Minimum Weight 2-Edge-Connected Spanning Subgraphs in Planar Graphs

Size: px
Start display at page:

Download "Minimum Weight 2-Edge-Connected Spanning Subgraphs in Planar Graphs"

Transcription

1 Minimum Weight 2-Edge-Connected Spanning Subgraphs in Planar Graphs André Berger 1 and Michelangelo Grigni 2 1 Technische Universität Berlin, Institut für Mathematik, Berlin, Germany. berger@math.tu-berlin.de 2 Department of Mathematics and Computer Science, Emory University, Atlanta GA 30322, USA. mic@mathcs.emory.edu Abstract. We present a linear time algorithm exactly solving the 2-edge connected spanning subgraph (2-ECSS) problem in a graph of bounded treewidth. Using this with Klein s diameter reduction technique [15], we find a linear time PTAS for the problem in unweighted planar graphs, and the first PTAS for the problem in weighted planar graphs. 1 Introduction A graph G = (V, E) is 2-edge-connected (2-EC) if G e is connected for every edge e E. Given G and an non-negative edge weight vector w R E +, the 2-edge-connected spanning subgraph (2-ECSS) problem is to find a 2-ECSS of G of minimum weight. We consider PTAS (polytime approximation scheme) results. For our purposes, a PTAS is an algorithm taking an instance (G, w) and ε > 0 as inputs, running in polynomial time for each fixed ε, and finding a solution with weight at most 1 + ε times optimal. Like metric TSP, the 2-ECSS problem is NP-hard even when restricted to unweighted planar graphs, and MAX-SNP-hard for bounded degree graphs [9]. In particular 2-ECSS is NP-hard to approximate within 1573/1572 on graphs of maximum degree 3 [7]. For unweighted graphs the best known polytime algorithm has approximation ratio 5/4 [12], and for weighted graphs the best known ratio is 2 [13, 14]. In Section 2 we present an algorithm solving the 2-ECSS problem exactly in linear time for weighted graphs of bounded treewidth. Graphs of bounded treewidth have particular algorithmic interest, since they adapt very well to dynamic programming techniques. Moreover, graphs of small treewidth are also related to fixed parameter algorithms and the bidimensionality theory [10].

2 A PTAS is known for 2-ECSS in complete geometric graphs of low dimension [9], and for unweighted planar graphs [8]. The latter PTAS runs in time n O(1/ε). We would prefer an efficient PTAS (EPTAS); that is, running in time f(ε) n c, where c is independent of ε. Recent examples of improving a PTAS to an EPTAS are the Euclidean TSP (n O(1/ε) [1] to O(ε O(ε) + n log n) [18]) and the weighted planar graph TSP (n O(1/ε2) [2] to 2 O(1/ε2) n [15]). In the same vein, we present a linear time EPTAS for the 2-ECSS problem in unweighted planar graphs in Section 3. For this we use the diameter reduction approach which Klein [15] applied to the metric TSP in weighted planar graphs. For the 2-ECSS problem in weighted planar graphs, the best previously known polytime approximation ratio is 2, achieved by the algorithm for general weighted graphs [13]. There is also an approximation scheme running in time n O(log n log(1/ε)/ε) [3], a quasi-polynomial. In Section 4 we present the first PTAS for this problem. Again we use ideas developed for the metric TSP; however, our situation is more complicated because it is insufficient to replace the input with a spanner (a subgraph which approximates the original metric). Instead, at each step of the dynamic program we must consider using a small number edges which are inessential to the metric. The bounding of this number (Lemma 10) is a key novelty in our approach. 2 An Exact Algorithm for Graphs of Bounded Treewidth In this section we design an exact algorithm for the 2-ECSS problem in weighted graphs of bounded treewidth. We first review basic notions for treewidth. Definition 1. Let G = (V, E) be a graph. A tree decomposition of G is a pair T D = ({X i : i I}, T = (I, F )), where {X i : i I} is a family of subsets of V and T is a tree, satisfying: (i) i I X i = V, (ii) for any edge uv E there exists an i I such that u, v X i, and (iii) for any v V, T [{i I : v X i }] is a subtree of T. The width of the tree decomposition is max i X i 1. The treewidth tw(g) of G is the minimum width of a tree decompositions of G. Finding the treewidth of a graph is NP-hard. However, for a given integer k, one can find a tree decomposition of a graph G of width at most k or decide that tw(g) > k in time O(f(k) n) [5]. In [19] it was 2

3 shown that f(k) = 2 O(k3), which makes the algorithm in [5] impractical for large values of k. For algorithmic purposes a especially simple tree decomposition is very useful. Definition 2. A tree decomposition ({X i : i I}, T ) of a graph G = (V, E) rooted at some r I is called nice, if the following conditions are satisfied: 1. Every node of the tree T has at most two children. 2. If a node i I has two children j and k, then X i = X j = X k (in this case i is called a JOIN NODE). 3. If a node i has one child j, then either (a) X i = X j + 1 and X j X i (in this case i is called an INSERT NODE), or (b) X i = X j 1 and X i X j (in this case i is called a FORGET NODE). Lemma 1 ([17, Lemma ]). Given a tree decomposition of a graph G = (V, E) that has width k and O(n) nodes, where n = V, we can find a nice tree decomposition of G that also has width k and O(n) nodes in time O(kn). From now on we assume G = (V, E) is a weighted graph with weight vector w R E +, and we let n = V. Furthermore, let ({X i : i I}, (I, F )) be a nice tree decomposition of G of width k and order I = O(n), rooted at r I. For any i I we define V i to be the set of vertices contained in X i and all X j, for which j is a descendant of i in T. Moreover, we let G i = G[V i ]. Note that V r = V and G r = G. We now consider the problem of finding a minimum weight 2-ECSS of G using this tree decomposition. For each i I we want to solve the following subproblem. Given a set of edges S E(X i ) and a graph H = (V H, E H ) with V H V = X i, find a subgraph G(i, S, H) of G i of minimum weight, such that E(G(i, S, H)) E(X i ) = and such that G(i, S, H) S H is 2-edge-connected. We let w(i, S, H) denote the weight of the graph G(i, S, H). In particular, the graph G(r, S, H ) S, where S minimizes w(r, S, H )+ w(s ) among all choices of S E(X r ) (H = (X r, )), is a minimum weight 2-ECSS of G. To determine w(r, S, H ) for S E(X r ), we will work up the tree decomposition from the leaves to the root solving the aforementioned subproblems. The graph H in a particular subproblem describes how the vertices in a set X i may be connected outside the graph 3

4 G i. We therefore refer to H as an external type of G i. We can think of H as a subgraph of G[(V \ V i ) X i ], that is part of a solution to the 2-ECSS problem on G, complementing the solution on G i. In general, there is an exponential number of subgraphs of G[(V \V i ) X i ] that can appear as an external type of the graph G i. However, as we will show, the number of external types H of the graph G i, which must be considered for a subproblem at the vertex i I of the tree decomposition, can be bounded only in terms of X i. Given a graph H = (V H, E H ), and a subset X V H, there is an algorithm [8] returning a simplified graph ext(h, X), which has the same 2-edge-connected properties when used as an external type for another graph, which shares only vertices in X with H. The detailed properties ext(h, X) are given in Lemma 2 and Corollary 1. Lemma 2. Let H = (V H, E H ) be a graph and X V H. Moreover, let H = ext(h, X). Then, if H is another graph with V H V H = X, then H H is 2-EC if and only if H H is 2-EC. Furthermore, V H 3 X. Lemma 2 implies that w(i, S, H) = G(i, S, ext(h, X i )), and that ext(h, X i ) does not have too many vertices. To make the computation efficient, we need that the number of possible external types that can appear during the algorithm is small. We let T ext (X) = {ext(h, X) : H = (V H, E H ), X V H } denote the set of all possible graphs ext(h, X) for a fixed set of vertices X. The following corollary gives us the bound on the number of possible external types. Corollary 1. For any set X we have T ext (X) = 2 O( X log X ). The algorithm for the 2-ECSS problem on G will compute the graphs G(i, S, H) and the values w(i, S, H) for every possible choice of i I, S E(X i ) and H T ext (X i ), working up from the leaves of T to the root r. The following four lemmata will help us prove the correctness of our algorithm. Lemma 3. Let i I be a leaf in T. Moreover let S E(X i ) and let H T ext (X i ). Then G(i, S, H) = and w(i, S, H) = 0. Proof. This is true by the definition of G(i, S, H). Lemma 4. Let i I be a join node with children j 1 and j 2, S E(X i ) and let H T ext (X i ). Then [ ] w(i, S, H) = min w(j 1, S, H 1 )+w(j 2, S, ext(h G(j 1, S, H 1 ), X j2 )). H 1 T ext(x j1 ) 4

5 Proof. The idea of the proof is as follows. If i is a join node with children j 1 and j 2, then the subgraph G(i, S, H) of G i achieving the minimum weight w(i, S, H) is a disjoint union of two subgraphs G 1 and G 2 of G j1 and G j2, respectively.we show that then G 1 and G 2 must be solutions to one of the previously computed subproblems at the nodes j 1 and j 2, related by the formula in the lemma. Let H 1 T ext (X j1 ) be the graph that achieves the minimum value W on the right hand side of the equality. Furthermore, let G 1 = G(j 1, S, H 1 ) and G 2 = G(j 2, S, ext(h G 1, X j2 )). Now G 2 S (H G 1 ) is 2-EC and thus G 1 G 2 is a feasible solution to the subproblem (i, S, H). Note that G 1 and G 2 are edge-disjoint, since X j1 = X j2 and G 1 and G 2 both do not contain any edges between vertices in X j1. Hence w(i, S, H) w(g 1 G 2 ) = w(g 1 ) + w(g 2 ) = W. To prove the reverse inequality, let H opt be a subgraph of G i with weight w(h opt ) = w(i, S, H) such that H opt S H is 2-EC. Let G 1 = H opt [V j1 \ X j1 ] and G 2 = H opt [V j2 \ X j2 ]. Also let H 1 = ext(h G 2, X j1 ). Since H opt = G 1 G 2, we have that G 1 S H 1 is 2-EC and thus w(j 1, S, H 1 ) w(g 1 ). Next, we let G 1 = G(j 1, S, H 1 ), i.e. G 1 S (H G 2) is 2-EC. Hence G 2 is a feasible solution to the subproblem (j 2, S, ext(h G 1, X j 2 )) and therefore w(j 2, S, ext(h G 1, X j 2 )) w(g 2 ). We finally get W w(j 1, S, H 1 )+w(j 2, S, ext(h G 1, X j 2 )) w(g 1 )+ w(g 2 ) = w(h opt ) = w(i, S, H). The proofs of the following two lemmas are similar to that of Lemma 4 and we omit them in this abstract. Lemma 5. Let i I be a insert node with child j, v V \ X j such that X i = X j {v}, S E[X i ] and let H T ext (X i ). Then w(i, S, H) = w(j, S E(X j ), ext(h (S E(v, X j )), X j )). Lemma 6. Let i I be a forget node with child j, v V such that X i = X j \ {v}, S E(X i ) and let H T ext (X i ). Then [ ] w(i, S, H) = min w(j, S S 1, H 1 = (V H {v}, E H )) + w(s 1 ). S 1 E(v,X j ) Using Lemmas 3-6, the values w(i, S, H) as well as the graphs G(i, S, H) can be computed by using the previously computed values at the children of i in the tree decomposition. We now analyze the running time of the algorithm. The tree decomposition has O(n) vertices and for each i I we have 2 k2 k O(k) subproblems, 5

6 since there are at most k 2 edges in E(X i ) and at most k O(k) external types for X i (Lemma 1). The most time consuming type of subproblem to solve is the one described in Lemma 4, which takes 2 O(k log k) time. Thus the total running time of the algorithm is O(2 O(k2) n). Theorem 1. The 2-ECSS problem in a weighted graph of with n vertices and a tree decomposition of width k can be solved exactly in O(2 O(k2) n) time. The constant in the running time of the algorithm can be improved using the following observation. Since G(i, S, H) is a graph with no edges in E(X i ), we have that G(i, S, H) = G(i, S, H) for all S that we can obtain from S by removing chords. The detailed arguments for this claim are identical to those in the proof of Lemma 2. Therefore, during the whole algorithm, we only have to consider edge sets S E(X i ) for some subproblem (i, S, H), which have no chords. We state the following lemma without proof. Lemma 7. Let G = (V, E) be a graph with no chords. Then E 2 V 2. For a given i I, therefore, we only need to consider at most (k 2 ) k = 2 O(k log k) different subsets S E(X i ) and thus we have the following improved running time. Theorem 2. Let a 2-EC graph G = (V, E) be given with a tree decomposition of width k, and let w R E + be a non-negative weight vector. Then, in time O(2 O(k log k) n), one can find a 2-ECSS of G of minimum weight. 3 A Linear-Time PTAS for Unweighted Planar Graphs We will now apply the methods from the previous section to obtain a linear time PTAS for the 2-ECSS problem in unweighted planar graphs. (More generally, the approach here would also work in weighted planar graphs whenever w(g) = O(OPT).) As in [8], we apply cycle contractions. Suppose G = (V, E) is a graph, C is a simple cycle in G, and G/C is the result of contracting C to a vertex (we remove any self-loops but leave parallel edges). First observe that G is 2-EC iff G/C is 2-EC. If H is a 2-ECSS in G, then its image (H C)/C is a 2-ECSS in G/C. Conversely, if H is a 2-ECSS in G/C, we can lift it back to a 2-ECSS H = H C in G (so H /C = H). Thus if we can find a near-optimal 2-ECSS H in G/C, then we also have a 6

7 near-optimal 2-ECSS H in G, with at most w(c) more additive error. Furthermore, if the vertex v representing C in G/C is a cut vertex, then the 2-ECSS problem in G/C decomposes into independent subproblems, one per block around v. The strategy to design the PTAS will be as follows. Given an unweighted planar 2-EC graph G, let OPT denote the size of a minimum 2-ECSS of G. We will first use methods from [4] and [15] to decompose the graph into parts of small treewidth by contracting a set of low weight cycles and committing only a small error to the solution. The 2-ECSS problems on the parts of small treewidth will be solved exactly using the algorithm from Section 2. So, let G = (V, E) be an embedded planar 2-EC graph. Our algorithm will make use of the following lemma, which can be derived from methods in [15] and [4]. Lemma 8. Let G = (V, E) be an embedded planar graph, w R E +, and k a positive integer. Then one can find a set of simple, edge-disjoint cycles C, such that w(e(c)) w(g)/k and such that the 2-ECSS problem on G/C can be decomposed into independent 2-ECSS problems on k-outerplanar graphs. Algorithm 1 gives a general overview of the PTAS for the unweighted planar graph 2-ECSS problem. The subproblems considered in step 1 are 2-EC planar graphs of outerplanarity k. Let H be such a graph. Since H is k-outerplanar, it can be triangulated in such a way that the outerplanarity of H is still k and such that every vertex of H has distance at most k to some vertex on the infinite face. Let H be the triangulated graph. By choosing an arbitrary v 0 on the infinite face and doing a breadth-first search on H, we can obtain a spanning tree T of H of radius at most k. This spanning tree will help us define a tree decomposition of H. We let I be the set of triangles in H and two triangles i, j I form an edge ij F if and only if they share an edge in E(H ) \ E(T ). Moreover, for a triangle i = {u, v, w}, we let X i be the set of all vertices on the three paths from u, v, and w, respectively, to the root v 0 of T. It is shown in [11], that T D = ({X i : i I}, (I, F )) is indeed a tree decomposition of H of width at most 3k, and that it can be found time O(kn). Since H is a subgraph of H, T D is also a tree decomposition of H. We can now apply the 2-ECSS algorithm for graphs of bounded treewidth to H and T D and we conclude the following theorem. Theorem 3. Given an unweighted 2-EC planar graph G = (V, E) and ε > 0, Algorithm 1 finds an (1 + ε)-approximate 2-ECSS of G in time 2 O( 1 ε log 1 ε ) n, where n = V. 7

8 Proof. We first prove the approximation guarantee of the algorithm. In any graph a 2-ECSS must use at least n edges, and therefore OP T n. Moreover, since G is planar, E 3n 6. The only error committed by the algorithm is adding the edges in E(C), for which by Lemma 8 we have E(C) E /k (3n 6)/ 3 ε ε n ε OP T. The running time is dominated by the application of the algorithm from Section 2, and is at most 2 O( 1 ε log 1 ε ) n by Theorem 2. Algorithm 1 A PTAS for the 2-ECSS problem on unweighted planar graphs Input: An embedded planar 2-EC graph G = (V, E), ε > 0. Output: A (1 + ε)-approximate 2-ECSS of G. 1: Apply Lemma 8 to the G with k = 3/ε and obtain a set of cycles C. 2: for all 2-EC components H of G/C do 3: Find a tree decomposition T D of H of width at most 3k as described above. 4: Transform T D to a nice tree decomposition using Lemma 1. 5: Apply the algorithm from Section 2 to H and T D and obtain S H. 6: end forthe solutions S H and E(C). Moreover, for planar graphs we can derive a better bound on the number of external types that can occur during the dynamic program (cf. [8]). We conclude this section with the following theorem. Theorem 4. Given an unweighted 2-EC planar graph G = (V, E) and ε > 0, one can find a (1+ε)-approximate 2-ECSS of G in time 2 O(1/ε) n, where n = V. 4 A PTAS for Weighted Planar Graphs In this section we will sketch the PTAS for the 2-ECSS problem in weighted planar graphs. The main idea of the algorithm is similar to Algorithm 1, the PTAS for unweighted planar graphs. If Algorithm 1 is applied to a weighted graph G with weights w R E +, then it will still run in linear time, however the approximation ratio will be 1 + O(ε w(g)/op T ), which may be unbounded for general weights. To avoid this problem we first construct a spanning subgraph of the input graph whose weight is at most constant times the optimal solution OP T. Similarly to the algorithm for unweighted planar graphs, we would like to decompose the spanner according to Lemma 8 into independent 8

9 2-ECSS problems of bounded outerplanarity, and solve these using a dynamic program similar to the algorithm from Section 2. However, we have to consider that certain edges that were deleted during the construction of the spanner have to appear in any (1 + ε)-approximate solution. In the following we will concentrate on the construction of the desired spanning subgraph. For a weighted graph G = (V, E), a weight vector w R E + and some s 1, a s-spanner of G is a spanning subgraph H of G, such that for any two vertices u, v V we have dist H s dist G (u, v), i.e. in H the distances between any two vertices are approximately the same as the respective distance in G. Of course G itself is a s-spanner of G for any s 1, but in many applications one seeks to find spanners with particular properties, such as few edges or small weight, so that instead of solving the given problem on G one can instead solve the problem on a much smaller graph. One example for this is the metric-tsp PTAS for weighted planar graphs, which starts with replacing the input graph by a s-spanner while effectively bounding w(g)/op T for the remainder of the algorithm. Suppose now that G is a weighted embedded planar graph and H is a subgraph. A chord e of H is an edge of G not in H. Note that H and e inherit embeddings from G. For each chord e we define w H (e) as the length of the shortest walk connecting the endpoints of e along the boundary of the face of H containing e. The spanner algorithm Augment(G, w, A, S) from [3] has as inputs a weighted planar graph G = (V, E), weights w R E +, a subgraph A of G and some s 1. Its properties are as follows: Lemma 9 ([3, Theorem 2]). Let H = Augment(G, w, A, s), where G, w, A and s are as above. Then H is a s-spanner of G containing A. In particular, w H (e) s w(e) for any e E \ E(H). Moreover, if A is connected, then w(h) (1 + 2/(s 1)) w(a). Given a face f in H, the chords of f are the edges of G embedded inside f, according to G s embedding. A face-edge e of f is an abstract edge connecting two vertices of f; unlike a chord, a face-edge is not necessarily an edge of G. (If vertices appear more than once on f, we must specify which appearances we want as the endpoints of e.) We say the face edge e crosses a chord c if c is a chord of the same face f and their endpoints are distinct vertex appearances on f, and they appear in cyclic ecec order around the boundary of f. Note that we may embed e inside f so e intersects only the crossed chords. The following lemma shows that the spanner has nice intersection properties with a near-optimal solution 9

10 to the 2-ECSS problem; it is key in bounding the size of the dynamic program described later. Lemma 10. Let G = (V, E) be a planar graph and w R E +. Let A be a 2- approximate 2-ECSS of G and H = Augment(G, w, A, 2). Furthermore, let S be an optimal 2-ECSS of G and D a positive integer. Then there exists a 2-ECSS S of G with the following properties: 1. w(s) (1 + 2/D) w(s ) 2. For any face f of H, and any face-edge e of f, the number of edges in S crossing e is at most 2D. Proof. Consider a face f of H and let C be the set of chords of f which are also edges of S. Moreover, let E f be the set of edges on the boundary of f. There are potentially very many edges in S crossing a particular face-edge of f. We will remove some edges from S and replace them by walks on E f as to maintain a 2-edge-connected subgraph S of G, while not increasing the weight of S by too much and ensuring the bound on the number of edges in S crossing any face-edge of f. Let Sf be the set of edges in S embedded inside the face f, i.e. these are edges which are not in H. The dual edges of all edges in Sf form a tree T, and we will orient each such dual edge in the following way. Any e = uv S f divides the boundary of f, i.e. E f, into two walks P 1 and P 2, and we will orient the dual edge e towards that part of the boundary of f which has smaller total weight, and we will call this walk P e. Using the definition above this means w H (e) = w(p e ). In particular we know by Lemma 9, w(p e ) 2 w(e). Moreover, any vertex of T has indegree at most 1, since if a vertex had indegree at least two, then two disjoint parts of E f would both have total weight greater than w(e f )/2, which is a contradiction. Therefore there must exist a root vertex r of T, such that all edges in T are oriented away from r. We will now modify S as follows. Let TD be the set of dual edges at distance D from r in T. For any e TD, we will delete from S all edges which are dual to the edges descendant to e in T. Moreover, we add P e to S. We call the resulting graph S. Note that P e P e = for any two edges e, e TD. It is clear that S remains a 2-ECSS of G. Further, the dual tree of edges in the newly constructed S which are embedded in f has now depth at most D. This shows that any face-edge of f is now crossed by at most 2D edges in S. It remains to show that adding the walks P e for all e TD increases the weight of S by at most 2/D w(sf ). If e TD and an edge ẽ 10

11 is on the path from e to r in T, we say ẽ e. For any fixed ẽ E(T ) of distance at most D from r we then have that e TD :ẽ e P e Pẽ. Thus e TD :ẽ e w(p e) w(pẽ). In turn, for any fixed e TD there are exactly D edges ẽ with ẽ e. Now the total weight by which we increase the weight of the solution S is w(p e ) 1/D e T D w(p e ) 1/D w(p e ) 1/D w(pẽ) e TD ẽ e ẽ T e TD :ẽ e ẽ T 2/D w(s f ). When doing this procedure for all faces of H, we obtain the desired 2-ECSS S of G. For each face f of H we added edges to S of weight at most 2/D w(sf ), so the total weight added to S for all faces of H is at most 2/D w(s ). Algorithm 2 gives a high-level description of the PTAS for the weighted 2-ECSS problem in planar graphs. Algorithm 3 is a dynamic program based on the ideas from Section 2 and is discussed more thoroughly below. In order for Algorithm 2 to be a PTAS for the 2-ECSS prob- Algorithm 2 A PTAS for the 2-ECSS problem on weighted planar graphs Input: A planar 2-EC graph G = (V, E), edge weights w R E +, ε > 0. Output: A (1 + ε)-approximate 2-ECSS of G. 1: Find a 2-approximate 2-ECSS A of G (e.g. using [13]). 2: H = Augment(G, 2, A) 3: Apply Lemma 8 to H with k = 24 and obtain a set of cycles C. ε 4: for all 2-EC components F = (V F, E F ) of H/C do 5: Apply Algorithm 3 to F with D = 4 and obtain SF. ε 6: end for the union of the solutions S F and E(C). lem on weighted planar graphs, we have to make use of Lemma 10, which guarantees that there exists a 2-ECSS of G of weight at most (1 + 2/D) OP T (1 + ε/2) OP T, which crosses each face edge of the spanner H at most 2D times. Algorithm 3 finds a minimum weight 2-ECSS of G which satisfies this condition for only a certain subset of all face-edges, and therefore it also has weight at most (1 + ε/2) OP T. Theorem 5. Given a planar 2-EC graph G = (V, E) and ε > 0, Algorithm 2 computes an (1 + ε)-approximate 2-ECSS of G in time n O(1/ε2), where n = V. Proof. Let us first prove the approximation guarantee of the algorithm. By Lemma 9 we have that w(h) (1 + 1/( 2 1)) w(a) 12 OP T, 11

12 since A is a 2-approximate 2-ECSS of G. Applying Lemma 8 in step 2 with k = 24 ε yields a collection of cycles of weight at most w(h)/k 12 OP T/ 24 ε ε 2 OP T. Therefore, by committing the cycles in C to our solution, we add an error of at most ε 2 OP T to the solution. For each 2-EC component F of H/C, we find a (1 + ε/2)-approximate 2-ECSS of F as argued above, and all the 2-ECSS problems on those 2-EC components are independent as described in Section 3. In step 2 we combine the solutions and their total weight is at most (1 + ε/2) OP T. Together with C they comprise a 2-ECSS of the original input graph G, whose total cost is at most (1 + ε) OP T. The overall running time is dominated by the application of Algorithm 3 in steps 2-2 and is bounded by n O(kD) = n O(1/ε2). We will now sketch the ideas for Algorithm 3. Let F = (V F, E F ) be a 2-EC component of H/C. We denote by C F the set of edges from the original input graph G which were deleted during the construction of the spanner H and are now missing from F. Note that F C F is still a planar graph. We can triangulate each face of F (as described in Section 3), to obtain a planar graph F and a spanning tree T of F of radius at most k. Moreover, let E = E(F ) \ E F. Some of the edges in T may be edges in E F and some may be edges in E, and we let T = E(T ) E. We will make use of the same tree decomposition of F as used in the PTAS for unweighted planar graphs. For a vertex u V F let P u denote the path from u to the root v r of T, and for a triangle i = uvw, we denote by P i = P u P v P w the union of the three paths from u, v and w to v r. Remember, we can obtain a tree decomposition T D = ({X i : i I}, T T D ) of F by letting I be the set of triangles in F and by letting X i = V (P i ) for each i I. Two triangles i, j I are connected by an edge in T T D if and only if they share an edge which is not in T. T D is a tree decomposition of F and therefore also of F of width at most 3k, and every i I has degree at most 3 in T T D. We apply Lemma 1 to T D and change it to a nice tree decomposition, then each X i (i I) is still the vertex set of the union of three paths P u, P v and P w, but now u, v and w do not necessarily lie on a triangle of F. We now define the subproblems that we want to solve in our dynamic program. For an edge e T, let C e C F be the set of edges in C F crossing e. Let now i I be a node of the tree decomposition. The set of edges in C F crossing any edge in P i is denoted by C i = C e. e P i Moreover, we let S E F (X i ) be a set of edges of F with both endpoints 12

13 in X i. Furthermore, let C C i be a set of crossing edges, such that C C e 2D for any e P i T. Finally, for any subproblem we also have to define an external type on the vertices in X i and V (C), the endpoints of edges in C. Thus let also a Ĥ T ext(x i V (C)) be given. The subproblem (i, S, C, Ĥ) is now the following. Find a subgraph F = F (i, S, C, Ĥ) of F [V i] C F of minimum weight w(i, S, C, Ĥ),, such that F H is 2-edge-connected, F E F (X i ) = S and such that F C i = C. If we let (S, C ) = argmin{w(r, S, C, Ĥ ) : S E F (X r ), C C i : C C e 2D for all e P r T }, then F (r, S, C, Ĥ ) is a minimum weight 2-ECSS of F C F which crosses each edge in P r T at most 2D times. Therefore the weight of that solution, i.e. w(r, S, C, Ĥ ), is at most the weight of a minimum weight 2-ECSS of F C F which crosses all edges in T at most 2D times. This is the property we needed in the proof of Theorem 5. Computing the values w(i, S, C, Ĥ) and the graphs F (i, S, C, Ĥ) can be done using Lemmas similar to Lemmas 3-6. The running time is n O(kD) ; we omit its analysis, except to remark it is dominated by considering all subsets C of O(kD) crossing edges. Algorithm 3 Generalized 2-ECSS for planar graphs Input: A planar graph F, a set of edges C F and D (as described above). Output: A 2-ECSS of F C F. 1: Triangulate F and obtain F of radius k. 2: Find a spanning tree T of F of radius k. 3: Find a tree decomposition T D of F of width at most 3k using T. 4: Refine T D and obtain T D, a nice tree decomposition with root r. 5: Solve the dynamic program on T D as described above. 6: (S, C ) = argmin{w(r, S, C, Ĥ ) : S E F (X r), C C i : C C e 2D for all e P r T }. F (r, S, C, Ĥ ) 5 Future Work The PTAS results of this paper are likely to carry over to the 2-VCSS problem (minimum weight 2-vertex-connected spanning subgraph), as happened before [8]. A question motivated by the TSP is whether we can find a PTAS for the Steiner or subset version of the 2-ECSS problem. That is, along with the graph G we are given a subset R of required vertices; we must find a 2-EC subgraph spanning at least the vertices in R. For the subset 13

14 version of planar metric TSP, Klein s subset spanner construction [16] yields an EPTAS. A similar subset spanner construction yields another EPTAS for the Steiner tree problem in planar graphs [6]. We can apply Klein s subset spanner construction to generalize our Lemma 9; however, the generalization of Lemma 10 remains a roadblock. 14

15 References 1. Sanjeev Arora. Polynomial time approximation schemes for Euclidean TSP and other geometric problems. In Proceedings of the Thirtyseventh Annual Symposium on Foundations of Computer Science (FOCS), pages IEEE Computer Society Press, Sanjeev Arora, Michelangelo Grigni, David Karger, Philip Klein, and Andrzej Woloszyn. A polynomial-time approximation scheme for weighted planar graph TSP. In Proceedings of the Ninth Annual ACM-SIAM Symposium on Discrete Algorithms (SODA), pages 33 41, New York, ACM Press. 3. André Berger, Artur Czumaj, Michelangelo Grigni, and Hairong Zhao. Approximation Schemes for Minimum 2-Connected Spanning Subgraphs in Weighted Planar Graphs. In Proceedings of the Thirteenth Annual European Symposium on Algorithms (ESA), volume 3669 of Lecture Notes in Computer Science, pages , Berlin, Springer. 4. Andre Berger, Michelangelo Grigni, and Hairong Zhao. A well-connected separator for planar graphs. Technical Report TR A, Emory University, Hans L. Bodlaender. A linear time algorithm for finding tree-decompositions of small treewidth. SIAM Journal on Computing, 25: , Glencora Borradaile, Claire Kenyon-Mathieu, and Philip Klein. A Polynomial- Time Approximation Scheme for Steiner Tree in Planar Graphs. In Proceedings of the Eighteenth Annual ACM-SIAM Symposium on Discrete Algorithms (SODA), New York, ACM Press. 7. Béla Csaba, Marek Karpinski, and Piotr Krysta. Approximability of dense and sparse instances of minimum 2-connectivity, TSP and path problems. In Proceedings of the 13th Annual ACM-SIAM Symposium On Discrete Mathematics (SODA), pages ACM Press, Artur Czumaj, Michelangelo Grigni, Papa Sissokho, and Hairong Zhao. Approximation schemes for minimum 2-edge-connected and biconnected subgraphs in planar graphs. In Proceedings of the Fifteenth Annual ACM-SIAM Symposium on Discrete Algorithms (SODA), pages , Artur Czumaj and Andrzej Lingas. On approximability of the minimumcost k- connected spanning subgraph problem. In Proceedings of the Tenth Annual ACM- SIAM Symposium on Discrete Algorithms (SODA), pages ACM Press, Erik D. Demaine and MohammadTaghi Hajiaghayi. Fast algorithms for hard graph problems: Bidimensionality, minors, and local treewidth. In Proceedings of the 12th International Symposium on Graph Drawing (GD 2004), volume 3383 of Lecture Notes in Computer Science, pages , David Eppstein. Subgraph isomorphism in planar graphs and related problems. Journal of Graph Algorithms and Applications, 3(3):1 27, Raja Jothi, Balaji Raghavachari, and Subramanian Varadarajan. A 5/4- approximation algorithm for minimum 2-edge-connectivity. In Proceedings of the Fourteenth Annual ACM-SIAM Symposium on Discrete Algorithms (SODA), pages , Samir Khuller. Approximation algorithms for finding highly connected subgraphs. In Dorit S. Hochbaum, editor, Approximation Algorithms for NP-hard Problems. PWS Publishing Company, Boston, Massachusetts, Samir Khuller and Uzi Vishkin. Biconnectivity approximations and graph carvings. Journal of the ACM, 41(2): ,

16 15. Philip N. Klein. A linear-time approximation scheme for planar weighted TSP. In Proceedings of the 46th Annual IEEE Symposium on Foundations of Computer Science (FOCS), pages IEEE Computer Society Press, Philip N. Klein. A subset spanner for planar graphs, with application to subset TSP. In Proceedings of the Thirty-Eighth Annual ACM Symposium on Theory of Computing (STOC), pages , Ton Kloks. Treewidth. Computations and Approximations. Number 842 in Lecture Notes in Computer Science. Springer-Verlag, Berlin, Satish B. Rao and Warren D. Smith. Approximating geometrical graphs via spanners and banyans. In Proceedings of the Thirtieth annual ACM Symposium on Theory of Computing (STOC 1998), pages , Hein Röhrig. Tree decomposition: A feasibility study. Master s thesis, Max-Planck- Institut für Informatik in Saarbrücken,

A Well-Connected Separator for Planar Graphs

A Well-Connected Separator for Planar Graphs A Well-Connected Separator for Planar Graphs André Berger Michelangelo Grigni Hairong Zhao Abstract We prove a separator theorem for planar graphs, which has been used to obtain polynomial time approximation

More information

Approximation Schemes for Minimum 2-Connected Spanning Subgraphs in Weighted Planar Graphs

Approximation Schemes for Minimum 2-Connected Spanning Subgraphs in Weighted Planar Graphs Approximation Schemes for Minimum 2-Connected Spanning Subgraphs in Weighted Planar Graphs André Berger 1, Artur Czumaj 2, Michelangelo Grigni 1, and Hairong Zhao 2 1 Department of Mathematics and Computer

More information

Approximation Schemes for Minimum 2-Edge-Connected and Biconnected Subgraphs in Planar Graphs

Approximation Schemes for Minimum 2-Edge-Connected and Biconnected Subgraphs in Planar Graphs Approximation Schemes for Minimum 2-Edge-Connected and Biconnected Subgraphs in Planar Graphs Artur Czumaj Michelangelo Grigni Papa Sissokho Hairong Zhao Abstract Given an undirected graph, finding either

More information

Approximation Schemes for Planar Graph Problems (1983, 1994; Baker)

Approximation Schemes for Planar Graph Problems (1983, 1994; Baker) Approximation Schemes for Planar Graph Problems (1983, 1994; Baker) Erik D. Demaine, MIT, theory.csail.mit.edu/ edemaine MohammadTaghi Hajiaghayi, MIT, www.mit.edu/ hajiagha INDEX TERMS: approximation

More information

Connecting face hitting sets in planar graphs

Connecting face hitting sets in planar graphs Connecting face hitting sets in planar graphs Pascal Schweitzer and Patrick Schweitzer Max-Planck-Institute for Computer Science Campus E1 4, D-66123 Saarbrücken, Germany pascal@mpi-inf.mpg.de University

More information

Fixed-Parameter Algorithms, IA166

Fixed-Parameter Algorithms, IA166 Fixed-Parameter Algorithms, IA166 Sebastian Ordyniak Faculty of Informatics Masaryk University Brno Spring Semester 2013 Introduction Outline 1 Introduction Algorithms on Locally Bounded Treewidth Layer

More information

arxiv: v1 [cs.ds] 11 Nov 2016

arxiv: v1 [cs.ds] 11 Nov 2016 A PTAS for Three-Edge Connectivity in Planar Graphs Glencora Borradaile and Baigong Zheng Oregon State University {glencora, zhengb}@eecs.oregonstate.edu arxiv:1611.03889v1 [cs.ds] 11 Nov 2016 Abstract

More information

PLANAR GRAPH BIPARTIZATION IN LINEAR TIME

PLANAR GRAPH BIPARTIZATION IN LINEAR TIME PLANAR GRAPH BIPARTIZATION IN LINEAR TIME SAMUEL FIORINI, NADIA HARDY, BRUCE REED, AND ADRIAN VETTA Abstract. For each constant k, we present a linear time algorithm that, given a planar graph G, either

More information

1 Introduction and Results

1 Introduction and Results On the Structure of Graphs with Large Minimum Bisection Cristina G. Fernandes 1,, Tina Janne Schmidt,, and Anusch Taraz, 1 Instituto de Matemática e Estatística, Universidade de São Paulo, Brazil, cris@ime.usp.br

More information

Strongly Connected Spanning Subgraph for Almost Symmetric Networks

Strongly Connected Spanning Subgraph for Almost Symmetric Networks CCC 2015, Kingston, Ontario, August 10 12, 2015 Strongly Connected Spanning Subgraph for Almost Symmetric Networks A. Karim Abu-Affash Paz Carmi Anat Parush Tzur Abstract In the strongly connected spanning

More information

6.889 Lecture 15: Traveling Salesman (TSP)

6.889 Lecture 15: Traveling Salesman (TSP) 6.889 Lecture 15: Traveling Salesman (TSP) Christian Sommer csom@mit.edu (figures by Philip Klein) November 2, 2011 Traveling Salesman Problem (TSP) given G = (V, E) find a tour visiting each 1 node v

More information

Treewidth and graph minors

Treewidth and graph minors Treewidth and graph minors Lectures 9 and 10, December 29, 2011, January 5, 2012 We shall touch upon the theory of Graph Minors by Robertson and Seymour. This theory gives a very general condition under

More information

Approximating Fault-Tolerant Steiner Subgraphs in Heterogeneous Wireless Networks

Approximating Fault-Tolerant Steiner Subgraphs in Heterogeneous Wireless Networks Approximating Fault-Tolerant Steiner Subgraphs in Heterogeneous Wireless Networks Ambreen Shahnaz and Thomas Erlebach Department of Computer Science University of Leicester University Road, Leicester LE1

More information

Grids containment and treewidth, PTASs

Grids containment and treewidth, PTASs Lecture 8 (30..0) Grids containment and treewidth, PTASs Author: Jakub Oćwieja Definition. A graph G is a t t grid if it consists of t vertices forming a square t t such that every two consecutive vertices

More information

Bottleneck Steiner Tree with Bounded Number of Steiner Vertices

Bottleneck Steiner Tree with Bounded Number of Steiner Vertices Bottleneck Steiner Tree with Bounded Number of Steiner Vertices A. Karim Abu-Affash Paz Carmi Matthew J. Katz June 18, 2011 Abstract Given a complete graph G = (V, E), where each vertex is labeled either

More information

Approximation Algorithms via Structural Results for Apex-Minor-Free Graphs

Approximation Algorithms via Structural Results for Apex-Minor-Free Graphs Approximation Algorithms via Structural Results for Apex-Minor-Free Graphs Erik D. Demaine 1, MohammadTaghi Hajiaghayi 2, and Ken-ichi Kawarabayashi 3 1 MIT Computer Science and Artificial Intelligence

More information

Embedded Width, A Variation of Treewidth for Planar Graphs

Embedded Width, A Variation of Treewidth for Planar Graphs Embedded Width, A Variation of Treewidth for Planar Graphs Robert Weber Advisor: Glencora Borradaile Abstract A commonly studied property of graphs is treewidth, a measure of how tree-like a graph is.

More information

Chordal deletion is fixed-parameter tractable

Chordal deletion is fixed-parameter tractable Chordal deletion is fixed-parameter tractable Dániel Marx Institut für Informatik, Humboldt-Universität zu Berlin, Unter den Linden 6, 10099 Berlin, Germany. dmarx@informatik.hu-berlin.de Abstract. It

More information

Approximating Node-Weighted Multicast Trees in Wireless Ad-Hoc Networks

Approximating Node-Weighted Multicast Trees in Wireless Ad-Hoc Networks Approximating Node-Weighted Multicast Trees in Wireless Ad-Hoc Networks Thomas Erlebach Department of Computer Science University of Leicester, UK te17@mcs.le.ac.uk Ambreen Shahnaz Department of Computer

More information

Jordan Curve. Salesman Tour

Jordan Curve. Salesman Tour A Polynomial-Time Approximation Scheme for Weighted Planar Graph TSP Sanjeev Arora Princeton U. Michelangelo Grigni y Emory U. Andrzej Woloszyn Emory U. David Karger z Philip Klein x Brown U. Abstract

More information

Parameterized coloring problems on chordal graphs

Parameterized coloring problems on chordal graphs Parameterized coloring problems on chordal graphs Dániel Marx Department of Computer Science and Information Theory, Budapest University of Technology and Economics Budapest, H-1521, Hungary dmarx@cs.bme.hu

More information

FEDOR V. FOMIN. Lectures on treewidth. The Parameterized Complexity Summer School 1-3 September 2017 Vienna, Austria

FEDOR V. FOMIN. Lectures on treewidth. The Parameterized Complexity Summer School 1-3 September 2017 Vienna, Austria FEDOR V. FOMIN Lectures on treewidth The Parameterized Complexity Summer School 1-3 September 2017 Vienna, Austria Why treewidth? Very general idea in science: large structures can be understood by breaking

More information

Subexponential algorithms for rectilinear Steiner tree and arborescence problems

Subexponential algorithms for rectilinear Steiner tree and arborescence problems Subexponential algorithms for rectilinear Steiner tree and arborescence problems Fedor V. Fomin Sudeshna Kolay Daniel Lokshtanov Fahad Panolan Saket Saurabh Abstract A rectilinear Steiner tree for a set

More information

Algorithms for Euclidean TSP

Algorithms for Euclidean TSP This week, paper [2] by Arora. See the slides for figures. See also http://www.cs.princeton.edu/~arora/pubs/arorageo.ps Algorithms for Introduction This lecture is about the polynomial time approximation

More information

The Square Root Phenomenon in Planar Graphs

The Square Root Phenomenon in Planar Graphs 1 The Square Root Phenomenon in Planar Graphs Survey and New Results Dániel Marx Institute for Computer Science and Control, Hungarian Academy of Sciences (MTA SZTAKI) Budapest, Hungary Satisfiability

More information

Planar graphs: multiple-source shortest paths, brick decomposition, and Steiner tree

Planar graphs: multiple-source shortest paths, brick decomposition, and Steiner tree Planar graphs: multiple-source shortest paths, brick decomposition, and Steiner tree Philip Klein joint work with Glencora Borradaile and Claire Mathieu Program: For fundamental optimization problems on

More information

Parameterized graph separation problems

Parameterized graph separation problems Parameterized graph separation problems Dániel Marx Department of Computer Science and Information Theory, Budapest University of Technology and Economics Budapest, H-1521, Hungary, dmarx@cs.bme.hu Abstract.

More information

Bicriteria Network Design via Iterative Rounding

Bicriteria Network Design via Iterative Rounding Bicriteria Network Design via Iterative Rounding Piotr Krysta Department of Computer Science, Dortmund University Baroper Str. 301, 44221 Dortmund, Germany piotr.krysta@cs.uni-dortmund.de Abstract. We

More information

Equivalence of Local Treewidth and Linear Local Treewidth and its Algorithmic Applications (extended abstract)

Equivalence of Local Treewidth and Linear Local Treewidth and its Algorithmic Applications (extended abstract) Equivalence of Local Treewidth and Linear Local Treewidth and its Algorithmic Applications (extended abstract) Erik D. Demaine and MohammadTaghi Hajiaghayi Laboratory for Computer Science, Massachusetts

More information

Randomized Algorithms 2017A - Lecture 10 Metric Embeddings into Random Trees

Randomized Algorithms 2017A - Lecture 10 Metric Embeddings into Random Trees Randomized Algorithms 2017A - Lecture 10 Metric Embeddings into Random Trees Lior Kamma 1 Introduction Embeddings and Distortion An embedding of a metric space (X, d X ) into a metric space (Y, d Y ) is

More information

Linear Kernel for Planar Connected Dominating Set

Linear Kernel for Planar Connected Dominating Set Linear Kernel for Planar Connected Dominating Set Daniel Lokshtanov Matthias Mnich Saket Saurabh Abstract We provide polynomial time data reduction rules for Connected Dominating Set on planar graphs and

More information

Subexponential Parameterized Odd Cycle Transversal on Planar Graphs

Subexponential Parameterized Odd Cycle Transversal on Planar Graphs Subexponential Parameterized Odd Cycle Transversal on Planar Graphs Daniel Lokshtanov 1, Saket Saurabh 2, and Magnus Wahlström 3 1 Department of Computer Science and Engineering, University of California,

More information

Theorem 2.9: nearest addition algorithm

Theorem 2.9: nearest addition algorithm There are severe limits on our ability to compute near-optimal tours It is NP-complete to decide whether a given undirected =(,)has a Hamiltonian cycle An approximation algorithm for the TSP can be used

More information

Notes on Binary Dumbbell Trees

Notes on Binary Dumbbell Trees Notes on Binary Dumbbell Trees Michiel Smid March 23, 2012 Abstract Dumbbell trees were introduced in [1]. A detailed description of non-binary dumbbell trees appears in Chapter 11 of [3]. These notes

More information

Arora s PTAS for the Euclidean TSP. R. Inkulu (Arora s PTAS for the Euclidean TSP) 1 / 23

Arora s PTAS for the Euclidean TSP. R. Inkulu   (Arora s PTAS for the Euclidean TSP) 1 / 23 Arora s PTAS for the Euclidean TSP R. Inkulu http://www.iitg.ac.in/rinkulu/ (Arora s PTAS for the Euclidean TSP) 1 / 23 Problem description Given a set P of points in Euclidean plane, find a tour of minimum

More information

Steiner tree in planar graphs: An O(n log n) approximation scheme with singly-exponential dependence on epsilon

Steiner tree in planar graphs: An O(n log n) approximation scheme with singly-exponential dependence on epsilon Steiner tree in planar graphs: An O(n log n) approximation scheme with singly-exponential dependence on epsilon Glencora Borradaile, Philip N. Klein, and Claire Mathieu Brown University, Providence RI

More information

On Structural Parameterizations of the Matching Cut Problem

On Structural Parameterizations of the Matching Cut Problem On Structural Parameterizations of the Matching Cut Problem N. R. Aravind, Subrahmanyam Kalyanasundaram, and Anjeneya Swami Kare Department of Computer Science and Engineering, IIT Hyderabad, Hyderabad,

More information

An O(log 2 k)-approximation Algorithm for the k-vertex Connected Spanning Subgraph Problem

An O(log 2 k)-approximation Algorithm for the k-vertex Connected Spanning Subgraph Problem An O(log 2 k)-approximation Algorithm for the k-vertex Connected Spanning Subgraph Problem ABSTRACT Jittat Fakcharoenphol Department of Computer Engineering Kasetsart University Bangkok, 0900, Thailand

More information

A note on Baker s algorithm

A note on Baker s algorithm A note on Baker s algorithm Iyad A. Kanj, Ljubomir Perković School of CTI, DePaul University, 243 S. Wabash Avenue, Chicago, IL 60604-2301. Abstract We present a corrected version of Baker s algorithm

More information

arxiv: v1 [cs.ds] 19 Feb 2014

arxiv: v1 [cs.ds] 19 Feb 2014 Turing Kernelization for Finding Long Paths and Cycles in Restricted Graph Classes Bart M. P. Jansen 1 University of Bergen, Norway. Bart.Jansen@ii.uib.no arxiv:1402.4718v1 [cs.ds] 19 Feb 2014 Abstract.

More information

Steiner Trees and Forests

Steiner Trees and Forests Massachusetts Institute of Technology Lecturer: Adriana Lopez 18.434: Seminar in Theoretical Computer Science March 7, 2006 Steiner Trees and Forests 1 Steiner Tree Problem Given an undirected graph G

More information

Solutions for the Exam 6 January 2014

Solutions for the Exam 6 January 2014 Mastermath and LNMB Course: Discrete Optimization Solutions for the Exam 6 January 2014 Utrecht University, Educatorium, 13:30 16:30 The examination lasts 3 hours. Grading will be done before January 20,

More information

Polynomial Time Approximation Schemes for the Euclidean Traveling Salesman Problem

Polynomial Time Approximation Schemes for the Euclidean Traveling Salesman Problem PROJECT FOR CS388G: ALGORITHMS: TECHNIQUES/THEORY (FALL 2015) Polynomial Time Approximation Schemes for the Euclidean Traveling Salesman Problem Shanshan Wu Vatsal Shah October 20, 2015 Abstract In this

More information

Planarity allowing few error vertices in linear time

Planarity allowing few error vertices in linear time Planarity allowing few error vertices in linear time Ken-ichi Kawarabayashi National Institute of Informatics 2-1-2, Hitotsubashi, Chiyoda-ku Tokyo 101-8430, Japan k keniti@nii.ac.jp Abstract We show that

More information

Topic: Local Search: Max-Cut, Facility Location Date: 2/13/2007

Topic: Local Search: Max-Cut, Facility Location Date: 2/13/2007 CS880: Approximations Algorithms Scribe: Chi Man Liu Lecturer: Shuchi Chawla Topic: Local Search: Max-Cut, Facility Location Date: 2/3/2007 In previous lectures we saw how dynamic programming could be

More information

Recent PTAS Algorithms on the Euclidean TSP

Recent PTAS Algorithms on the Euclidean TSP Recent PTAS Algorithms on the Euclidean TSP by Leonardo Zambito Submitted as a project for CSE 4080, Fall 2006 1 Introduction The Traveling Salesman Problem, or TSP, is an on going study in computer science.

More information

A Fast Algorithm for Optimal Alignment between Similar Ordered Trees

A Fast Algorithm for Optimal Alignment between Similar Ordered Trees Fundamenta Informaticae 56 (2003) 105 120 105 IOS Press A Fast Algorithm for Optimal Alignment between Similar Ordered Trees Jesper Jansson Department of Computer Science Lund University, Box 118 SE-221

More information

Embedding Bounded Bandwidth Graphs into l p

Embedding Bounded Bandwidth Graphs into l p UCLA Tech Report 070002 Embedding Bounded Bandwidth Graphs into l p Douglas E. Carroll Adam Meyerson February 16, 2007 Abstract We prove that graphs can be embedded randomly into l p (for any p 1) with

More information

On Covering a Graph Optimally with Induced Subgraphs

On Covering a Graph Optimally with Induced Subgraphs On Covering a Graph Optimally with Induced Subgraphs Shripad Thite April 1, 006 Abstract We consider the problem of covering a graph with a given number of induced subgraphs so that the maximum number

More information

Bidimensionality Theory and Algorithmic Graph Minor Theory

Bidimensionality Theory and Algorithmic Graph Minor Theory Bidimensionality Theory and Algorithmic Graph Minor Theory Lecture Notes for MohammadTaghi Hajiaghayi s Tutorial Mareike Massow, Jens Schmidt, Daria Schymura, Siamak Tazari MDS Fall School Blankensee October

More information

Necessary edges in k-chordalizations of graphs

Necessary edges in k-chordalizations of graphs Necessary edges in k-chordalizations of graphs Hans L. Bodlaender Abstract In this note, we look at which edges must always be added to a given graph G = (V, E), when we want to make it a chordal graph

More information

Linearity of Grid Minors in Treewidth with Applications through Bidimensionality

Linearity of Grid Minors in Treewidth with Applications through Bidimensionality Linearity of Grid Minors in Treewidth with Applications through Bidimensionality Erik D. Demaine MohammadTaghi Hajiaghayi Abstract We prove that any H-minor-free graph, for a fixed graph H, of treewidth

More information

The Encoding Complexity of Network Coding

The Encoding Complexity of Network Coding The Encoding Complexity of Network Coding Michael Langberg Alexander Sprintson Jehoshua Bruck California Institute of Technology Email: mikel,spalex,bruck @caltech.edu Abstract In the multicast network

More information

Straight-Line Drawings of 2-Outerplanar Graphs on Two Curves

Straight-Line Drawings of 2-Outerplanar Graphs on Two Curves Straight-Line Drawings of 2-Outerplanar Graphs on Two Curves (Extended Abstract) Emilio Di Giacomo and Walter Didimo Università di Perugia ({digiacomo,didimo}@diei.unipg.it). Abstract. We study how to

More information

Lecture 2 - Graph Theory Fundamentals - Reachability and Exploration 1

Lecture 2 - Graph Theory Fundamentals - Reachability and Exploration 1 CME 305: Discrete Mathematics and Algorithms Instructor: Professor Aaron Sidford (sidford@stanford.edu) January 11, 2018 Lecture 2 - Graph Theory Fundamentals - Reachability and Exploration 1 In this lecture

More information

Improved algorithms for constructing fault-tolerant spanners

Improved algorithms for constructing fault-tolerant spanners Improved algorithms for constructing fault-tolerant spanners Christos Levcopoulos Giri Narasimhan Michiel Smid December 8, 2000 Abstract Let S be a set of n points in a metric space, and k a positive integer.

More information

Vertex Cover is Fixed-Parameter Tractable

Vertex Cover is Fixed-Parameter Tractable Vertex Cover is Fixed-Parameter Tractable CS 511 Iowa State University November 28, 2010 CS 511 (Iowa State University) Vertex Cover is Fixed-Parameter Tractable November 28, 2010 1 / 18 The Vertex Cover

More information

A subexponential parameterized algorithm for Subset TSP on planar graphs

A subexponential parameterized algorithm for Subset TSP on planar graphs A subexponential parameterized algorithm for Subset TSP on planar graphs Philip N. Klein Dániel Marx Abstract Given a graph G and a subset S of vertices, the Subset TSP problem asks for a shortest closed

More information

Bidimensional Parameters and Local Treewidth

Bidimensional Parameters and Local Treewidth Bidimensional Parameters and Local Treewidth Erik D. Demaine 1, Fedor V. Fomin 2, MohammadTaghi Hajiaghayi 1, and Dimitrios M. Thilikos 3 1 MIT Laboratory for Computer Science, 200 Technology Square, Cambridge,

More information

Graphs and Discrete Structures

Graphs and Discrete Structures Graphs and Discrete Structures Nicolas Bousquet Louis Esperet Fall 2018 Abstract Brief summary of the first and second course. É 1 Chromatic number, independence number and clique number The chromatic

More information

[8] that this cannot happen on the projective plane (cf. also [2]) and the results of Robertson, Seymour, and Thomas [5] on linkless embeddings of gra

[8] that this cannot happen on the projective plane (cf. also [2]) and the results of Robertson, Seymour, and Thomas [5] on linkless embeddings of gra Apex graphs with embeddings of face-width three Bojan Mohar Department of Mathematics University of Ljubljana Jadranska 19, 61111 Ljubljana Slovenia bojan.mohar@uni-lj.si Abstract Aa apex graph is a graph

More information

arxiv: v1 [math.co] 7 Dec 2018

arxiv: v1 [math.co] 7 Dec 2018 SEQUENTIALLY EMBEDDABLE GRAPHS JACKSON AUTRY AND CHRISTOPHER O NEILL arxiv:1812.02904v1 [math.co] 7 Dec 2018 Abstract. We call a (not necessarily planar) embedding of a graph G in the plane sequential

More information

Call Control with k Rejections

Call Control with k Rejections Call Control with k Rejections R. Sai Anand, Thomas Erlebach, Alexander Hall, and Stamatis Stefanakos Computer Engineering and Networks Laboratory (TIK), ETH Zürich, CH-8092 Zürich, Switzerland {anand

More information

Approximation scheme for the traveling salesman problem

Approximation scheme for the traveling salesman problem Chapter 15 Approximation scheme for the traveling salesman problem 15.1 Multiterminal cut In this chapter, we describe a methodology that can be used for some problems for which the previous chapter s

More information

The geometric generalized minimum spanning tree problem with grid clustering

The geometric generalized minimum spanning tree problem with grid clustering 4OR (2006) 4:319 329 DOI 10.1007/s10288-006-0012-6 REGULAR PAPER The geometric generalized minimum spanning tree problem with grid clustering Corinne Feremans Alexander Grigoriev René Sitters Received:

More information

Exercise set 2 Solutions

Exercise set 2 Solutions Exercise set 2 Solutions Let H and H be the two components of T e and let F E(T ) consist of the edges of T with one endpoint in V (H), the other in V (H ) Since T is connected, F Furthermore, since T

More information

Faster parameterized algorithms for Minimum Fill-In

Faster parameterized algorithms for Minimum Fill-In Faster parameterized algorithms for Minimum Fill-In Hans L. Bodlaender Pinar Heggernes Yngve Villanger Abstract We present two parameterized algorithms for the Minimum Fill-In problem, also known as Chordal

More information

Hamiltonian cycles in bipartite quadrangulations on the torus

Hamiltonian cycles in bipartite quadrangulations on the torus Hamiltonian cycles in bipartite quadrangulations on the torus Atsuhiro Nakamoto and Kenta Ozeki Abstract In this paper, we shall prove that every bipartite quadrangulation G on the torus admits a simple

More information

Applied Mathematics Letters. Graph triangulations and the compatibility of unrooted phylogenetic trees

Applied Mathematics Letters. Graph triangulations and the compatibility of unrooted phylogenetic trees Applied Mathematics Letters 24 (2011) 719 723 Contents lists available at ScienceDirect Applied Mathematics Letters journal homepage: www.elsevier.com/locate/aml Graph triangulations and the compatibility

More information

Towards more efficient infection and fire fighting

Towards more efficient infection and fire fighting Towards more efficient infection and fire fighting Peter Floderus Andrzej Lingas Mia Persson The Centre for Mathematical Sciences, Lund University, 00 Lund, Sweden. Email: pflo@maths.lth.se Department

More information

Planarity: dual graphs

Planarity: dual graphs : dual graphs Math 104, Graph Theory March 28, 2013 : dual graphs Duality Definition Given a plane graph G, the dual graph G is the plane graph whose vtcs are the faces of G. The correspondence between

More information

Monotone Paths in Geometric Triangulations

Monotone Paths in Geometric Triangulations Monotone Paths in Geometric Triangulations Adrian Dumitrescu Ritankar Mandal Csaba D. Tóth November 19, 2017 Abstract (I) We prove that the (maximum) number of monotone paths in a geometric triangulation

More information

Improved Results on Geometric Hitting Set Problems

Improved Results on Geometric Hitting Set Problems Improved Results on Geometric Hitting Set Problems Nabil H. Mustafa nabil@lums.edu.pk Saurabh Ray saurabh@cs.uni-sb.de Abstract We consider the problem of computing minimum geometric hitting sets in which,

More information

Lecture 4: September 11, 2003

Lecture 4: September 11, 2003 Algorithmic Modeling and Complexity Fall 2003 Lecturer: J. van Leeuwen Lecture 4: September 11, 2003 Scribe: B. de Boer 4.1 Overview This lecture introduced Fixed Parameter Tractable (FPT) problems. An

More information

Faces in the Crowd: On the Algorithmic Utility of Disks and Covers

Faces in the Crowd: On the Algorithmic Utility of Disks and Covers Faces in the Crowd: On the Algorithmic Utility of Disks and Covers Faisal N. Abu-Khzam and Michael A. Langston University of Tennessee, Knoxville, TN 3793, USA Abstract. We study a pair of N P-hard problems

More information

Edge-disjoint Spanning Trees in Triangulated Graphs on Surfaces and application to node labeling 1

Edge-disjoint Spanning Trees in Triangulated Graphs on Surfaces and application to node labeling 1 Edge-disjoint Spanning Trees in Triangulated Graphs on Surfaces and application to node labeling 1 Arnaud Labourel a a LaBRI - Universite Bordeaux 1, France Abstract In 1974, Kundu [4] has shown that triangulated

More information

BIDIMENSIONAL PARAMETERS AND LOCAL TREEWIDTH

BIDIMENSIONAL PARAMETERS AND LOCAL TREEWIDTH BIDIMENSIONAL PARAMETERS AND LOCAL TREEWIDTH ERIK D. DEMAINE, FEDOR V. FOMIN, MOHAMMADTAGHI HAJIAGHAYI, AND DIMITRIOS M. THILIKOS Abstract. For several graph-theoretic parameters such as vertex cover and

More information

11. APPROXIMATION ALGORITHMS

11. APPROXIMATION ALGORITHMS 11. APPROXIMATION ALGORITHMS load balancing center selection pricing method: vertex cover LP rounding: vertex cover generalized load balancing knapsack problem Lecture slides by Kevin Wayne Copyright 2005

More information

Unit 8: Coping with NP-Completeness. Complexity classes Reducibility and NP-completeness proofs Coping with NP-complete problems. Y.-W.

Unit 8: Coping with NP-Completeness. Complexity classes Reducibility and NP-completeness proofs Coping with NP-complete problems. Y.-W. : Coping with NP-Completeness Course contents: Complexity classes Reducibility and NP-completeness proofs Coping with NP-complete problems Reading: Chapter 34 Chapter 35.1, 35.2 Y.-W. Chang 1 Complexity

More information

1 The Traveling Salesman Problem

1 The Traveling Salesman Problem Comp 260: Advanced Algorithms Tufts University, Spring 2018 Prof. Lenore Cowen Scribe: Duc Nguyen Lecture 3a: The Traveling Salesman Problem 1 The Traveling Salesman Problem The Traveling Salesman Problem

More information

Math 443/543 Graph Theory Notes 11: Graph minors and Kuratowski s Theorem

Math 443/543 Graph Theory Notes 11: Graph minors and Kuratowski s Theorem Math 443/543 Graph Theory Notes 11: Graph minors and Kuratowski s Theorem David Glickenstein November 26, 2008 1 Graph minors Let s revisit some de nitions. Let G = (V; E) be a graph. De nition 1 Removing

More information

Solving Dominating Set in Larger Classes of Graphs: FPT Algorithms and Polynomial Kernels

Solving Dominating Set in Larger Classes of Graphs: FPT Algorithms and Polynomial Kernels Solving Dominating Set in Larger Classes of Graphs: FPT Algorithms and Polynomial Kernels Geevarghese Philip, Venkatesh Raman, and Somnath Sikdar The Institute of Mathematical Sciences, Chennai, India.

More information

Dual-Based Approximation Algorithms for Cut-Based Network Connectivity Problems

Dual-Based Approximation Algorithms for Cut-Based Network Connectivity Problems Dual-Based Approximation Algorithms for Cut-Based Network Connectivity Problems Benjamin Grimmer bdg79@cornell.edu arxiv:1508.05567v2 [cs.ds] 20 Jul 2017 Abstract We consider a variety of NP-Complete network

More information

val(y, I) α (9.0.2) α (9.0.3)

val(y, I) α (9.0.2) α (9.0.3) CS787: Advanced Algorithms Lecture 9: Approximation Algorithms In this lecture we will discuss some NP-complete optimization problems and give algorithms for solving them that produce a nearly optimal,

More information

Approximation Algorithms: The Primal-Dual Method. My T. Thai

Approximation Algorithms: The Primal-Dual Method. My T. Thai Approximation Algorithms: The Primal-Dual Method My T. Thai 1 Overview of the Primal-Dual Method Consider the following primal program, called P: min st n c j x j j=1 n a ij x j b i j=1 x j 0 Then the

More information

Totally odd subdivisions and parity subdivisions: Structures and Coloring

Totally odd subdivisions and parity subdivisions: Structures and Coloring Totally odd subdivisions and parity subdivisions: Structures and Coloring Downloaded 02/06/18 to 46.3.204.37. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php

More information

Exponential time algorithms for the minimum dominating set problem on some graph classes

Exponential time algorithms for the minimum dominating set problem on some graph classes Exponential time algorithms for the minimum dominating set problem on some graph classes Serge Gaspers University of Bergen Department of Informatics N-500 Bergen, Norway. gaspers@ii.uib.no Dieter Kratsch

More information

Lecture 1: TSP on graphs of bounded branch width

Lecture 1: TSP on graphs of bounded branch width CS523: Advanced Algorithms Spring 2014 Lecture 1: TSP on graphs of bounded branch width Lecturer: Glencora Borradaile Scribes: Hung Le 1.1 Branch Decomposition Let G = (V (G), E(G)) be an undirected graph.

More information

Faster parameterized algorithms for Minimum Fill-In

Faster parameterized algorithms for Minimum Fill-In Faster parameterized algorithms for Minimum Fill-In Hans L. Bodlaender Pinar Heggernes Yngve Villanger Technical Report UU-CS-2008-042 December 2008 Department of Information and Computing Sciences Utrecht

More information

V1.0: Seth Gilbert, V1.1: Steven Halim August 30, Abstract. d(e), and we assume that the distance function is non-negative (i.e., d(x, y) 0).

V1.0: Seth Gilbert, V1.1: Steven Halim August 30, Abstract. d(e), and we assume that the distance function is non-negative (i.e., d(x, y) 0). CS4234: Optimisation Algorithms Lecture 4 TRAVELLING-SALESMAN-PROBLEM (4 variants) V1.0: Seth Gilbert, V1.1: Steven Halim August 30, 2016 Abstract The goal of the TRAVELLING-SALESMAN-PROBLEM is to find

More information

/633 Introduction to Algorithms Lecturer: Michael Dinitz Topic: Approximation algorithms Date: 11/27/18

/633 Introduction to Algorithms Lecturer: Michael Dinitz Topic: Approximation algorithms Date: 11/27/18 601.433/633 Introduction to Algorithms Lecturer: Michael Dinitz Topic: Approximation algorithms Date: 11/27/18 22.1 Introduction We spent the last two lectures proving that for certain problems, we can

More information

Discrete Mathematics I So Practice Sheet Solutions 1

Discrete Mathematics I So Practice Sheet Solutions 1 Discrete Mathematics I So 2016 Tibor Szabó Shagnik Das Practice Sheet Solutions 1 Provided below are possible solutions to the questions from the practice sheet issued towards the end of the course. Exercise

More information

On the Convexity Number of Graphs

On the Convexity Number of Graphs On the Convexity Number of Graphs Mitre C. Dourado 1, Fábio Protti, Dieter Rautenbach 3, and Jayme L. Szwarcfiter 4 1 ICE, Universidade Federal Rural do Rio de Janeiro and NCE - UFRJ, Brazil, email: mitre@nce.ufrj.br

More information

A Framework for Network Reliability Problems on Graphs of Bounded Treewidth

A Framework for Network Reliability Problems on Graphs of Bounded Treewidth A Framework for Network Reliability Problems on Graphs of Bounded Treewidth Thomas Wolle Institute of Information and Computing Sciences, Utrecht University P.O.Box 80.089, 3508 TB Utrecht, The Netherlands

More information

CS261: A Second Course in Algorithms Lecture #16: The Traveling Salesman Problem

CS261: A Second Course in Algorithms Lecture #16: The Traveling Salesman Problem CS61: A Second Course in Algorithms Lecture #16: The Traveling Salesman Problem Tim Roughgarden February 5, 016 1 The Traveling Salesman Problem (TSP) In this lecture we study a famous computational problem,

More information

Max-Cut and Max-Bisection are NP-hard on unit disk graphs

Max-Cut and Max-Bisection are NP-hard on unit disk graphs Max-Cut and Max-Bisection are NP-hard on unit disk graphs Josep Díaz and Marcin Kamiński 2 Llenguatges i Sistemes Informàtics 2 RUTCOR, Rutgers University Universitat Politècnica de Catalunya 640 Bartholomew

More information

Geometric Unique Set Cover on Unit Disks and Unit Squares

Geometric Unique Set Cover on Unit Disks and Unit Squares CCCG 2016, Vancouver, British Columbia, August 3 5, 2016 Geometric Unique Set Cover on Unit Disks and Unit Squares Saeed Mehrabi Abstract We study the Unique Set Cover problem on unit disks and unit squares.

More information

1 The Traveling Salesperson Problem (TSP)

1 The Traveling Salesperson Problem (TSP) CS 598CSC: Approximation Algorithms Lecture date: January 23, 2009 Instructor: Chandra Chekuri Scribe: Sungjin Im In the previous lecture, we had a quick overview of several basic aspects of approximation

More information

Subexponential Algorithms for Partial Cover Problems

Subexponential Algorithms for Partial Cover Problems Subexponential Algorithms for Partial Cover Problems Fedor V. Fomin Daniel Lokshtanov Venkatesh Raman Saket Saurabh July 4, 2009 Abstract Partial Cover problems are optimization versions of fundamental

More information

Planar Graphs. 1 Graphs and maps. 1.1 Planarity and duality

Planar Graphs. 1 Graphs and maps. 1.1 Planarity and duality Planar Graphs In the first half of this book, we consider mostly planar graphs and their geometric representations, mostly in the plane. We start with a survey of basic results on planar graphs. This chapter

More information