Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

IEEE International Symposium on Information Theory (ISIT)

ML decoding via mixed-integer adaptive linear programming

Stark Christiaan Draper, Jonathan S. Yedidia, Yige Wang

Published Jun 1, 2007• 43 citations
Publisher

Abstract

Linear programming (LP) decoding was introduced by Feldman et al. (IEEE Trans. Inform. Theory Mar. 2005) as a novel way to decode binary low-density parity-check codes. Taghavi and Siegel (Proc. ISIT 2006) describe a computationally simplified decoding approach they term "adaptive" LP decoding. Adaptive LP decoding starts with a sub-set of the LP constraints, and iteratively adds violated constraints until an optimum of the original LP is found. Usually only a tiny fraction of the original constraints need to be reinstated, leading to huge efficiency gains compared to ordinary LP decoding. Here we describe a modification of the adaptive LP decoder that results in a maximum likelihood (ML) decoder. Whenever the adaptive LP decoder returns a pseudo-codeword rather than a codeword, we add an integer constraint on the least certain symbol of the pseudo-codeword. For certain codes, and especially in the high-SNR (error floor) regime, only a few integer constraints are required to force the resultant mixed-integer LP to the ML solution. We demonstrate that our approach can efficiently achieve the optimal ML decoding performance on a (155,64) LDPC code introduced by Tanner et al.

DOI
10.1109/isit.2007.4557459

Related papers

Adaptive cut generation for improved linear programming decoding of binary linear codes

Xiaojie Zhang, Paul H. Siegel

Linear programming (LP) decoding approximates optimal maximum-likelihood (ML) decoding of a linear block code by relaxing the equivalent ML integer programming (IP) problem into a more easily solved LP problem. The LP problem is defined by a set of linear inequalities derived from the constraints represented by the rows of a parity-check matrix of the code. Adaptive linear programming (ALP) decoding significantly reduces the complexity of LP decoding by iteratively and adaptively adding necessary constraints in a sequence of smaller LP problems. Adaptive introduction of constraints derived from certain additional redundant parity check (RPC) constraints can further improve ALP performance. In this paper, we propose a new and effective algorithm to identify RPCs that produce linear constraints, referred to as “cuts,” that can eliminate non-ML solutions generated by the ALP decoder, often significantly improving the decoder error-rate performance. The cut-finding algorithm is based upon a specific transformation of an initial parity-check matrix of the linear block code. Simulation results for several low-density parity-check codes demonstrate that the modified ALP decoding algorithm significantly narrows the performance gap between LP decoding and ML decoding.

IEEE International Symposium on Information Theory (ISIT)•Jul 1, 2011•3 citations

Multistep linear programming approaches for decoding low-density parity-check codes

Haiyang Liu, Lianrong Ma, Jie Chen

The problem of improving the performance of linear programming (LP) decoding of low-density parity-check (LDPC) codes is considered in this paper. A multistep linear programming (MLP) algorithm was developed for decoding LDPC codes that includes a slight increase in computational complexity. The MLP decoder adaptively adds new constraints which are compatible with a selected check node to refine the results when an error is reported by the original LP decoder. The MLP decoder result is shown to have the maximum-likelihood (ML) certificate property. Simulations with moderate block length LOpe codes suggest that the M LP decoder gives better performance than both the original LP decoder and the conventional sum-product (SP) decoder.

Tsinghua Science & Technology•Oct 1, 2009•2 citations

Adaptive Cut Generation Algorithm for Improved Linear Programming Decoding of Binary Linear Codes

Xiaojie Zhang, Paul H. Siegel

Linear programming (LP) decoding approximates maximum-likelihood (ML) decoding of a linear block code by relaxing the equivalent ML integer programming (IP) problem into a more easily solved LP problem. The LP problem is defined by a set of box constraints together with a set of linear inequalities called “parity inequalities” that are derived from the constraints represented by the rows of a parity-check matrix of the code and can be added iteratively and adaptively. In this paper, we first derive a new necessary condition and a new sufficient condition for a violated parity inequality constraint, or “cut,” at a point in the unit hypercube. Then, we propose a new and effective algorithm to generate parity inequalities derived from certain additional redundant parity check (RPC) constraints that can eliminate pseudocodewords produced by the LP decoder, often significantly improving the decoder error-rate performance. The cut-generating algorithm is based upon a specific transformation of an initial parity-check matrix of the linear block code. We also design two variations of the proposed decoder to make it more efficient when it is combined with the new cut-generating algorithm. Simulation results for several low-density parity-check (LDPC) codes demonstrate that the proposed decoding algorithms significantly narrow the performance gap between LP decoding and ML decoding.

arXiv:1105.0703•IEEE Transactions on Information Theory•Jun 16, 2012•66 citations

Adaptive linear programming for decoding LDPC codes

Abdullah Sarıduman, Alí Emre Pusane, Z. Caner Taşkın

When linear programming is used to decode low-density parity-check (LDPC) codes, the outcome is a codeword or a pseudocodeword that contains fractional symbol values. It is possible to make pseudocodewords infeasible and increase the performance of linear programming decoders by generating redundant parity check equations. In this paper, redundant parity check equations that can eliminate pseudocodewords are searched using an integer programming based optimization approach. We show that the generated parity check equations increase the performance of the adaptive linear programming decoder and that its performance can converge that of a maximum-likelihood decoder.

Signal Processing and Communications Applications Conference (SIU)•Apr 1, 2014•2 citations

Using Linear Programming to Decode Binary Linear Codes

J. Feldman, Martin J. Wainwright, David Ron Karger

A new method is given for performing approximate maximum-likelihood (ML) decoding of an arbitrary binary linear code based on observations received from any discrete memoryless symmetric channel. The decoding algorithm is based on a linear programming (LP) relaxation that is defined by a factor graph or parity-check representation of the code. The resulting "LP decoder" generalizes our previous work on turbo-like codes. A precise combinatorial characterization of when the LP decoder succeeds is provided, based on pseudocodewords associated with the factor graph. Our definition of a pseudocodeword unifies other such notions known for iterative algorithms, including "stopping sets," "irreducible closed walks," "trellis cycles," "deviation sets," and "graph covers." The fractional distance d/sub frac/ of a code is introduced, which is a lower bound on the classical distance. It is shown that the efficient LP decoder will correct up to /spl lceil/d/sub frac//2/spl rceil/-1 errors and that there are codes with d/sub frac/=/spl Omega/(n/sup 1-/spl epsi//). An efficient algorithm to compute the fractional distance is presented. Experimental evidence shows a similar performance on low-density parity-check (LDPC) codes between LP decoding and the min-sum and sum-product algorithms. Methods for tightening the LP relaxation to improve performance are also provided.

IEEE Transactions on Information Theory•Feb 28, 2005•547 citations

Efficient implementation of linear programming decoding

Mohammad Hossein Taghavi, Amin Shokrollahi, Paul H. Siegel

This paper explores ideas for fast linear programming (LP) decoding of low-density parity-check (LDPC) codes. We first propose a modification of adaptive LP decoding, and prove that it performs LP decoding by solving a number of linear programs that contain at most one linear constraint derived from each of the parity-check constraints. Then, as a step toward designing an efficient LP solver that exploits the structure of LDPC codes, we study a sparse interior-point implementation for solving this sequence of linear programs. Since the most complex part of each iteration of the interior-point algorithms is to solve a (usually ill-conditioned) system of linear equations for finding the step direction, we propose a framework for designing preconditioners to be used with the iterative methods for solving these systems. The effectiveness of the proposed approach is demonstrated via both analytical and simulation results.

Allerton Conference on Communication, Control, and Computing•Sep 1, 2008•13 citations
Linear programming (LP) decoding was introduced by Feldman et al. (IEEE Trans. Inform. Theory Mar. 2005) as a novel way to decode binary low-density parity-check codes. Taghavi and Siegel (Proc. ISIT 2006) describe a computationally simplified decoding approach they term "adaptive" LP decoding. Adaptive LP decoding starts with a sub-set of the LP constraints, and iteratively adds violated constraints until an optimum of the original LP is found. Usually only a tiny fraction of the original constraints need to be reinstated, leading to huge efficiency gains compared to ordinary LP decoding. Here we describe a modification of the adaptive LP decoder that results in a maximum likelihood (ML) decoder. Whenever the adaptive LP decoder returns a pseudo-codeword rather than a codeword, we add an integer constraint on the least certain symbol of the pseudo-codeword. For certain codes, and especially in the high-SNR (error floor) regime, only a few integer constraints are required to force the resultant mixed-integer LP to the ML solution. We demonstrate that our approach can efficiently achieve the optimal ML decoding performance on a (155,64) LDPC code introduced by Tanner et al.
Linear programming (LP) decoding approximates optimal maximum-likelihood (ML) decoding of a linear block code by relaxing the equivalent ML integer programming (IP) problem into a more easily solved LP problem. The LP problem is defined by a set of linear inequalities derived from the constraints represented by the rows of a parity-check matrix of the code. Adaptive linear programming (ALP) decoding significantly reduces the complexity of LP decoding by iteratively and adaptively adding necessary constraints in a sequence of smaller LP problems. Adaptive introduction of constraints derived from certain additional redundant parity check (RPC) constraints can further improve ALP performance. In this paper, we propose a new and effective algorithm to identify RPCs that produce linear constraints, referred to as “cuts,” that can eliminate non-ML solutions generated by the ALP decoder, often significantly improving the decoder error-rate performance. The cut-finding algorithm is based upon a specific transformation of an initial parity-check matrix of the linear block code. Simulation results for several low-density parity-check codes demonstrate that the modified ALP decoding algorithm significantly narrows the performance gap between LP decoding and ML decoding.
The problem of improving the performance of linear programming (LP) decoding of low-density parity-check (LDPC) codes is considered in this paper. A multistep linear programming (MLP) algorithm was developed for decoding LDPC codes that includes a slight increase in computational complexity. The MLP decoder adaptively adds new constraints which are compatible with a selected check node to refine the results when an error is reported by the original LP decoder. The MLP decoder result is shown to have the maximum-likelihood (ML) certificate property. Simulations with moderate block length LOpe codes suggest that the M LP decoder gives better performance than both the original LP decoder and the conventional sum-product (SP) decoder.
Linear programming (LP) decoding approximates maximum-likelihood (ML) decoding of a linear block code by relaxing the equivalent ML integer programming (IP) problem into a more easily solved LP problem. The LP problem is defined by a set of box constraints together with a set of linear inequalities called “parity inequalities” that are derived from the constraints represented by the rows of a parity-check matrix of the code and can be added iteratively and adaptively. In this paper, we first derive a new necessary condition and a new sufficient condition for a violated parity inequality constraint, or “cut,” at a point in the unit hypercube. Then, we propose a new and effective algorithm to generate parity inequalities derived from certain additional redundant parity check (RPC) constraints that can eliminate pseudocodewords produced by the LP decoder, often significantly improving the decoder error-rate performance. The cut-generating algorithm is based upon a specific transformation of an initial parity-check matrix of the linear block code. We also design two variations of the proposed decoder to make it more efficient when it is combined with the new cut-generating algorithm. Simulation results for several low-density parity-check (LDPC) codes demonstrate that the proposed decoding algorithms significantly narrow the performance gap between LP decoding and ML decoding.
When linear programming is used to decode low-density parity-check (LDPC) codes, the outcome is a codeword or a pseudocodeword that contains fractional symbol values. It is possible to make pseudocodewords infeasible and increase the performance of linear programming decoders by generating redundant parity check equations. In this paper, redundant parity check equations that can eliminate pseudocodewords are searched using an integer programming based optimization approach. We show that the generated parity check equations increase the performance of the adaptive linear programming decoder and that its performance can converge that of a maximum-likelihood decoder.
A new method is given for performing approximate maximum-likelihood (ML) decoding of an arbitrary binary linear code based on observations received from any discrete memoryless symmetric channel. The decoding algorithm is based on a linear programming (LP) relaxation that is defined by a factor graph or parity-check representation of the code. The resulting "LP decoder" generalizes our previous work on turbo-like codes. A precise combinatorial characterization of when the LP decoder succeeds is provided, based on pseudocodewords associated with the factor graph. Our definition of a pseudocodeword unifies other such notions known for iterative algorithms, including "stopping sets," "irreducible closed walks," "trellis cycles," "deviation sets," and "graph covers." The fractional distance d/sub frac/ of a code is introduced, which is a lower bound on the classical distance. It is shown that the efficient LP decoder will correct up to /spl lceil/d/sub frac//2/spl rceil/-1 errors and that there are codes with d/sub frac/=/spl Omega/(n/sup 1-/spl epsi//). An efficient algorithm to compute the fractional distance is presented. Experimental evidence shows a similar performance on low-density parity-check (LDPC) codes between LP decoding and the min-sum and sum-product algorithms. Methods for tightening the LP relaxation to improve performance are also provided.
This paper explores ideas for fast linear programming (LP) decoding of low-density parity-check (LDPC) codes. We first propose a modification of adaptive LP decoding, and prove that it performs LP decoding by solving a number of linear programs that contain at most one linear constraint derived from each of the parity-check constraints. Then, as a step toward designing an efficient LP solver that exploits the structure of LDPC codes, we study a sparse interior-point implementation for solving this sequence of linear programs. Since the most complex part of each iteration of the interior-point algorithms is to solve a (usually ill-conditioned) system of linear equations for finding the step direction, we propose a framework for designing preconditioners to be used with the iterative methods for solving these systems. The effectiveness of the proposed approach is demonstrated via both analytical and simulation results.