Evolutionary Algorithm-Based Approximation of the Capacity of Full-Surface Channels

摘要

Ideas of Evolutionary algorithms can be used to improve the Maximum Likelihood (ML) estimate of the full-surface data, this improved estimate is used to compute the channel capacity of a full-surface communication channel, i.e. a noisy two-dimensional ISI channel. Channel capacity is computed as a function of the entropy rate. Using a Shannon-McMillan-Breimann theorem, the problem is further reduced to the computation of the probability associated with the output. This density function is estimated by using the improved ML data obtained.