13,763 Hits in 6.7 sec

Horizontally Scalable Submodular Maximization [article]

Mario Lucic and Olivier Bachem and Morteza Zadimoghaddam and Andreas Krause
2016 arXiv   pre-print
The proposed framework applies to a broad class of algorithms and constraints and provides theoretical guarantees on the approximation factor for any available capacity.  ...  We empirically evaluate the proposed algorithm on a variety of data sets and demonstrate that it achieves performance competitive with the centralized greedy solution.  ...  Acknowledgements We thank the reviewers for their insightful comments. This research was partially supported by ERC StG 307036 and the Zurich Information Security Center.  ... 
arXiv:1605.09619v1 fatcat:nc62eqrmnjgy5g5eluppxhztb4

An Information-Theoretic Characterization of Channels That Die

Lav R. Varshney, Sanjoy K. Mitter, Vivek K Goyal
2012 IEEE Transactions on Information Theory  
We provide a partial ordering of communication channels. A dynamic programming formulation is used to show the structural result that channel state feedback does not improve performance.  ...  We show that communication with arbitrarily small probability of error is not possible.  ...  This can be used to show that the capacity of a channel that dies is zero. Proposition 5: The Shannon capacity, , of a channel that dies is zero.  ... 
doi:10.1109/tit.2012.2199078 fatcat:yjgk4yeraza4rip74wyaf46vla

A Modified OMP Algorithm with Reduced Feedback Overhead for Massive MIMO System

Jagadeesh Chandra Prasad Matta, Acarya Nagarjuna University, Guntur, Andhra Pradesh, India, P Siddiah
2021 Indian Journal of Science and Technology  
The evaluation parameters used in the experimental analysis are Normalized Mean Square Error and channel capacity.  ...  Conventional greedy search algorithms like Matching Pursuit (MP) and Orthogonal Matching Pursuit (OMP) are based on the prior knowledge of the signal sparsity, which enhances overhead.  ...  Conventional Greedy Algorithms The function of a greedy search algorithm is to sequentially find the support of the signal to be reconstructed through an iterative method, and find one or more constituent  ... 
doi:10.17485/ijst/v14i33.1442 fatcat:rspyq36wdfcs5cifc22gueii6q

Evolving heuristically difficult instances of combinatorial problems

Bryant A. Julstrom
2009 Proceedings of the 11th Annual conference on Genetic and evolutionary computation - GECCO '09  
As an example, a genetic algorithm searches for instances of the quadratic knapsack problem that are difficult for a straightforward greedy heuristic.  ...  An evolutionary algorithm can search a space of problem instances for cases that are heuristically difficult.  ...  An instance of the problem the GA addresses is thus defined by the parameters listed in Section 5: the number n of objects, the approximate density p of the value matrix, the objects' weights, the capacity  ... 
doi:10.1145/1569901.1569941 dblp:conf/gecco/Julstrom09 fatcat:tkjwcysavza7vklzsu3ln63gnm

Analysis of a Classification-based Policy Iteration Algorithm

Alessandro Lazaric, Mohammad Ghavamzadeh, Rémi Munos
2010 International Conference on Machine Learning  
The analysis reveals a tradeoff between the estimation and approximation errors in this classification-based policy iteration setting.  ...  Our results state a performance bound in terms of the number of policy improvement steps, the number of rollouts used in each iteration, the capacity of the considered policy space, and a new capacity  ...  Acknowledgments This work was supported by French National Research Agency (ANR) (project EXPLO-RA n • ANR-08-COSI-004).  ... 
dblp:conf/icml/LazaricGM10 fatcat:lkahsrgbsfefjmuvxevxajocd4

Algorithms for the bounded set-up knapsack problem

Laura A. McLay, Sheldon H. Jacobson
2007 Discrete Optimization  
One of the dynamic programming algorithms presented solves BKP with the same time and space bounds of the best known dynamic programming algorithm for BKP.  ...  Moreover, the FPTAS improves the worst-case time bound for obtaining approximate solutions to BKP as compared to using FPTASs designed for BKP or the 0-1 Knapsack Problem. Published by Elsevier B.V.  ...  This research has been supported in part by the National Science Foundation (DMI-0114499) and the Air Force Office of Scientific Research (FA9550-04-1-0110).  ... 
doi:10.1016/j.disopt.2006.11.002 fatcat:zzybpj3orrf5pfzxouf7zusff4

Clonal Selection from First Principles [chapter]

Chris McEwan, Emma Hart
2010 Lecture Notes in Computer Science  
For the latter, clonal selection is often interpreted as an asexual variant of natural selection, and thus, tend to be variations on evolutionary strategies.  ...  We show our algorithm to be highly competitive with well-established approximation and learning techniques; notably for large, high-dimensional problems.  ...  Reproducible Research All source code used to produce the results and figures in this paper is available from∼cs268/mcewan2010principles.tgz.  ... 
doi:10.1007/978-3-642-14547-6_3 fatcat:ultgk7nhrzclfhi4zzf46jwyyu

Channels that die

Lav R. Varshney, Sanjoy K. Mitter, Vivek K Goyal
2009 2009 47th Annual Allerton Conference on Communication, Control, and Computing (Allerton)  
We show that communication with arbitrarily small probability of error is not possible.  ...  Given the possibility of communication systems failing catastrophically, we investigate limits to communicating over channels that fail at random times.  ...  A Greedy Algorithm To try to solve the optimization problem (5), we propose a greedy algorithm that optimizes blocklengths n i one by one.  ... 
doi:10.1109/allerton.2009.5394908 fatcat:6wmbkx3a3vbwrkhae23krvoys4

On zero-error source coding with decoder side information

P. Koulgi, E. Tuncel, S.L. Regunathan, K. Rose
2003 IEEE Transactions on Information Theory  
The minimum zero-error asymptotic rate of transmission is shown to be the complementary graph entropy of an associated graph.  ...  The algorithmic aspects of instantaneous code design are considered next. It is shown that optimal code design is -hard. An optimal code design algorithm is derived.  ...  In particular, the suggestions of one of the reviewers helped to clarify the proof of Lemma 3, and the exposition of Example 1, while the comments of another reviewer resulted in simplification of the  ... 
doi:10.1109/tit.2002.806154 fatcat:kc2twvkwjvf3xooxin54bk2r3m

Greedy Methods For Simultaneous Sparse Approximation

Leila Belmerhnia, David Brie, El-hadi Djermoune
2014 Zenodo  
Publication in the conference proceedings of EUSIPCO, Lisbon, Portugal, 2014  ...  This sparse representation allows a good approximation of a set of sparse signal vectors sharing common non-zeros support.  ...  CONCLUSION We presented extensions of greedy algorithms to solve the simultaneous sparse approximation problem.  ... 
doi:10.5281/zenodo.43996 fatcat:56skaojryvf2vdswhkmod7hb3q

Exact Joint Distribution Analysis of Zero-Forcing V-BLAST Gains with Greedy Ordering [article]

Serdar Ozyurt, Murat Torlak
2013 arXiv   pre-print
We derive the joint probability distribution of zero-forcing (ZF) V-BLAST gains under a greedy selection of decoding order and no error propagation.  ...  Unlike the previous approximated analyses, a mathematical framework is built by applying order statistics rules and an exact closed-form joint probability density function expression for squared layer  ...  CONCLUSION We have presented an exact statistical analysis on zeroforcing V-BLAST algorithm under no error propagation and a greedy selection of decoding order at the receiver.  ... 
arXiv:1309.5836v1 fatcat:r6b3r432nnazbpo7poxe5n7j5y

A new approach to greedy multiuser detection

A. AlRustamani, B.R. Vojcic
2002 IEEE Transactions on Communications  
The complexity of the algorithm is approximately 2 log per bit, where is the number of users.  ...  is achieved by the greedy detection.  ...  The greedy method suggests that one can think of an algorithm which works in stages, considering one input at a time.  ... 
doi:10.1109/tcomm.2002.801493 fatcat:cg5zfu77xbajdiszryqwat6gky

Downlink Scheduling and Power Allocation in Cognitive Femtocell Networks [chapter]

Hesham M. Elmaghraby, Dongrun Qin, Zhi Ding
2015 Lecture Notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering  
We analyze the distributions of signal to interference and noise ratio (SINR) of MBS users and signal to interference ratio (SIR) of FBS users.  ...  We consider resource assignment and power allocation problem in femtocells under channel estimation errors.  ...  Fig. 3 . 3 The Hungarian and Greedy algorithms results. Fig. 4 .Fig. 5 . 45 Results due to zero and non-zero estimation error assumptions. Results according to error assumptions for different β.  ... 
doi:10.1007/978-3-319-24540-9_8 fatcat:zwenu4lrgbhkhiyifronini5py

Greedy, genetic, and greedy genetic algorithms for the quadratic knapsack problem

Bryant A. Julstrom
2005 Proceedings of the 2005 conference on Genetic and evolutionary computation - GECCO '05  
Augmenting an evolutionary algorithm with knowledge of its target problem can yield a more effective algorithm, as this presentation illustrates.  ...  In these problems, an object's value density is the sum of the values associated with it divided by its weight.  ...  Instance WITH GREEDY OPERATORS The performance of the genetic algorithm just described can be improved by introducing into it greedy techniques from the algorithms of Section 5.  ... 
doi:10.1145/1068009.1068111 dblp:conf/gecco/Julstrom05a fatcat:gatoyxfifbbgjokc35dr3s6y3u

Correcting Errors in Linear Measurements and Compressed Sensing of Multiple Sources [article]

Alexander Petukhov, Inna Kozlov
2013 arXiv   pre-print
We present an algorithm for finding sparse solutions of the system of linear equations Φx=y with rectangular matrices Φ of size n× N, where n<N, when measurement vector y is corrupted by a sparse vector  ...  We call our algorithm the ℓ^1-greedy-generous (LGGA) since it combines both greedy and generous strategies in decoding.  ...  The idea of such algorithm relies on the reweighting strategy and some knowledge about an approximate error rate. Algorithm A is based on the greedy strategy.  ... 
arXiv:1304.4191v1 fatcat:pxhd443rq5aedddaqfkfpryoxe
« Previous Showing results 1 — 15 out of 13,763 results