Lecture and notes by: Nate Chenette, Brent Myers, Hari Prasad November 8, Property Testing

Size: px
Start display at page:

Download "Lecture and notes by: Nate Chenette, Brent Myers, Hari Prasad November 8, Property Testing"

Transcription

1 Property Testing 1 Introduction Broadly, property testing is the study of the following class of problems: Given the ability to perform (local) queries concerning a particular object (e.g., a function, or a graph), the task is to determine whether the object has a predetermined (global) property (e.g., linearity or bipartiteness), or is far from having the property. The task should be performed by inspecting only a small (possibly randomly selected) part of the whole object, where a small probability of failure is allowed.[2] 1.1 Testing Linearity A Motivating Example To give the reader an intuitive sense of why and how property testing is useful, we begin with a motivating example. Let F be a finite field. Consider the definition of functional linearity: Definition. A function f : F m F is linear if and only if for every x, y F m, f(x) + f(y) = f(x + y). Now, suppose you are given a function f : F m F, and are asked to determine whether it is linear. More specifically, you are given a sort of black box representing the function, into which you can feed a value x F m, and receive back the value f(x) F (such an action is called a query. A naive approach to this problem might be to take every pair of values x, y in F m, and, using the black box, test whether f(x)+f(y) = f(x+y). In fact, if we wish to know absolutely whether f is linear or not, then something like this naive approach is necessary. But one can readily see that this approach is impractical if F is of a nontrivial size. And besides, if we feed every value of F m into the black box, then this actually determines the value of f, which seems like an exorbitantly strong intermediate result in determining a property of f (linearity). However, suppose we are faced with a relaxation of the problem: given that you can compute any value of f, determine whether f is linear or far from linear with high probability. Now, we can exploit the black box model to answer this question quickly, using the following approach. Algorithm (Sketch). Let us pick a set S of n random elements of F m. Now, check the linearity of the subset, by comparing f(x) + f(y) and f(x + y) for every pair x, y S. If any such comparison is an inequality, answer REJECT. Otherwise, if all comparisons are equalities, answer ACCEPT. It is easy to see that, the larger S is, the more accurate our answer can be, and in particular, the probability of success can be brought arbitrarily close to 1 by increasing S. A few more notes about this particular algorithm: This particular algorithm has one-sided error, meaning that if f is linear, then the algorithm always answers ACCEPT. Equivalently, the contrapositive is true: if the algorithm answers REJECT, then f is not linear with probability 1. If f is determined to be not linear, then this algorithm even provides a witness, i.e., two elements x, y F m such that f(x) + f(y) f(x + y). 1

2 1.2 Property testing definitions We see that the above motivating example fits into the broad definition of property testing given at the beginning of this paper. The object in question is a function, and we wish to determine whether it has, or is far from having, the property of linearity. Our algorithm randomly selected a small part of the object (namely, f(s)) to inspect and determined an answer that is true with high probability. Let us formalize some of the notions in our definition of property testing. Let U be a family of objects of the same type and representation (for example, U could be the family of graphs under the adjacency matrix model.) Let us also define a distance function for elements of U, δ : U U [0, 1]. Let P be a property defined on elements of U. Then for any ɛ [0, 1], we say an object X U is ɛ far from having property P if δ(x, Y ) > ɛ for all objects Y U having property P. Otherwise, X is ɛ-close. A property testing algorithm for a property P is given a distance parameter ɛ and an object. The algorithm should accept with probability at least p > 1/2 every object that has property P, and should reject with probability q > 1/2 every object that is ɛ-far from having property P. (Note: p and q are usually arbitrarily set to 2/3. Recall that, as in random algorithms, we can repeat the algorithm in order to achieve an arbitrarily high assurance bound.) If one of p or q is 1, then we say the algorithm has one-sided-error, and if both are strictly less that 1, the algorithm has twosided-error. The efficiency of a property testing algorithm can be measured using two parameters, the running time and the query complexity (the number of queries the algorithm performs on an object.) Most property testing algorithms actually achieve query and time complexities that depend on ɛ rather than the size of the object! 1.3 Formalization of the Linearity Testing Example Here is a formalization of the algorithm for testing the linearity of a function f : F m F. We say that a pair of elements x, y is a violating pair if f(x) + f(y) f(x + y). The distance of a function from linearity will be defined as the number of violating pairs over the domain of the function. Linearity Test 1. Uniformly and independently select m = Θ(ɛ 1 ) pairs of elements x, y F. 2. For every pair of elements selected, check whether f(x) + f(y) = f(x + y). 3. If any of the selected pairs is a violating pair, then reject, otherwise, accept. By the definition of linearity, if f is linear then it is always accepted. It thus remains to prove: Theorem 1. If f is ɛ-far from linear, then with probability at least 2/3, Linearity Test rejects it. Proof. We define the distance from linearity of a nonlinear function in this setting to be the number of points f(x) needed to be changed in order to make f a linear function. Let δ denote the (exact) distance of f from linearity (so that in particular, δ > ɛ). We will prove the theorem for the case where δ is bounded away from 1 2, i.e., δ 1 2 γ for some constant γ > 0. We shall show that the probability that a single uniformly selected pair of elements is a violating pair is at least 3δ(1 2δ). For δ bounded away from 1 2, this probability is Ω(δ). Since the test 2

3 selects Θ(1/ɛ) = ω(1/δ) pairs, the probability that no violating pair is selected is at most 1/3 (for the appropriate choice of the constant in the Ω( ) notation). Let g be a linear function at distance δ from f. Let G {x : f(x) = g(x)} be the set of good elements in F on which f and g agree. For any pair x, y, if among the three elements x, y, and x + y, two of them belong to G while the third doesn t, then x, y are a violating pair. Hence, Pr[x, y are a violating pair] Pr[x / G, y G, (x + y) G] +Pr[x G, y / G, (x + y) G] +Pr[x G, y G, (x + y) / G] (1) Consider the first probability in the above sum (the treatment of the other two is analogous, as the important property of any triplet is that every two of the elements are pairwise independent). Pr[x / G, y G, (x + y) G] = Pr[x / G] Pr[y G, (x + y) G x / G] = δ (1 Pr[y / G or (x + y) / G x / G] (2) By using a probability union bound, and the fact that both y and (x+y) are uniformly distributed, 1 Pr[y / G or (x + y) / G x / G] 1 2Pr[y / G x / G] (3) Since x and y are chosen independently, Pr[y / G x / G] = Pr[y / G] = δ (4) and so by combining Equations (1) - (4), the probability of selecting a violating pair is at least 3 δ (1 2δ). 2 Property Testing in Graphs 2.1 Graph Representations Adjacency Matrix Model One way to represent a graph is by its adjacency matrix. An adjacency matrix is is a 0 1 matrix with a ij = 1 if and only if (i, j) E. A property testing algorithm under the adjacency matrix model is allowed to query whether there is an edge between any two vertices of its choice. That is, the algorithm queries vertices i, j and is given the value of a ij = a ji. In this representation, the distance between graphs is the fraction of entries in the adjacency matrix on which the two graphs differ. By this definition, for a given distance parameter ɛ, the algorithm should reject every graph that requires more than ɛ 2 V 2 edge modifications in order to acquire the tested property (the factor of 1 2 is because each edge is represented twice in the matrix.) This representation is most appropriate for dense graphs, and the results for testing in this model are most meaningful for dense graphs Incidence-List Model In this model, a graph is represented by lists of length d, where d is a bound on the degree of the graph. Here, the testing algorithm can query, for every vertex v and index i {1,..., d}, which vertex is the ith neighbor of v. If no such neighbor exists, then the answer is 0. Analogously to the adjacency-matrix model, the distance between graphs is defined to be the fraction of entries on 3

4 which the graphs differ according to this representation. Since the total number of incidence-list entries is d V, a graph should be rejected if the number of edge modifications required in order to obtain the property is greater than ɛ 2 d V. (Once again, the factor of 1 2 is because each edge (u, v) is represented in the list for u and the list for v.) This model is best for sparse graphs that have low or bounded maximum degree. A variant of the above model allows the incidence lists to be of varying lengths. In such a case, the distance between graphs is defined with respect to the total number of edges in the graph (or an upper bound on this number). This model is suitable for testing graphs that are not dense but for which there is a large variance in the degrees of the graph vertices. Furthermore, some problems are more interesting in this model, in the sense that removing the degree bound makes them less restricted. For example, testing whether a graph has a diameter of at most a bounded size, is less interesting in the bounded degree model, since a bound d on the degree implies a lower bound on the diameter of a graph. Intuitively, testing in this model is at least as hard as testing in the bounded-degree model described above, and in fact in some cases it is strictly harder. 2.2 Summary of Results The following graph properties were studied in [1] and were shown to have testing algorithms with query complexity poly(1/ɛ) and time complexity at most exp(poly(1/ɛ)). In what follows, N denotes the number of graph vertices. [Note: Õ is the same as O, but without logarithmic factors.] 1. Bipartiteness. The algorithm has query complexity and running time Õ(ɛ 3 ). Alon and Krivelevich [3] improved the analysis of the algorithm and obtained a bound Õ(ɛ 2 ) on the query complexity and running time. 2. k-colorability, ) k 3. The algorithm has query complexity Õ(k4 /ɛ 6 ) and running time exp (Õ(k 2 /ɛ 3 ). Alon and Krivelevich [3] improved the analysis of the algorithm and obtained ) a bound of Õ(k2 /ɛ 4 ) on the query complexity, and exp (Õ(k/ɛ 2 ) on the running time. 3. ρ-clique. The property is having a clique of size ρ N, where 0 < ρ < 1 is a constant. ) The query complexity of the algorithm is Õ(ρ2 /ɛ 6 ) and the running time is exp (Õ(ρ/ɛ 2 ). 4. ρ-cut. The property is having a 2-way cut with ρn 2 crossing) edges. The query complexity of the algorithm is Õ(ɛ 7 ) and the running time is exp (Õ(ɛ 3 ). The algorithm generalizes to k-way cuts, at a multiplicative cost of O(log 2 (k)) in the query complexity and in the exponent of the running time. The algorithm can also be modified to test ρ-bisection. This property is similar to ρ-cut except that the partition ) is to equal size subsets. The query complexity is Õ(ɛ 8 ) and the running time is exp (Õ(ɛ 3 ). For all the above properties (except bipartiteness) it is very unlikely that there is a testing algorithm having running time poly(1/ɛ). If such an algorithm exists, by setting ɛ = 1/N we would be able to obtain an exact (randomized) decision procedure that runs in polynomial time, and this would imply that NP BP P. (BP P is the class of decision problems solvable by a probabilistic Turing machine in polynomial time, with an error probability of at most 1/3 for all instances.) The bipartiteness and k-colorability algorithms have one-sided error. Furthermore, whenever a graph is rejected, the algorithm supplies evidence that it does not have the property. Evidence is 4

5 in the form of a subgraph that is not bipartite/k-colorable. All other algorithms have two-sided error, and this can be shown to be unavoidable within o(n) query-complexity. 2.3 Testing Bipartiteness We would like to be able to use testing algorithms to test properties of large graphs. One common property of graphs is bipartiteness. Definition. A graph is bipartite if it can be partitioned into two disjoint subsets with no edges within each set. Recall that a graph is bipartite if and only if it has no odd length cycles. Clearly bipartiteness can be tested in O( V + E ) with no error by running BFS to find odd cycles. However, we can speed this up considerably if we can afford some inaccuracy. If we just take a random sample from the graph and test the induced subgraph for bipartiteness, what do we know about the full graph? If any induced subgraph contains an odd cycle, then the full graph also must contain an odd cycle. Any induced subgraph of a bipartite graph will always be bipartite. These two facts are encouraging for the sampling approach. This approach will always correctly label bipartite graphs. Also, if we find that the subgraph is not bipartite, the sample will contain proof that the graph is not bipartite. However, we still have to show that we find this proof with high probability. Fortunately, we do not have to reject any non-bipartite graph, only graphs that are ɛ-far from bipartite must be rejected. The distance from bipartiteness is defined in terms of the number of violating edges in the graph. Definition. For some two-way partition of a graph, an edge is a violating edge if both of its ends are in the same partition. Definition. A two-way partition of a graph is ɛ-bad if the number of violating edges is greater than ɛn 2. Otherwise it is ɛ-good. Definition. A graph is ɛ-far from bipartite if all possible partitions are ɛ-bad. If the partitions are known, how many pairs of vertices must be sampled to find a violating edge with high probability? In some ɛ-bad partitioning of a graph, there are at least ɛn 2 violating edges by definition. Also, there are at most N(N 1)/2 < N 2 edges in any simple graph. The probability of discovering a violating edge from a uniform sample is then at least ɛn 2 = ɛ. So to N 2 to find a violating edge with high probability we must sample O(ɛ 1 ) edges. Unfortunately we do not know the partitioning, so we must sample enough vertices to find a violating edge in every partitioning. This results in a sample size linear in the number of vertices. It turns out that we can do a little better in the analysis to obtain a better query complexity. First, let s state the algorithm more formally. Algorithm 1. Tests a graph for bipartiteness. 1. Uniformly and independently select m = Θ( log(1/ɛ) ) vertices. (m will be derived later) ɛ 2 5

6 2. Obtain the subgraph induced by the sampled vertices. 3. Use BFS to determine if the induced subgraph is bipartite. If it is bipartite, ACCEPT, otherwise, REJECT. Theorem 2. Algorithm 1 is a testing algorithm for bipartiteness. Proof. We have already discussed the one-sided error of our algorithm above. We only need to show that the algorithm rejects graphs ɛ-far from bipartite with probability 2/3. Definition. A vertex is influential if its degree is at least ɛ 4N. Otherwise it is not influential. In a bipartite graph, E < mn where m and n are the sizes of the two partitions. E is maximized when the sets are both the same size. So E < N 2 /4 for all bipartite graphs. Since there are at most N non-influential vertices in a graph, there are at most (ɛ/4)n 2 edges incident to non-influential vertices. This gives us the intuition that influential vertices will tell us much more than non-influential vertices about the bipartiteness of the graph. Definition. A set of vertices U covers a vertex v if v has at least one neighbor in U. Lemma 3. Let U be a set of vertices of size Θ(log(1/ɛ)/ɛ) selected uniformly and independently from a simple graph G. With probability at least 5/6 over the choice of the vertices in U, all but at most (ɛ/4)n influential vertices are covered by U. Proof. Let v be an influential vertex. If we randomly pick a set U of vertices uniformly and independently, what are the chances that the set covers v? Since v has at least (ɛ/4)n neighbors, the probability of finding none of them is at most (1 ɛ 4 ) U < exp( ɛ 4 U ). If we let U = 4 ɛ ln(24/ɛ), then the probability above reduces to ɛ/24. Since there are at most N influential vertices, the expected number of influential vertices not covered by U is at most (ɛ/24)n. Applying Markov s inequality shows that the probability of having greater than (ɛ/4)n vertices is at most 1/6 = 1 5/6. Lemma 4. Let G = (V, E) be a graph that is ɛ-far from bipartite, U a subset of V that covers all but at most (ɛ/4)n of the influential vertices in G, and (U 1, U 2 ) a fixed partition of U. Let S be a uniformly and independently selected sample of size Θ( U /ɛ) vertices. Then, with probability at least 1 2 U /6 over the choice of S, for every partition (S 1, S 2 ) of S, there is some edge between vertices in U S that is violating with respect to (U 1 S 1, U 2 S 2 ). Proof. Let C be the set of vertices in V that are covered by U and let R be the remaining vertices. R contains at most (ɛ/4)n vertices by Lemma 3. Break up U into a partition (U 1, U 2 ). Now there is a similar partiton (C 1, C 2 ) of C based on the cover sets of (U 1, U 2 ). Let C 2 be the set of vertices covered by U 1, and C 1 be C C 2. Also partition R into (R 1, R 2 ) with U 1 R R 1 and U 2 R R 2, and the rest of R distributed arbitrarily. The partition (C 1 R 1, C 2 R 2 ) must be ɛ-bad since the graph is ɛ-far from bipartite. Which edges in the graph are violating with respect to this partition? There are at most (ɛ/4)n influential vertices in R, and at most N non-influential vertices in the entire graph. Also there could be at most N non-influential edges each adjacent to ɛ 4N edges. So the number of edges incident to R is at most (ɛ/4)n N + N (ɛ/4)n = (ɛ/2)n 2. This shows that for all possible partitionings of R, there are at most (ɛ/2)n 2 violating edges incident to R. This leaves (ɛ)n 2 (ɛ/2)n 2 = (ɛ/2)n 2 violating edges that are not incident to R, and must be within C 1 or within C 2. 6

7 If S is viewed as pairs of vertices, the probability that each pair is a violating edge with respect to (C 1, C 2 ) is at least ɛ/2. So the probability there is no violating edge among all S /2 pairs is at most (1 (ɛ/2)) S /2. If S = 16 U /ɛ, the probability is less than 2 U /6. To complete the claim we need to show that if S contains such a pair, then it is not possible to partition S into (S 1, S 2 ) so that (U 1 S 1, U 2 S 2 ) has no violating edges. Consider an edge (v, w) such that v, w S that violates (C 1, C 2 ), and without loss of generality, assume v, w C 2. If we put both vertices either in S 1 or S 2 then (v, w) is violating with respect to (S 1, S 2 ). However, since v and w belong to C 2, by our definition of (C 1, C 2 ), v has some neighbor u U 1 and w has some neighbor u U 1. Therefore if we put v S 1 and w S 2, then the edge (u, v) will be violating, and if we put w S 1 and v S 2, then the edge (u, w) will be violating. The lemma follows. Combining Lemma 4 with the fact that there are 2 U partitions of U, it follows that with probability 5/6 over the choice of S, for every partition (U 1, U 2 ) of U, and for every partition (S 1, S 2 ) of S, the sample contains edges that violate (U 1 S 1, U 2 S 2 ). In other words, the sample U S cannot be partitioned without violations. Combining this with Lemma 3, the theorem follows. As described, Algorithm 1 has query and time complexities that are quadratic in the size of Figure 1: Diagram of graph structure induced by (U 1, U 2 ). Vertices in C 2 have at least one neighbor in U 1, and vertices in C 1 have at least one neighbor in C 1. R has no neighbors in U 1 or U 2, with the remaining elements partitioned arbitrarily. Edges incedent to R (at most (ɛ/2)/n 2 ) are dotted. Violating edges within C 1 or within C 2 (at most (ɛ/2)/n 2 ) are dashed. 7

8 the sample. That is, Θ(log 2 (1/ɛ)/ɛ 4 ). However, given the analysis, we can slightly improve on this bound. Assume the algorithm actually partitions the sample into two parts U and S of sizes m 1 = Θ(log(1/ɛ)/ɛ) and m 2 = Θ(log(1/ɛ)/ɛ 2 ) respectively. It views S as consisting of m 2 /2 pairs of vertices, and queries whether an edge exists only between all pairs of vertices in U S, and between the m 2 /2 pairs in S. It then checks whether the resulting subgraph is bipartite. Then by the analysis this suffices to obtain the desired success probability while decreasing the complexities to Θ(log 2 (1/ɛ)/ɛ 3 ) 3 Other Applications 3.1 Constructing (Good) Graph Partitions For many graph properties, if the graph has the desired property, the testing algorithm outputs some auxiliary information which allows to construct, in poly(1/ɛ) N time, a partition that approximately obeys the property. For example, for ρ-clique, the algorithm will find a subset of vertices of size ρn, such that at most ɛn 2 edges need to be added so that it becomes a clique. In the case of ρ-cut, the algorithm will construct a partition with at least (ρ ɛ)n 2 crossing edges. The basic idea is that the partition of the sample that caused the algorithm to accept is used to partition the whole graph. 3.2 A General Class of Partition Properties A 1998 paper of Goldreich, Goldwasser, and Ron [1], which introduced the notion of graph property testing, also provided an property testing algorithm and bounds for an entire class of property testing problems they called partition problems. What follows is a statement of this result, without description of the algorithm or proof (which are quite complex), but with a short list of problems for which the algorithm applies. The following framework of a general partition problem captures any graph property which requires the existence of partitions satisfying certain fixed density constraints. These constraints may refer both to the number of vertices in each component of the partition and to the number of edges between each pair of components. Let Φ def = { ρ LB j, ρ UB } k j j=1 { ϱ LB } k j,j, ϱub j,j be a set of non-negative parameters so that j,j ρlb =1 j ρ UB j for all j and ϱ LB j,j ϱ UB j,j for all j and j. (LB stands for Lower Bound, UB stands for Upper Bound.) Let GP Φ be the class of graphs which have a k-way partition (V 1,..., V k ) with the following conditions being satisfied. j, ρ LB j N V j ρ UB j N and j, j, ϱ LB j,j N 2 E(V j, V j ) ϱ UB j,j N 2 where, recall that E(V j, V j ) is the set of edges between vertices in V 1 and vertices in V j. That is, the first relation places lower and upper bounds on the relative sizes of the various components of the partition; whereas the second relation imposes lower and upper bounds on the density of edges among the various pairs of components. Theorem 5. There exists an algorithm A such that for every given set of parameters Φ, algorithm A is a property testing algorithm for the class GP Φ with query complexity ( O(k 2 )/ɛ ) k+5 k 2 log(k/ɛδ), and running time exp ( (O(k 2 )/ɛ) k+2 log(k/ɛδ) ). 8

9 For a description of the algorithm and proof of this theorem, see [1]. What follows is a list of property testing problems (along with values for the density constraints) for which the above theorem provides an algorithm to solve, with the stated bounds. 1. Bipartiteness. We set k = 2 (as we are interested in a two-way partition), ρ LB 1 = ρ LB 2 = 0, and ρ UB 1 = ρ UB 2 = 1 (as there are no restrictions on the sizes of the partition subsets), ϱ LB 1,1 = ϱlb 2,2 = ϱub 1,1 = ϱub 2,2 = 0, enforcing the main constraint that there be no edges within the partition subsets, and finally ϱ LB 1,2 = 0, ϱub 1,2 = 1, since the number of edges between the two subsets is not restricted. 2. k-colorability, k 3. This is a generalization of bipartiteness, and the important density bounds are ϱ UB j,j = 0 for every 1 j k. All lower bounds are 0 and all other upper bounds are ρ-clique. Here, k = 1, ρ LB 1 = ρ UB 1 = ρ, enforcing the restriction that one subset should be of size ρn, and ϱ LB 1 = 1 2 (ρ2 ρ/n), enforcing the restriction that the subset be a clique. 4. ρ-cut. Here k = 2, and ϱ LB 1,2 = ρ UB j = 1/2 for j {1, 2}. ρ LB l 4 Conclusion and Open Problems = ρ. For the case of ρ-bisection we add the constraints that Property testing is a useful tool that allows for sublinear algorithms when some accuracy can be sacrificed. However there are many properties that are hard to test. For example, First Order Graph Properties are hard to test. Let A(x 1,, x t, y 1,, y s ) be a quantifier free graph expression. That is, it contains equality of vertices, adjacency relations between vertices, and boolean connectives. Properties that are expresed in this form are known as First Order Graph Properties. There are many classes of first order graph properties, most of which are hard to test. For example, an expression is an EA First Order Graph Property if the expression is of the form x 1,, x t y 1,, y s A(x 1,, x t, y 1,, y s ). An expression of the type AE is defined analogously. k-colorability is known to be equivalent to an EA type problem. Properties that can be described as an EA statement have been shown to have query complexity and running time independent of the size of the graph. However, the best known upper bound on the number of queries is a tower of towers of poly(1/ɛ) exponents. It is not known if this can be improved upon. [4] References [1] O. Goldreich, S. Goldwasser, and D. Ron. Property testing and its connection to learning and approximation. JACM, 45(4): , [2] Ron, Dana. Property Testing. Appeared: Handbook on Randomization, Volume II. Editors: S. Rajasekaran, P. M. Pardalos, J. H. Reif, J. D. P. Rolim. [3] N. Alon and M. Krivelevich. Testing k-colorability. Manuscript, [4] N. Alon, E. Fischer, M. Krivelevich, and M. Szegedy. Efficient testing of large graphs. In Proceedings of the Fortieth Annual Symposium on Foundations of Computer Science, pages ,

Comparing the strength of query types in property testing: The case of testing k-colorability

Comparing the strength of query types in property testing: The case of testing k-colorability Comparing the strength of query types in property testing: The case of testing k-colorability Ido Ben-Eliezer Tali Kaufman Michael Krivelevich Dana Ron Abstract We study the power of four query models

More information

Theoretical Computer Science. Testing Eulerianity and connectivity in directed sparse graphs

Theoretical Computer Science. Testing Eulerianity and connectivity in directed sparse graphs Theoretical Computer Science 412 (2011) 6390 640 Contents lists available at SciVerse ScienceDirect Theoretical Computer Science journal homepage: www.elsevier.com/locate/tcs Testing Eulerianity and connectivity

More information

Sublinear Algorithms Lectures 1 and 2. Sofya Raskhodnikova Penn State University

Sublinear Algorithms Lectures 1 and 2. Sofya Raskhodnikova Penn State University Sublinear Algorithms Lectures 1 and 2 Sofya Raskhodnikova Penn State University 1 Tentative Topics Introduction, examples and general techniques. Sublinear-time algorithms for graphs strings basic properties

More information

Exact Algorithms Lecture 7: FPT Hardness and the ETH

Exact Algorithms Lecture 7: FPT Hardness and the ETH Exact Algorithms Lecture 7: FPT Hardness and the ETH February 12, 2016 Lecturer: Michael Lampis 1 Reminder: FPT algorithms Definition 1. A parameterized problem is a function from (χ, k) {0, 1} N to {0,

More information

Comparing the strength of query types in property testing: The case of testing k-colorability

Comparing the strength of query types in property testing: The case of testing k-colorability Comparing the strength of query types in property testing: The case of testing k-colorability Ido Ben-Eliezer Tali Kaufman Michael Krivelevich Dana Ron October 1, 2007 Abstract We study the power of four

More information

Testing Eulerianity and Connectivity in Directed Sparse Graphs

Testing Eulerianity and Connectivity in Directed Sparse Graphs Testing Eulerianity and Connectivity in Directed Sparse Graphs Yaron Orenstein School of EE Tel-Aviv University Ramat Aviv, Israel yaronore@post.tau.ac.il Dana Ron School of EE Tel-Aviv University Ramat

More information

Testing subgraphs in large graphs

Testing subgraphs in large graphs Testing subgraphs in large graphs Noga Alon Abstract Let H be a fixed graph with h vertices, let G be a graph on n vertices and suppose that at least ɛn 2 edges have to be deleted from it to make it H-free.

More information

Notes and Comments for [1]

Notes and Comments for [1] Notes and Comments for [1] Zhang Qin July 14, 007 The purpose of the notes series Good Algorithms, especially for those natural problems, should be simple and elegant. Natural problems are those with universal

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

/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

MAXIMAL PLANAR SUBGRAPHS OF FIXED GIRTH IN RANDOM GRAPHS

MAXIMAL PLANAR SUBGRAPHS OF FIXED GIRTH IN RANDOM GRAPHS MAXIMAL PLANAR SUBGRAPHS OF FIXED GIRTH IN RANDOM GRAPHS MANUEL FERNÁNDEZ, NICHOLAS SIEGER, AND MICHAEL TAIT Abstract. In 99, Bollobás and Frieze showed that the threshold for G n,p to contain a spanning

More information

Three Theorems Regarding Testing Graph Properties*

Three Theorems Regarding Testing Graph Properties* Three Theorems Regarding Testing Graph Properties* Oded Goldreich, 1, Luca Trevisan 1 Department of Computer Science and Applied Mathematics, Weizmann Institute of Science, Rehovot, Israel; e-mail: oded@wisdom.weizmann.ac.il

More information

PCP and Hardness of Approximation

PCP and Hardness of Approximation PCP and Hardness of Approximation January 30, 2009 Our goal herein is to define and prove basic concepts regarding hardness of approximation. We will state but obviously not prove a PCP theorem as a starting

More information

arxiv: v1 [math.co] 28 Sep 2010

arxiv: v1 [math.co] 28 Sep 2010 Densities of Minor-Closed Graph Families David Eppstein Computer Science Department University of California, Irvine Irvine, California, USA arxiv:1009.5633v1 [math.co] 28 Sep 2010 September 3, 2018 Abstract

More information

The clique number of a random graph in (,1 2) Let ( ) # -subgraphs in = 2 =: ( ) We will be interested in s.t. ( )~1. To gain some intuition note ( )

The clique number of a random graph in (,1 2) Let ( ) # -subgraphs in = 2 =: ( ) We will be interested in s.t. ( )~1. To gain some intuition note ( ) The clique number of a random graph in (,1 2) Let () # -subgraphs in = 2 =:() We will be interested in s.t. ()~1. To gain some intuition note ()~ 2 =2 and so ~2log. Now let us work rigorously. () (+1)

More information

W[1]-hardness. Dániel Marx. Recent Advances in Parameterized Complexity Tel Aviv, Israel, December 3, 2017

W[1]-hardness. Dániel Marx. Recent Advances in Parameterized Complexity Tel Aviv, Israel, December 3, 2017 1 W[1]-hardness Dániel Marx Recent Advances in Parameterized Complexity Tel Aviv, Israel, December 3, 2017 2 Lower bounds So far we have seen positive results: basic algorithmic techniques for fixed-parameter

More information

The clique number of a random graph in (,1 2) Let ( ) # -subgraphs in = 2 =: ( ) 2 ( ) ( )

The clique number of a random graph in (,1 2) Let ( ) # -subgraphs in = 2 =: ( ) 2 ( ) ( ) 1 The clique number of a random graph in (,1 2) Let () # -subgraphs in = 2 =:() We will be interested in s.t. ()~1. To gain some intuition note ()~ 2 =2 and so ~2log. Now let us work rigorously. () (+1)

More information

Notes for Lecture 24

Notes for Lecture 24 U.C. Berkeley CS170: Intro to CS Theory Handout N24 Professor Luca Trevisan December 4, 2001 Notes for Lecture 24 1 Some NP-complete Numerical Problems 1.1 Subset Sum The Subset Sum problem is defined

More information

Small Survey on Perfect Graphs

Small Survey on Perfect Graphs Small Survey on Perfect Graphs Michele Alberti ENS Lyon December 8, 2010 Abstract This is a small survey on the exciting world of Perfect Graphs. We will see when a graph is perfect and which are families

More information

3 No-Wait Job Shops with Variable Processing Times

3 No-Wait Job Shops with Variable Processing Times 3 No-Wait Job Shops with Variable Processing Times In this chapter we assume that, on top of the classical no-wait job shop setting, we are given a set of processing times for each operation. We may select

More information

Approximation slides 1. An optimal polynomial algorithm for the Vertex Cover and matching in Bipartite graphs

Approximation slides 1. An optimal polynomial algorithm for the Vertex Cover and matching in Bipartite graphs Approximation slides 1 An optimal polynomial algorithm for the Vertex Cover and matching in Bipartite graphs Approximation slides 2 Linear independence A collection of row vectors {v T i } are independent

More information

Introduction to Testing Graph Properties

Introduction to Testing Graph Properties Introduction to Testing Graph Properties Oded Goldreich Abstract. The aim of this article is to introduce the reader to the study of testing graph properties, while focusing on the main models and issues

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

Number Theory and Graph Theory

Number Theory and Graph Theory 1 Number Theory and Graph Theory Chapter 6 Basic concepts and definitions of graph theory By A. Satyanarayana Reddy Department of Mathematics Shiv Nadar University Uttar Pradesh, India E-mail: satya8118@gmail.com

More information

Testing Hypergraph Coloring

Testing Hypergraph Coloring Testing Hypergraph Coloring Artur Czumaj 1 and Christian Sohler 2 1 Department of Computer and Information Science, New Jersey Institute of Technology, University Heights, Newark, NJ 07102-1982, USA. czumaj@cis.njit.edu

More information

On the Relationships between Zero Forcing Numbers and Certain Graph Coverings

On the Relationships between Zero Forcing Numbers and Certain Graph Coverings On the Relationships between Zero Forcing Numbers and Certain Graph Coverings Fatemeh Alinaghipour Taklimi, Shaun Fallat 1,, Karen Meagher 2 Department of Mathematics and Statistics, University of Regina,

More information

Towards the Proof of the PCP Theorem

Towards the Proof of the PCP Theorem CS640 Computational Complexity Towards the Proof of the PCP Theorem Instructor: Manindra Agrawal Scribe: Ramprasad Saptharishi Last class we saw completed our discussion on expander graphs. We shall now

More information

2 The Fractional Chromatic Gap

2 The Fractional Chromatic Gap C 1 11 2 The Fractional Chromatic Gap As previously noted, for any finite graph. This result follows from the strong duality of linear programs. Since there is no such duality result for infinite linear

More information

On the Max Coloring Problem

On the Max Coloring Problem On the Max Coloring Problem Leah Epstein Asaf Levin May 22, 2010 Abstract We consider max coloring on hereditary graph classes. The problem is defined as follows. Given a graph G = (V, E) and positive

More information

Graph theory - solutions to problem set 1

Graph theory - solutions to problem set 1 Graph theory - solutions to problem set 1 1. (a) Is C n a subgraph of K n? Exercises (b) For what values of n and m is K n,n a subgraph of K m? (c) For what n is C n a subgraph of K n,n? (a) Yes! (you

More information

Adjacent: Two distinct vertices u, v are adjacent if there is an edge with ends u, v. In this case we let uv denote such an edge.

Adjacent: Two distinct vertices u, v are adjacent if there is an edge with ends u, v. In this case we let uv denote such an edge. 1 Graph Basics What is a graph? Graph: a graph G consists of a set of vertices, denoted V (G), a set of edges, denoted E(G), and a relation called incidence so that each edge is incident with either one

More information

Mathematical and Algorithmic Foundations Linear Programming and Matchings

Mathematical and Algorithmic Foundations Linear Programming and Matchings Adavnced Algorithms Lectures Mathematical and Algorithmic Foundations Linear Programming and Matchings Paul G. Spirakis Department of Computer Science University of Patras and Liverpool Paul G. Spirakis

More information

Definition For vertices u, v V (G), the distance from u to v, denoted d(u, v), in G is the length of a shortest u, v-path. 1

Definition For vertices u, v V (G), the distance from u to v, denoted d(u, v), in G is the length of a shortest u, v-path. 1 Graph fundamentals Bipartite graph characterization Lemma. If a graph contains an odd closed walk, then it contains an odd cycle. Proof strategy: Consider a shortest closed odd walk W. If W is not a cycle,

More information

CPSC 536N: Randomized Algorithms Term 2. Lecture 10

CPSC 536N: Randomized Algorithms Term 2. Lecture 10 CPSC 536N: Randomized Algorithms 011-1 Term Prof. Nick Harvey Lecture 10 University of British Columbia In the first lecture we discussed the Max Cut problem, which is NP-complete, and we presented a very

More information

Stanford University CS359G: Graph Partitioning and Expanders Handout 18 Luca Trevisan March 3, 2011

Stanford University CS359G: Graph Partitioning and Expanders Handout 18 Luca Trevisan March 3, 2011 Stanford University CS359G: Graph Partitioning and Expanders Handout 8 Luca Trevisan March 3, 20 Lecture 8 In which we prove properties of expander graphs. Quasirandomness of Expander Graphs Recall that

More information

Testing Hereditary Properties of Non-Expanding Bounded-Degree Graphs

Testing Hereditary Properties of Non-Expanding Bounded-Degree Graphs Electronic Colloquium on Computational Complexity, Report No. 83 (2007) Testing Hereditary Properties of Non-Expanding Bounded-Degree Graphs Artur Czumaj Asaf Shapira Christian Sohler Abstract We study

More information

PACKING DIGRAPHS WITH DIRECTED CLOSED TRAILS

PACKING DIGRAPHS WITH DIRECTED CLOSED TRAILS PACKING DIGRAPHS WITH DIRECTED CLOSED TRAILS PAUL BALISTER Abstract It has been shown [Balister, 2001] that if n is odd and m 1,, m t are integers with m i 3 and t i=1 m i = E(K n) then K n can be decomposed

More information

Finding k-paths in Cycle Free Graphs

Finding k-paths in Cycle Free Graphs Finding k-paths in Cycle Free Graphs Aviv Reznik Under the Supervision of Professor Oded Goldreich Department of Computer Science and Applied Mathematics Weizmann Institute of Science Submitted for the

More information

Lecture 9. Semidefinite programming is linear programming where variables are entries in a positive semidefinite matrix.

Lecture 9. Semidefinite programming is linear programming where variables are entries in a positive semidefinite matrix. CSE525: Randomized Algorithms and Probabilistic Analysis Lecture 9 Lecturer: Anna Karlin Scribe: Sonya Alexandrova and Keith Jia 1 Introduction to semidefinite programming Semidefinite programming is linear

More information

Introduction to Algorithms / Algorithms I Lecturer: Michael Dinitz Topic: Approximation algorithms Date: 11/18/14

Introduction to Algorithms / Algorithms I Lecturer: Michael Dinitz Topic: Approximation algorithms Date: 11/18/14 600.363 Introduction to Algorithms / 600.463 Algorithms I Lecturer: Michael Dinitz Topic: Approximation algorithms Date: 11/18/14 23.1 Introduction We spent last week proving that for certain problems,

More information

Clustering Using Graph Connectivity

Clustering Using Graph Connectivity Clustering Using Graph Connectivity Patrick Williams June 3, 010 1 Introduction It is often desirable to group elements of a set into disjoint subsets, based on the similarity between the elements in the

More information

MapReduce Algorithms. Barna Saha. March 28, 2016

MapReduce Algorithms. Barna Saha. March 28, 2016 MapReduce Algorithms Barna Saha March 28, 2016 Complexity Model for MapReduce Minimum Spanning Tree in MapReduce Computing Dense Subgraph in MapReduce Complexity Model for MapReduce:MRC i Input: finite

More information

1 Unweighted Set Cover

1 Unweighted Set Cover Comp 60: Advanced Algorithms Tufts University, Spring 018 Prof. Lenore Cowen Scribe: Yuelin Liu Lecture 7: Approximation Algorithms: Set Cover and Max Cut 1 Unweighted Set Cover 1.1 Formulations There

More information

1 Linear programming relaxation

1 Linear programming relaxation Cornell University, Fall 2010 CS 6820: Algorithms Lecture notes: Primal-dual min-cost bipartite matching August 27 30 1 Linear programming relaxation Recall that in the bipartite minimum-cost perfect matching

More information

Randomness and Computation March 25, Lecture 5

Randomness and Computation March 25, Lecture 5 0368.463 Randomness and Computation March 25, 2009 Lecturer: Ronitt Rubinfeld Lecture 5 Scribe: Inbal Marhaim, Naama Ben-Aroya Today Uniform generation of DNF satisfying assignments Uniform generation

More information

Testing st-connectivity

Testing st-connectivity Testing st-connectivity Sourav Chakraborty Eldar Fischer Oded Lachish Arie Matsliah Ilan Newman Abstract We continue the study, started in [9], of property testing of graphs in the orientation model. A

More information

Algorithmica 2002 Springer-Verlag New York Inc.

Algorithmica 2002 Springer-Verlag New York Inc. Algorithmica (2002) 32: 302 343 DOI: 10.1007/s00453-001-0078-7 Algorithmica 2002 Springer-Verlag New York Inc. Property Testing in Bounded Degree Graphs 1 O. Goldreich 2 and D. Ron 3 Abstract. We further

More information

Decision Problems. Observation: Many polynomial algorithms. Questions: Can we solve all problems in polynomial time? Answer: No, absolutely not.

Decision Problems. Observation: Many polynomial algorithms. Questions: Can we solve all problems in polynomial time? Answer: No, absolutely not. Decision Problems Observation: Many polynomial algorithms. Questions: Can we solve all problems in polynomial time? Answer: No, absolutely not. Definition: The class of problems that can be solved by polynomial-time

More information

Module 11. Directed Graphs. Contents

Module 11. Directed Graphs. Contents Module 11 Directed Graphs Contents 11.1 Basic concepts......................... 256 Underlying graph of a digraph................ 257 Out-degrees and in-degrees.................. 258 Isomorphism..........................

More information

Coloring 3-Colorable Graphs

Coloring 3-Colorable Graphs Coloring -Colorable Graphs Charles Jin April, 015 1 Introduction Graph coloring in general is an etremely easy-to-understand yet powerful tool. It has wide-ranging applications from register allocation

More information

Matching Algorithms. Proof. If a bipartite graph has a perfect matching, then it is easy to see that the right hand side is a necessary condition.

Matching Algorithms. Proof. If a bipartite graph has a perfect matching, then it is easy to see that the right hand side is a necessary condition. 18.433 Combinatorial Optimization Matching Algorithms September 9,14,16 Lecturer: Santosh Vempala Given a graph G = (V, E), a matching M is a set of edges with the property that no two of the edges have

More information

Matching Theory. Figure 1: Is this graph bipartite?

Matching Theory. Figure 1: Is this graph bipartite? Matching Theory 1 Introduction A matching M of a graph is a subset of E such that no two edges in M share a vertex; edges which have this property are called independent edges. A matching M is said to

More information

Disjoint directed cycles

Disjoint directed cycles Disjoint directed cycles Noga Alon Abstract It is shown that there exists a positive ɛ so that for any integer k, every directed graph with minimum outdegree at least k contains at least ɛk vertex disjoint

More information

A graph is finite if its vertex set and edge set are finite. We call a graph with just one vertex trivial and all other graphs nontrivial.

A graph is finite if its vertex set and edge set are finite. We call a graph with just one vertex trivial and all other graphs nontrivial. 2301-670 Graph theory 1.1 What is a graph? 1 st semester 2550 1 1.1. What is a graph? 1.1.2. Definition. A graph G is a triple (V(G), E(G), ψ G ) consisting of V(G) of vertices, a set E(G), disjoint from

More information

1. Lecture notes on bipartite matching

1. Lecture notes on bipartite matching Massachusetts Institute of Technology 18.453: Combinatorial Optimization Michel X. Goemans February 5, 2017 1. Lecture notes on bipartite matching Matching problems are among the fundamental problems in

More information

The Structure of Bull-Free Perfect Graphs

The Structure of Bull-Free Perfect Graphs The Structure of Bull-Free Perfect Graphs Maria Chudnovsky and Irena Penev Columbia University, New York, NY 10027 USA May 18, 2012 Abstract The bull is a graph consisting of a triangle and two vertex-disjoint

More information

Efficiently decodable insertion/deletion codes for high-noise and high-rate regimes

Efficiently decodable insertion/deletion codes for high-noise and high-rate regimes Efficiently decodable insertion/deletion codes for high-noise and high-rate regimes Venkatesan Guruswami Carnegie Mellon University Pittsburgh, PA 53 Email: guruswami@cmu.edu Ray Li Carnegie Mellon University

More information

Polynomial-Time Approximation Algorithms

Polynomial-Time Approximation Algorithms 6.854 Advanced Algorithms Lecture 20: 10/27/2006 Lecturer: David Karger Scribes: Matt Doherty, John Nham, Sergiy Sidenko, David Schultz Polynomial-Time Approximation Algorithms NP-hard problems are a vast

More information

CSE200: Computability and complexity Interactive proofs

CSE200: Computability and complexity Interactive proofs CSE200: Computability and complexity Interactive proofs Shachar Lovett January 29, 2018 1 What are interactive proofs Think of a prover trying to convince a verifer that a statement is correct. For example,

More information

The strong chromatic number of a graph

The strong chromatic number of a graph The strong chromatic number of a graph Noga Alon Abstract It is shown that there is an absolute constant c with the following property: For any two graphs G 1 = (V, E 1 ) and G 2 = (V, E 2 ) on the same

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

Massachusetts Institute of Technology Department of Electrical Engineering and Computer Science Algorithms For Inference Fall 2014

Massachusetts Institute of Technology Department of Electrical Engineering and Computer Science Algorithms For Inference Fall 2014 Suggested Reading: Massachusetts Institute of Technology Department of Electrical Engineering and Computer Science 6.438 Algorithms For Inference Fall 2014 Probabilistic Modelling and Reasoning: The Junction

More information

Advanced Algorithms Class Notes for Monday, October 23, 2012 Min Ye, Mingfu Shao, and Bernard Moret

Advanced Algorithms Class Notes for Monday, October 23, 2012 Min Ye, Mingfu Shao, and Bernard Moret Advanced Algorithms Class Notes for Monday, October 23, 2012 Min Ye, Mingfu Shao, and Bernard Moret Greedy Algorithms (continued) The best known application where the greedy algorithm is optimal is surely

More information

11.1 Facility Location

11.1 Facility Location CS787: Advanced Algorithms Scribe: Amanda Burton, Leah Kluegel Lecturer: Shuchi Chawla Topic: Facility Location ctd., Linear Programming Date: October 8, 2007 Today we conclude the discussion of local

More information

6 Randomized rounding of semidefinite programs

6 Randomized rounding of semidefinite programs 6 Randomized rounding of semidefinite programs We now turn to a new tool which gives substantially improved performance guarantees for some problems We now show how nonlinear programming relaxations can

More information

CPSC 536N: Randomized Algorithms Term 2. Lecture 5

CPSC 536N: Randomized Algorithms Term 2. Lecture 5 CPSC 536N: Randomized Algorithms 2011-12 Term 2 Prof. Nick Harvey Lecture 5 University of British Columbia In this lecture we continue to discuss applications of randomized algorithms in computer networking.

More information

NP-Hardness. We start by defining types of problem, and then move on to defining the polynomial-time reductions.

NP-Hardness. We start by defining types of problem, and then move on to defining the polynomial-time reductions. CS 787: Advanced Algorithms NP-Hardness Instructor: Dieter van Melkebeek We review the concept of polynomial-time reductions, define various classes of problems including NP-complete, and show that 3-SAT

More information

Using the FGLSS-reduction to Prove Inapproximability Results for Minimum Vertex Cover in Hypergraphs

Using the FGLSS-reduction to Prove Inapproximability Results for Minimum Vertex Cover in Hypergraphs Using the FGLSS-reduction to Prove Inapproximability Results for Minimum Vertex Cover in Hypergraphs Oded Goldreich Abstract. Using known results regarding PCP, we present simple proofs of the inapproximability

More information

Lecture notes on the simplex method September We will present an algorithm to solve linear programs of the form. maximize.

Lecture notes on the simplex method September We will present an algorithm to solve linear programs of the form. maximize. Cornell University, Fall 2017 CS 6820: Algorithms Lecture notes on the simplex method September 2017 1 The Simplex Method We will present an algorithm to solve linear programs of the form maximize subject

More information

by conservation of flow, hence the cancelation. Similarly, we have

by conservation of flow, hence the cancelation. Similarly, we have Chapter 13: Network Flows and Applications Network: directed graph with source S and target T. Non-negative edge weights represent capacities. Assume no edges into S or out of T. (If necessary, we can

More information

Recitation 4: Elimination algorithm, reconstituted graph, triangulation

Recitation 4: Elimination algorithm, reconstituted graph, triangulation Massachusetts Institute of Technology Department of Electrical Engineering and Computer Science 6.438 Algorithms For Inference Fall 2014 Recitation 4: Elimination algorithm, reconstituted graph, triangulation

More information

Lecture 7: Asymmetric K-Center

Lecture 7: Asymmetric K-Center Advanced Approximation Algorithms (CMU 18-854B, Spring 008) Lecture 7: Asymmetric K-Center February 5, 007 Lecturer: Anupam Gupta Scribe: Jeremiah Blocki In this lecture, we will consider the K-center

More information

9.1 Cook-Levin Theorem

9.1 Cook-Levin Theorem CS787: Advanced Algorithms Scribe: Shijin Kong and David Malec Lecturer: Shuchi Chawla Topic: NP-Completeness, Approximation Algorithms Date: 10/1/2007 As we ve already seen in the preceding lecture, two

More information

Local Algorithms for Sparse Spanning Graphs

Local Algorithms for Sparse Spanning Graphs Local Algorithms for Sparse Spanning Graphs Reut Levi Dana Ron Ronitt Rubinfeld Intro slides based on a talk given by Reut Levi Minimum Spanning Graph (Spanning Tree) Local Access to a Minimum Spanning

More information

Part II. Graph Theory. Year

Part II. Graph Theory. Year Part II Year 2017 2016 2015 2014 2013 2012 2011 2010 2009 2008 2007 2006 2005 2017 53 Paper 3, Section II 15H Define the Ramsey numbers R(s, t) for integers s, t 2. Show that R(s, t) exists for all s,

More information

Preimages of Small Geometric Cycles

Preimages of Small Geometric Cycles Preimages of Small Geometric Cycles Sally Cockburn Department of Mathematics Hamilton College, Clinton, NY scockbur@hamilton.edu Abstract A graph G is a homomorphic preimage of another graph H, or equivalently

More information

1 Definition of Reduction

1 Definition of Reduction 1 Definition of Reduction Problem A is reducible, or more technically Turing reducible, to problem B, denoted A B if there a main program M to solve problem A that lacks only a procedure to solve problem

More information

Property Testing of Massively Parametrized problems - A survey

Property Testing of Massively Parametrized problems - A survey Property Testing of Massively Parametrized problems - A survey Ilan Newman April 21, 2010 Abstract We survey here property testing results for the so called massively parametrized model (or problems).

More information

Algorithms for Grid Graphs in the MapReduce Model

Algorithms for Grid Graphs in the MapReduce Model University of Nebraska - Lincoln DigitalCommons@University of Nebraska - Lincoln Computer Science and Engineering: Theses, Dissertations, and Student Research Computer Science and Engineering, Department

More information

ACO Comprehensive Exam October 12 and 13, Computability, Complexity and Algorithms

ACO Comprehensive Exam October 12 and 13, Computability, Complexity and Algorithms 1. Computability, Complexity and Algorithms Given a simple directed graph G = (V, E), a cycle cover is a set of vertex-disjoint directed cycles that cover all vertices of the graph. 1. Show that there

More information

6. Lecture notes on matroid intersection

6. Lecture notes on matroid intersection Massachusetts Institute of Technology 18.453: Combinatorial Optimization Michel X. Goemans May 2, 2017 6. Lecture notes on matroid intersection One nice feature about matroids is that a simple greedy algorithm

More information

1. Lecture notes on bipartite matching February 4th,

1. Lecture notes on bipartite matching February 4th, 1. Lecture notes on bipartite matching February 4th, 2015 6 1.1.1 Hall s Theorem Hall s theorem gives a necessary and sufficient condition for a bipartite graph to have a matching which saturates (or matches)

More information

γ 2 γ 3 γ 1 R 2 (b) a bounded Yin set (a) an unbounded Yin set

γ 2 γ 3 γ 1 R 2 (b) a bounded Yin set (a) an unbounded Yin set γ 1 γ 3 γ γ 3 γ γ 1 R (a) an unbounded Yin set (b) a bounded Yin set Fig..1: Jordan curve representation of a connected Yin set M R. A shaded region represents M and the dashed curves its boundary M that

More information

The Rainbow Connection of a Graph Is (at Most) Reciprocal to Its Minimum Degree

The Rainbow Connection of a Graph Is (at Most) Reciprocal to Its Minimum Degree The Rainbow Connection of a Graph Is (at Most) Reciprocal to Its Minimum Degree Michael Krivelevich 1 and Raphael Yuster 2 1 SCHOOL OF MATHEMATICS, TEL AVIV UNIVERSITY TEL AVIV, ISRAEL E-mail: krivelev@post.tau.ac.il

More information

Property Testing for Sparse Graphs: Structural graph theory meets Property testing

Property Testing for Sparse Graphs: Structural graph theory meets Property testing Property Testing for Sparse Graphs: Structural graph theory meets Property testing Ken-ichi Kawarabayashi National Institute of Informatics(NII)& JST, ERATO, Kawarabayashi Large Graph Project Joint work

More information

Lecture 10 October 7, 2014

Lecture 10 October 7, 2014 6.890: Algorithmic Lower Bounds: Fun With Hardness Proofs Fall 2014 Lecture 10 October 7, 2014 Prof. Erik Demaine Scribes: Fermi Ma, Asa Oines, Mikhail Rudoy, Erik Waingarten Overview This lecture begins

More information

Superconcentrators of depth 2 and 3; odd levels help (rarely)

Superconcentrators of depth 2 and 3; odd levels help (rarely) Superconcentrators of depth 2 and 3; odd levels help (rarely) Noga Alon Bellcore, Morristown, NJ, 07960, USA and Department of Mathematics Raymond and Beverly Sackler Faculty of Exact Sciences Tel Aviv

More information

Finding a -regular Supergraph of Minimum Order

Finding a -regular Supergraph of Minimum Order Finding a -regular Supergraph of Minimum Order Hans L. Bodlaender a, Richard B. Tan a,b and Jan van Leeuwen a a Department of Computer Science Utrecht University Padualaan 14, 3584 CH Utrecht The Netherlands

More information

CS 598CSC: Approximation Algorithms Lecture date: March 2, 2011 Instructor: Chandra Chekuri

CS 598CSC: Approximation Algorithms Lecture date: March 2, 2011 Instructor: Chandra Chekuri CS 598CSC: Approximation Algorithms Lecture date: March, 011 Instructor: Chandra Chekuri Scribe: CC Local search is a powerful and widely used heuristic method (with various extensions). In this lecture

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

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

Extremal Graph Theory: Turán s Theorem

Extremal Graph Theory: Turán s Theorem Bridgewater State University Virtual Commons - Bridgewater State University Honors Program Theses and Projects Undergraduate Honors Program 5-9-07 Extremal Graph Theory: Turán s Theorem Vincent Vascimini

More information

{ 1} Definitions. 10. Extremal graph theory. Problem definition Paths and cycles Complete subgraphs

{ 1} Definitions. 10. Extremal graph theory. Problem definition Paths and cycles Complete subgraphs Problem definition Paths and cycles Complete subgraphs 10. Extremal graph theory 10.1. Definitions Let us examine the following forbidden subgraph problems: At most how many edges are in a graph of order

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

On the packing chromatic number of some lattices

On the packing chromatic number of some lattices On the packing chromatic number of some lattices Arthur S. Finbow Department of Mathematics and Computing Science Saint Mary s University Halifax, Canada BH C art.finbow@stmarys.ca Douglas F. Rall Department

More information

Sublinear Algorithms January 12, Lecture 2. First, let us look at a couple of simple testers that do not work well.

Sublinear Algorithms January 12, Lecture 2. First, let us look at a couple of simple testers that do not work well. Sublinear Algorithms January 12, 2012 Lecture 2 Lecturer: Sofya Raskhodnikova Scribe(s): Madhav Jha 1 Testing if a List is Sorted Input: Goal: a list x 1,..., x n of arbitrary numbers. an -tester for sortedness

More information

Summary of Raptor Codes

Summary of Raptor Codes Summary of Raptor Codes Tracey Ho October 29, 2003 1 Introduction This summary gives an overview of Raptor Codes, the latest class of codes proposed for reliable multicast in the Digital Fountain model.

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

Lecture Notes on Karger s Min-Cut Algorithm. Eric Vigoda Georgia Institute of Technology Last updated for Advanced Algorithms, Fall 2013.

Lecture Notes on Karger s Min-Cut Algorithm. Eric Vigoda Georgia Institute of Technology Last updated for Advanced Algorithms, Fall 2013. Lecture Notes on Karger s Min-Cut Algorithm. Eric Vigoda Georgia Institute of Technology Last updated for 4540 - Advanced Algorithms, Fall 2013. Today s topic: Karger s min-cut algorithm [3]. Problem Definition

More information

Acyclic Edge Colorings of Graphs

Acyclic Edge Colorings of Graphs Acyclic Edge Colorings of Graphs Noga Alon Ayal Zaks Abstract A proper coloring of the edges of a graph G is called acyclic if there is no 2-colored cycle in G. The acyclic edge chromatic number of G,

More information