Problems of robustness for universal coding schemes (Q2487076)

Well-known compression schemes (such as Lempel-Ziv algorithms), that are universal for classes of stationary ergodic sources, are considered in the paper. An important property of these universal coding schemes is that they are asymptotically optimal for the class of all stationary ergodic sources. A problem of robustness of this property under small violations of ergodicity is studied. Algorithms working with constructive objects (that is, integer and rational numbers, or words in a finite alphabet) are under consideration. A theorem is given that asserts that ''optimal compression scheme'', corresponding to the Kolmogorov complexity is non robust in the class of all stationary ergodic sources. As a consequence of this theorem, a result on non robustness of computable universal coding schemes is given. The notion of deficiency of algorithmic randomness is used as a measure of disagreement between data sequence and probability measure. It is proven that universal compression schemes form a large class which is nonrobust in the following sense: If the randomness deficiency grows arbitrarily slowly on initial fragments of an infinitive sequence, then the property of asymptotic optimality of any universal compression algorithm can be violated. Particularly Lempel-Ziv compression algorithms are robust on infinitive sequences generated by ergodic Markov chains when the randomness deficiency of their initial fragments of length \(n\) grows as \(o(n)\).