We revisit the so-called sampling and discarding approach used to quantify the probability of constraint violation of a solution to convex scenario programs when some of the original samples are allowed to be discarde...
详细信息
We revisit the so-called sampling and discarding approach used to quantify the probability of constraint violation of a solution to convex scenario programs when some of the original samples are allowed to be discarded. Motivated by two scenario programs that possess analytic solutions and the fact that the existing bound for scenario programs with discarded constraints is not tight, we analyze a removal scheme that consists of a cascade of optimization problems, where, at each step, we remove a superset of the active constraints. By relying on results from compression learning theory, we show that such a removal scheme leads to less conservative bounds for the probability of constraint violation than the existing ones. We also show that the proposed bound is tight by characterizing a class of optimization problems which achieves the given upper bound. The performance improvement of the proposed methodology is illustrated by an example that involves a resource sharing linear program.
One popular approach to access the importance/influence of a group of nodes in a network is based on the notion of centrality. For a given group, its group betweenness centrality is computed, first, by evaluating a ra...
详细信息
One popular approach to access the importance/influence of a group of nodes in a network is based on the notion of centrality. For a given group, its group betweenness centrality is computed, first, by evaluating a ratio of shortest paths between each node pair in a network that are "covered" by at least one node in the considered group, and then summing all these ratios for all node pairs. In this paper we study the problem of finding the most influential (or central) group of nodes (of some predefined size) in a network based on the concept of betweenness centrality. One known approach to solve this problem exactly relies on using a linear mixed-integer programming (linear MIP) model. However, the size of this MIP model (with respect to the number of variables and constraints) is exponential in the worst case as it requires computing all (or almost all) shortest paths in the network. We address this limitation by considering randomized approaches that solve a single linear MIP (or a series of linear MIPs) of a much smaller size(s) by sampling a sufficiently large number of shortest paths. Some probabilistic estimates of the solution quality provided by our approaches are also discussed. Finally, we illustrate the performance of our methods in a computational study.
This paper studies the online pricing problem in which there is a sequence of users who want to buy items from one seller. The single seller has k types of items and each type has limited copies. These users are arriv...
详细信息
This paper studies the online pricing problem in which there is a sequence of users who want to buy items from one seller. The single seller has k types of items and each type has limited copies. These users are arriving one by one at different times and are single-minded. When arriving, each user would announce her interested bundle of items and a non-increasing acceptable price function, which specifies how much she is willing to pay for a certain number of bundles. Upon the arrival of a user, the seller needs to determine immediately the number of bundles to be sold to the user and the price he would charge her. His goal is to maximize the sum of money received from users. When items are indivisible, we show that no deterministic algorithm for the problem could have competitive ratio better than O (hk) where h is the highest unit price that at least some user is willing to pay. Thus we focus on randomized algorithms. We derive a lower bound Omega(logh + root k) on the competitive ratio of any randomized algorithm for solving this problem. Then we give the first competitive randomized algorithm Rp-MULTI which is O (psi(h)root k Delta/delta)-competitive, where Delta and delta are respectively the maximum and minimum copies a type of items can have and psi(h) is a function growing slightly faster than logh. When h is known ahead of time, the ratio is decreased to O ((logh)root k Delta/delta). When items are divisible and can be sold fractionally, we concentrate on designing competitive deterministic algorithms. We give the first competitive deterministic algorithm Dp-MULTI which is O (psi(h)root k Delta/delta)-competitive. When h is known, the ratio can be decreased to O ((logh)root k Delta/delta). We also study randomized algorithms for the problem. (C) 2015 Elsevier B.V. All rights reserved.
Motivated by the distributed binary consensus algorithm inPerron et al. [(2009) UsingThree States for Binary Consensus on Complete Graphs. INFOCOM 2009, IEEE, April, pp. 2527-2535], we propose a distributed algorithm ...
详细信息
Motivated by the distributed binary consensus algorithm inPerron et al. [(2009) UsingThree States for Binary Consensus on Complete Graphs. INFOCOM 2009, IEEE, April, pp. 2527-2535], we propose a distributed algorithm for the multivalued consensus problem. In the multivalued consensus problem, each node initially chooses from one of k available choices and the objective of all nodes is to find the choice which was initially chosen by the majority in a distributed fashion. Although the voter model (e. g. Hassin, Y. and Peleg, D. (2002) Distributed probabilistic polling and applications to proportionate agreement. Inf. Comput., 171, 248-268) can be used to find a consensus on multiple choices, it only guarantees consensus and not the consensus on the majority. We derive the time of convergence and an upper bound for the probability of error of our proposed algorithm which shows that, similar to Perron et al. [(2009) Using Three States for Binary Consensus on Complete Graphs. INFOCOM 2009, IEEE, pp. 2527-2535], having an additional state would result in significant improvement of both the convergence time and the probability of error for complete graphs. We also show that our algorithm could be used in Erd"s-Renyi and regular graphs by simulations.
A radio network (RN for short) is a distributed system with no central arbiter, consisting of n radio transceivers, henceforth referred to as stations. We assume that the stations run on batteries and expends power wh...
详细信息
A radio network (RN for short) is a distributed system with no central arbiter, consisting of n radio transceivers, henceforth referred to as stations. We assume that the stations run on batteries and expends power while broadcasting/receiving a data packet. Thus, the most important measure to evaluate protocols on the radio network is the number of awake time slots, in which a station is broadcasting/receiving a data packet. We also assume that the stations are identical and have no unique ID number, and no station knows the number n of the stations. For given n keys one for each station, the ranking problem asks each station to determine the number of keys in the RN smaller than its own key. The main contribution of this paper is to present an optimal randomized ranking protocol on the k-channel RN. Our protocol solves the ranking problem, with high probability, in O(n/k + log n) time slots with every station being awake for at most O(log n) time slots. We also prove that any randomized ranking protocol is required to run in expected Omega(n/k + log n) time slots with at least one station being awake for expected Q(Iog n) time slots. Therefore, our ranking protocol is optimal.
作者:
Agarwal, PKSharir, MDuke Univ
Dept Comp Sci Ctr Geometr Comp Durham NC 27708 USA Tel Aviv Univ
Sch Math Sci IL-69978 Tel Aviv Israel NYU
Courant Inst Math Sci New York NY 10012 USA
We review the recent progress in the design of efficient algorithms for various problems in geometric optimization. We present several techniques used to attack these problems, such as parametric searching, geometric ...
详细信息
We review the recent progress in the design of efficient algorithms for various problems in geometric optimization. We present several techniques used to attack these problems, such as parametric searching, geometric alternatives to parametric searching, prune-and-search techniques for linear programming and related problems, and LP-type problems and their efficient solution. We then describe a wide range of applications of these and other techniques to numerous problems in geometric optimization, including facility location, proximity problems, statistical estimators and metrology, placement and intersection of polygons and polyhedra, and ray shooting and other query-type problems.
This paper addresses the design of robust track-following dynamic output feedback controller for hard disk drives (HDDs) in face of parameter uncertainties which can enter into problem description in a possibly non-li...
详细信息
This paper addresses the design of robust track-following dynamic output feedback controller for hard disk drives (HDDs) in face of parameter uncertainties which can enter into problem description in a possibly non-linear way. The design is performed in a probabilistic framework where the uncertain parameters are treated as random variables and the design specification is met with a given probability level. In particular, a sequential algorithm based on gradient iteration is employed to find a probabilistic robust feasible solution to the formulated problem. The design procedure is computationally tractable and its computational complexity does not depend on the number of uncertain parameters. Our case study allows natural frequency and damping ratio to vary within 8% and 10% from their nominal values for rigid body and all resonance modes. The designed controller achieves robustness in the presence of these uncertainties. Furthermore, the designed controller is implemented in real time on a commercial HDD. (C) 2014 Elsevier Ltd. All rights reserved.
In this letter, fractional calculus is used to propose the fractional entropy (FE) and the fractional mutual information (FMI) as the new forms of the information measure in a generalized Euclidean metric space. Being...
详细信息
In this letter, fractional calculus is used to propose the fractional entropy (FE) and the fractional mutual information (FMI) as the new forms of the information measure in a generalized Euclidean metric space. Being position-related and causal, FE and FMI are natural extensions and more generalized forms of Shannon entropy and mutual information, respectively. (C) 2012 Elsevier B.V. All rights reserved.
We consider the problem of maximizing a nonnegative submodular set function f : 2(N) -> R+ over a ground set N subject to a variety of packing-type constraints including (multiple) matroid constraints, knapsack con...
详细信息
We consider the problem of maximizing a nonnegative submodular set function f : 2(N) -> R+ over a ground set N subject to a variety of packing-type constraints including (multiple) matroid constraints, knapsack constraints, and their intersections. In this paper we develop a general framework that allows us to derive a number of new results, in particular, when f may be a nonmonotone function. Our algorithms are based on (approximately) maximizing the multilinear extension F of f over a polytope P that represents the constraints, and then effectively rounding the fractional solution. Although this approach has been used quite successfully, it has been limited in some important ways. We overcome these limitations as follows. First, we give constant factor approximation algorithms to maximize F over a downward-closed polytope P described by an efficient separation oracle. Previously this was known only for monotone functions. For nonmonotone functions, a constant factor was known only when the polytope was either the intersection of a fixed number of knapsack constraints or a matroid polytope. Second, we show that contention resolution schemes are an effective way to round a fractional solution, even when f is nonmonotone. In particular, contention resolution schemes for different polytopes can be combined to handle the intersection of different constraints. Via linear programming duality we show that a contention resolution scheme for a constraint is related to the correlation gap of weighted rank functions of the constraint. This leads to an optimal contention resolution scheme for the matroid polytope. Our results provide a broadly applicable framework for maximizing linear and submodular functions subject to independence constraints. We give several illustrative examples. Contention resolution schemes may find other applications.
A stochastic approximation algorithm is a recursive procedure to find the solution to an unknown nonlinear equation via noisy measurements. In this paper, we present a stopping rule for a stochastic approximation. We ...
详细信息
A stochastic approximation algorithm is a recursive procedure to find the solution to an unknown nonlinear equation via noisy measurements. In this paper, we present a stopping rule for a stochastic approximation. We show that there is a high probability that the distance between the exact solution and the candidate solution is less than a specified tolerance level when the stochastic approximation stops according to our stopping rule. Furthermore, the number of recursions required by the stopping rule is a polynomial function of the problem size. (C) 2015 Elsevier Ltd. All rights reserved.
暂无评论