## Weighted Capacitated, Priority, and Geometric Set Cover via Improved Quasi-Uniform Sampling (2011)

### Cached

### Download Links

Citations: | 3 - 1 self |

### BibTeX

@MISC{Chan11weightedcapacitated,,

author = {Timothy M. Chan and Elyot Grant and Jochen Könemann and Malcolm Sharpe},

title = {Weighted Capacitated, Priority, and Geometric Set Cover via Improved Quasi-Uniform Sampling},

year = {2011}

}

### OpenURL

### Abstract

The minimum-weight set cover problem is widely known to be O(log n)-approximable, with no improvement possible in the general case. We take the approach of exploiting problem structure to achieve better results, by providing a geometry-inspired algorithm whose approximation guarantee depends solely on an instance-specific combinatorial property known as shallow cell complexity (SCC). Roughly speaking, a set cover instance has low SCC if any column-induced submatrix of the corresponding element-set incidence matrix has few distinct rows. By adapting and improving Varadarajan’s recent quasi-uniform random sampling method for weighted geometric covering problems, we obtain strong approximation algorithms for a structurally rich class of weighted covering problems with low SCC. We also show how to derandomize our algorithm. Our main result has several immediate consequences. Among them, we settle an open question of Chakrabarty et al. [8] by showing that weighted instances of the capacitated covering problem with underlying network structure have O(1)-approximations. Additionally, our improvements to Varadarajan’s sampling framework yield several new results for weighted geometric set cover, hitting set, and dominating set problems. In particular, for weighted covering problems exhibiting linear (or near-linear) union complexity, we obtain approximability results agreeing with those known for the unweighted case. For example, we obtain a constant approximation for the weighted disk cover problem, improving upon the 2 O(log ∗ n)-approximation known prior to our work and matching the O(1)approximation known for the unweighted variant.

### Citations

1872 | Randomized Algorithms
- Motwani, Raghavan
- 1995
(Show Context)
Citation Context ...niform sample of the set of original columns, with a certain sampling probability Pi. To produce this sample, we take the subset of all columns S such that XS ∈ [0, PiU). By a well known construction =-=[20, 23]-=-, a b-wise independent sequence X1, . . . , XN ∈ [0, U) can be generated from b truly random integers in [0, U) for a given prime U ≥ N. Deterministically, we can try all O(U b ) possible choices for ... |

972 |
O.: Computational Geometry: Algorithms and Applications
- Berg, Kreveld, et al.
- 1997
(Show Context)
Citation Context ...ion is the complement of a convex polyhedron. We thus obtain O(1)approximations for octants, unit cubes, and halfspaces in R3 . Points and halfspaces can be mapped to halfspaces and points by duality =-=[12]-=-, and so we also get an O(1)-approximation for the weighted hitting set problem for halfspaces in R3 . Disks in R2 can be mapped to halfspaces in R3 by a well-known lifting transformation [12], and so... |

946 |
On the uniform convergence of relative frequencies of events to their probabilities. Theory Probab
- Vapnik, Chervonenkis
- 1971
(Show Context)
Citation Context ... SCC at most ( ) n k . Binary matrices that do not contain the submatrix [0, 1] have SCC at most k + 1. A set cover matrix has SCC at most O(nd ) whenever its underlying set system has VC-codimension =-=[28]-=- at most d (see [17]). Finally, binary network matrices (a subclass of totally unimodular matrices [26]) have SCC O(n) as we show in Lemma 4.1. Our main result shows that set cover instances with smal... |

628 | A threshold of ln n for approximating set cover
- Feige
- 1998
(Show Context)
Citation Context ...: Ax ≥ 1, x ≥ 0, x integer}. The set cover problem is rather well understood in terms of approximability; it admits efficient O(log M) approximation algorithms and this is best possible unless NP = P =-=[15]-=-. Nevertheless, there are many cases where one can get better approximations by exploiting the structure of matrix A. A classical example for this is given by the case where matrix A is totally unimod... |

566 | Voronoi Diagrams - A Survey of a Fundamental Geometric Data Structure
- Aurenhammer
- 1991
(Show Context)
Citation Context ...′}, so that the two disks intersect if and only if pσ is covered by Sσ ′. The union of the Sσ ′’s corresponds to a planar additively weighted Voronoi diagram, which is known to have linear complexity =-=[2]-=-. We immediately obtain the following improvement to Gibson and Pirwani’s result: Corollary 1.5. There is a randomized polynomialtime O(1)-approximation algorithm for weighted dominating set for disks... |

478 | Matroid Theory - Oxley - 1992 |

387 | Applications of random sampling in computational geometry
- Clarkson, Shor
- 1989
(Show Context)
Citation Context ...k a minimum cost collection of these objects to cover all points. The study of geometric covering problems is a vibrant area in itself, and much progress has been made in the last 20 years (e.g., see =-=[1, 6, 9, 10, 13, 11, 29, 30]-=-). Until recently, much of the work in this area has focused on the unweighted setting, where wS = 1 for all sets S. A key insight here links the integrality gap of the canonical linear programming re... |

260 | Epsilon-nets and simplex range queries
- Haussler, Welzl
- 1987
(Show Context)
Citation Context ...L)), for some function g (optis the optimum value of (SC-LP)). Hence, improving bounds on the size of ɛ-nets is a central theme in the area of geometric covering. Clarkson [9] and Haussler and Welzl =-=[17]-=- showed that L/N-nets of size O(N/L · log N/L) exist for fairly general set families, including triangles, rectangles, and disks among others. A few years ago, Clarkson and Varadarajan [11] exhibited ... |

161 | Almost Optimal Set Covers in Finite VC-Dimension
- Brönnimann, Goodrich
- 1995
(Show Context)
Citation Context ...k a minimum cost collection of these objects to cover all points. The study of geometric covering problems is a vibrant area in itself, and much progress has been made in the last 20 years (e.g., see =-=[1, 6, 9, 10, 13, 11, 29, 30]-=-). Until recently, much of the work in this area has focused on the unweighted setting, where wS = 1 for all sets S. A key insight here links the integrality gap of the canonical linear programming re... |

140 |
On the union of Jordan regions and collision-free translational motion amidst polygonal obstacles, Discrete Comput
- Kedem, Livne, et al.
- 1986
(Show Context)
Citation Context ...taining to weightedgeometric set cover for objects with linear or near linear union complexity. It is, for example, well-known that a collection of n disks (or pseudodisks) has union complexity O(n) =-=[21]-=-. Corollary 1.1 thus gives us an O(1) approximation for weighted disk cover, and improving the 2 O(log∗ n) -approximation obtained in [30]: Corollary 1.2. There is a randomized polynomialtime O(1)-app... |

139 | Combinatorial Optimization
- Schrijver
- 2003
(Show Context)
Citation Context ...ertheless, there are many cases where one can get better approximations by exploiting the structure of matrix A. A classical example for this is given by the case where matrix A is totally unimodular =-=[26]-=-, or where A exhibits other useful combinatorial structure (e.g., sparsity [18, 22, 27]). Another relevant example for this paper is when (SC) encodes a geometric covering problem. Here X usually cons... |

102 | Chernoff-Hoeffding Bounds for Applications with Limited Independence - Schmidt, Siegel, et al. - 1994 |

56 |
Approximating covering and packing problems: set cover, vertex cover, independent set, and related problems, in Approximation algorithms for NP-hard problems
- HOCHBAUM
- 1997
(Show Context)
Citation Context ...oiting the structure of matrix A. A classical example for this is given by the case where matrix A is totally unimodular [26], or where A exhibits other useful combinatorial structure (e.g., sparsity =-=[18, 22, 27]-=-). Another relevant example for this paper is when (SC) encodes a geometric covering problem. Here X usually consists of points in some fixed-dimensional Euclidean space, S contains geometric objects ... |

55 | Improved approximation algorithms for geometric set cover
- Clarkson, Varadarajan
(Show Context)
Citation Context ...k a minimum cost collection of these objects to cover all points. The study of geometric covering problems is a vibrant area in itself, and much progress has been made in the last 20 years (e.g., see =-=[1, 6, 9, 10, 13, 11, 29, 30]-=-). Until recently, much of the work in this area has focused on the unweighted setting, where wS = 1 for all sets S. A key insight here links the integrality gap of the canonical linear programming re... |

54 | Voronoi diagrams in higher dimensions under certain polyhedral distance functions, Discrete Comput
- Boissonnat, Sharir, et al.
- 1998
(Show Context)
Citation Context ...ollary 1.3. There is a randomized polynomialtime O(log log ∗ n)-approximation algorithm for weighted fat triangle cover in R 2 . Axis-aligned octants and unit cubes in R3 have linear union complexity =-=[5]-=-. Halfspaces in R3 have linear union complexity, since the union is the complement of a convex polyhedron. We thus obtain O(1)approximations for octants, unit cubes, and halfspaces in R3 . Points and ... |

47 |
On a Set of Almost Deterministic k-Independent Random Variables. The Annals of Probability
- Joffe
- 1974
(Show Context)
Citation Context ...niform sample of the set of original columns, with a certain sampling probability Pi. To produce this sample, we take the subset of all columns S such that XS ∈ [0, PiU). By a well known construction =-=[20, 23]-=-, a b-wise independent sequence X1, . . . , XN ∈ [0, U) can be generated from b truly random integers in [0, U) for a given prime U ≥ N. Deterministically, we can try all O(U b ) possible choices for ... |

46 | Flows and generalized coloring theorems in graphs - Jaeger - 1979 |

43 | Improved Approximation Guarantees for Packing and Covering Integer Programs
- Srinivasan
- 1999
(Show Context)
Citation Context ...oiting the structure of matrix A. A classical example for this is given by the case where matrix A is totally unimodular [26], or where A exhibits other useful combinatorial structure (e.g., sparsity =-=[18, 22, 27]-=-). Another relevant example for this paper is when (SC) encodes a geometric covering problem. Here X usually consists of points in some fixed-dimensional Euclidean space, S contains geometric objects ... |

34 | The geometry of scheduling - Bansal, Pruhs - 2010 |

34 | Uniquely representable combinatorial geometries - Brylawski, Lucas - 1976 |

20 |
Hitting sets when the vc-dimension is small
- Even, Rawitz, et al.
- 2005
(Show Context)
Citation Context |

19 | Approximation algorithms for covering/packing integer programs
- Kolliopoulos, Young
(Show Context)
Citation Context ...oiting the structure of matrix A. A classical example for this is given by the case where matrix A is totally unimodular [26], or where A exhibits other useful combinatorial structure (e.g., sparsity =-=[18, 22, 27]-=-). Another relevant example for this paper is when (SC) encodes a geometric covering problem. Here X usually consists of points in some fixed-dimensional Euclidean space, S contains geometric objects ... |

8 | Epsilon nets and union complexity, in
- Varadarajan
- 2009
(Show Context)
Citation Context |

8 |
Weighted geometric set cover via quasi-uniform sampling
- Varadarajan
- 2010
(Show Context)
Citation Context |

7 | On column-restricted and priority covering integer programs
- Chakrabarty, Grant, et al.
- 2010
(Show Context)
Citation Context ...eighted covering problems with low SCC. We also show how to derandomize our algorithm. Our main result has several immediate consequences. Among them, we settle an open question of Chakrabarty et al. =-=[8]-=- by showing that weighted instances of the capacitated covering problem with underlying network structure have O(1)-approximations. Additionally, our improvements to Varadarajan’s sampling framework y... |

7 | Improved bound for the union of fat triangles
- Aronov, Ezra, et al.
- 2011
(Show Context)
Citation Context ...on algorithm for weighted geometric disk (or pseudodisk) cover in R 2 . Similarly, fat triangles are known to have union complexity O(n · 2α(n) log ∗ n) by a recent result of Ezra, Aronov, and Sharir =-=[14]-=- (with de Berg, these authors have apparently improved the bound to O(n log ∗ n)). Applying Varadarajan’s result readily gives a 2 O(log∗ n) -approximation for the weighted fat triangle cover. Corolla... |

3 |
Small-size ɛ-nets for axis-parallel rectangles and boxes
- Aronov, Ezra, et al.
(Show Context)
Citation Context |

2 | On capacitated set cover problems
- Bansal, Krishnaswamy, et al.
- 2011
(Show Context)
Citation Context ...n result in [8] shows that if the integrality gaps of the two families above are bounded by α and β, then there is an O(α + β) approximation for the given capacitated covering instance. Bansal et al. =-=[3]-=- very recently showed that if the integrality gap bound α of the underlying multicover family is hereditary in the sense that it also holds for row-induced sub-systems, then the integrality gap of the... |

1 |
Algorithms for dominating set in disk graphs: Breaking the log n barrier
- Gibson, Pirwani
- 2010
(Show Context)
Citation Context ...oximation algorithm for weighted set cover for axis-aligned octants and unit cubes in R 3 , and halfspaces in R 3 , as well as for weighted hitting set for disks in R 2 . Recently, Gibson and Pirwani =-=[16]-=- have applied Varadarajan’s technique to the weighted dominating set problem in the intersection graph of a set of disks in R2 . We can map a disk σ with center (a, b) and radius c to a point pσ = (a,... |