Fast enumeration algorithm for words with given constraints on run lengths of ones

We propose an algorithm for enumeration and denumeration of words with given constraints on run lengths of ones ($dklr$-sequences). For large $n$, operation time of the algorithm (per symbol of a sequence) is at most $O(\log^3n\log\log n)$, where $n$ is the length of enumerated words, whereas for the best known methods it is at least $cn$, $c>0$.