *

A Linear Interpolation-Based Algorithm for Path Planning and Replanning on Girds * Advance in Linear Algebra & Matrix Theory, 2012, 2, 20-24 http://dx.doi.org/10.4236/alamt.2012.22003 Publihed Online June 2012 (http://www.scirp.org/journal/alamt) A Linear Interpolation-Baed Algorithm

More information

Computer Arithmetic Homework Solutions. 1 An adder for graphics. 2 Partitioned adder. 3 HDL implementation of a partitioned adder

Computer Arithmetic Homework Solutions. 1 An adder for graphics. 2 Partitioned adder. 3 HDL implementation of a partitioned adder Computer Arithmetic Homework 3 2016 2017 Solution 1 An adder for graphic In a normal ripple carry addition of two poitive number, the carry i the ignal for a reult exceeding the maximum. We ue thi ignal

More information

Optimal Peer-to-Peer Technique for Massive Content Distribution

Optimal Peer-to-Peer Technique for Massive Content Distribution 1 Optimal Peer-to-Peer Technique for Maive Content Ditribution Xiaoying Zheng, Chunglae Cho and Ye Xia Computer and Information Science and Engineering Department Univerity of Florida Email: {xiazheng,

More information

Localized Minimum Spanning Tree Based Multicast Routing with Energy-Efficient Guaranteed Delivery in Ad Hoc and Sensor Networks

Localized Minimum Spanning Tree Based Multicast Routing with Energy-Efficient Guaranteed Delivery in Ad Hoc and Sensor Networks Localized Minimum Spanning Tree Baed Multicat Routing with Energy-Efficient Guaranteed Delivery in Ad Hoc and Senor Network Hanne Frey Univerity of Paderborn D-3398 Paderborn hanne.frey@uni-paderborn.de

More information

Distance Optimal Formation Control on Graphs with a Tight Convergence Time Guarantee

Distance Optimal Formation Control on Graphs with a Tight Convergence Time Guarantee Ditance Optimal Formation Control on Graph with a Tight Convergence Time Guarantee Jingjin Yu Steven M. LaValle Abtract For the tak of moving a et of inditinguihable agent on a connected graph with unit

More information

Trainable Context Model for Multiscale Segmentation

Trainable Context Model for Multiscale Segmentation Trainable Context Model for Multicale Segmentation Hui Cheng and Charle A. Bouman School of Electrical and Computer Engineering Purdue Univerity Wet Lafayette, IN 47907-1285 {hui, bouman}@ ecn.purdue.edu

More information

Midterm 2 March 10, 2014 Name: NetID: # Total Score

Midterm 2 March 10, 2014 Name: NetID: # Total Score CS 3 : Algorithm and Model of Computation, Spring 0 Midterm March 0, 0 Name: NetID: # 3 Total Score Max 0 0 0 0 Grader Don t panic! Pleae print your name and your NetID in the boxe above. Thi i a cloed-book,

More information

A PROBABILISTIC NOTION OF CAMERA GEOMETRY: CALIBRATED VS. UNCALIBRATED

A PROBABILISTIC NOTION OF CAMERA GEOMETRY: CALIBRATED VS. UNCALIBRATED A PROBABILISTIC NOTION OF CAMERA GEOMETRY: CALIBRATED VS. UNCALIBRATED Jutin Domke and Yianni Aloimono Computational Viion Laboratory, Center for Automation Reearch Univerity of Maryland College Park,

More information

The Set Constraint/CFL Reachability Connection in Practice

The Set Constraint/CFL Reachability Connection in Practice The Set Contraint/CFL Reachability Connection in Practice John Kodumal EECS Department Univerity of California, Berkeley jkodumal@c.berkeley.edu Alex Aiken Computer Science Department Stanford Univerity

More information

mapping reult. Our experiment have revealed that for many popular tream application, uch a networking and multimedia application, the number of VC nee

mapping reult. Our experiment have revealed that for many popular tream application, uch a networking and multimedia application, the number of VC nee Reolving Deadlock for Pipelined Stream Application on Network-on-Chip Xiaohang Wang 1,2, Peng Liu 1 1 Department of Information Science and Electronic Engineering, Zheiang Univerity Hangzhou, Zheiang,

More information

A Fast Association Rule Algorithm Based On Bitmap and Granular Computing

A Fast Association Rule Algorithm Based On Bitmap and Granular Computing A Fat Aociation Rule Algorithm Baed On Bitmap and Granular Computing T.Y.Lin Xiaohua Hu Eric Louie Dept. of Computer Science College of Information Science IBM Almaden Reearch Center San Joe State Univerity

More information

Landmark Selection and Greedy Landmark-Descent Routing for Sensor Networks

Landmark Selection and Greedy Landmark-Descent Routing for Sensor Networks 1 Landmark Selection and Greedy Landmark-Decent Routing for Senor Network An Nguyen Nikola Miloavljević Qing Fang Jie Gao Leonida J. Guiba Department of Computer Science, Stanford Univerity. {anguyen,nikolam,guiba}@c.tanford.edu

More information

The Comparison of Neighbourhood Set and Degrees of an Interval Graph G Using an Algorithm

The Comparison of Neighbourhood Set and Degrees of an Interval Graph G Using an Algorithm The Comparion of Neighbourhood Set and Degree of an Interval Graph G Uing an Algorithm Dr.A.Sudhakaraiah, K.Narayana Aitant Profeor, Department of Mathematic, S.V. Univerity, Andhra Pradeh, India Reearch

More information

Floating Point CORDIC Based Power Operation

Floating Point CORDIC Based Power Operation Floating Point CORDIC Baed Power Operation Kazumi Malhan, Padmaja AVL Electrical and Computer Engineering Department School of Engineering and Computer Science Oakland Univerity, Rocheter, MI e-mail: kmalhan@oakland.edu,

More information

Keywords Cloud Computing, Service Level Agreements (SLA), CloudSim, Monitoring & Controlling SLA Agent, JADE

Keywords Cloud Computing, Service Level Agreements (SLA), CloudSim, Monitoring & Controlling SLA Agent, JADE Volume 5, Iue 8, Augut 2015 ISSN: 2277 128X International Journal of Advanced Reearch in Computer Science and Software Engineering Reearch Paper Available online at: www.ijarce.com Verification of Agent

More information

Refining SIRAP with a Dedicated Resource Ceiling for Self-Blocking

Refining SIRAP with a Dedicated Resource Ceiling for Self-Blocking Refining SIRAP with a Dedicated Reource Ceiling for Self-Blocking Mori Behnam, Thoma Nolte Mälardalen Real-Time Reearch Centre P.O. Box 883, SE-721 23 Väterå, Sweden {mori.behnam,thoma.nolte}@mdh.e ABSTRACT

More information

Robust Routing in Interdependent Networks

Robust Routing in Interdependent Networks Robut Routing in Interdependent Network Jianan Zhang and Eytan Modiano Laboratory for Information and Deciion Sytem, Maachuett Intitute of Technology Abtract We conider a model of two interdependent network,

More information

Evolution of Non-Deterministic Incremental Algorithms. Hugues Juille. Volen Center for Complex Systems. Brandeis University. Waltham, MA

Evolution of Non-Deterministic Incremental Algorithms. Hugues Juille. Volen Center for Complex Systems. Brandeis University. Waltham, MA Evolution of Non-Determinitic Incremental Algorithm a a New Approach for Search in State Space Hugue Juille Computer Science Department Volen Center for Complex Sytem Brandei Univerity Waltham, MA 02254-9110

More information

A Study of a Variable Compression Ratio and Displacement Mechanism Using Design of Experiments Methodology

A Study of a Variable Compression Ratio and Displacement Mechanism Using Design of Experiments Methodology A Study of a Variable Compreion Ratio and Diplacement Mechanim Uing Deign of Experiment Methodology Shugang Jiang, Michael H. Smith, Maanobu Takekohi Abtract Due to the ever increaing requirement for engine

More information

Keywords: Defect detection, linear phased array transducer, parameter optimization, phased array ultrasonic B-mode imaging testing.

Keywords: Defect detection, linear phased array transducer, parameter optimization, phased array ultrasonic B-mode imaging testing. Send Order for Reprint to reprint@benthamcience.ae 488 The Open Automation and Control Sytem Journal, 2014, 6, 488-492 Open Acce Parameter Optimization of Linear Phaed Array Tranducer for Defect Detection

More information

CS 467/567: Divide and Conquer on the PRAM

CS 467/567: Divide and Conquer on the PRAM CS 467/567: Divide and Conquer on the PRAM Stefan D. Bruda Winter 2017 BINARY SEARCH Problem: Given a equence S 1..n orted in nondecreaing order and a value x, find the ubcript k uch that S i x If n proceor

More information