Core Concepts
The authors derive explicit formulas for the average decoding error probabilities of the random parity-check matrix ensemble under unambiguous decoding, maximum likelihood decoding, and list decoding over the erasure channel. They also compute the error exponents of these average decoding error probabilities and establish a strong concentration result for the unsuccessful decoding probability under unambiguous decoding.
Abstract
The paper studies the decoding error probability of the random parity-check matrix ensemble over the erasure channel. The key points are:
Explicit Formulas for Average Decoding Error Probabilities:
The authors derive explicit formulas for the average unsuccessful decoding probability under list decoding, unambiguous decoding, and maximum likelihood decoding.
These formulas are expressed in terms of the Gaussian q-binomial coefficients and a function ψm(i) defined in the paper.
The formulas show that the random [n, k]q code ensemble behaves slightly better than the random parity-check matrix ensemble in terms of average decoding error probabilities.
Error Exponents:
The authors compute the error exponents of the average decoding error probabilities for the ensemble series {R(1-R)n,n} as n goes to infinity.
The error exponents are identical to those obtained for the random [n, nR]q code ensemble in prior work.
Concentration Result:
For the unambiguous decoding case, the authors establish a strong concentration result, showing that the unsuccessful decoding probability of a random code in the ensemble {R(1-R)n,n} converges to the average unsuccessful decoding probability with high probability as n goes to infinity, under certain conditions on the rate R.
This implies that the unsuccessful decoding probability tends to zero exponentially fast with high probability for the ensemble {R(1-R)n,n} under the specified conditions.
The paper provides a detailed analysis of the decoding error probabilities for the random parity-check matrix ensemble, which is an important code ensemble in coding theory and has wide applications in communication and storage systems.