Improved lower bounds on the length of Davenport-Schinzel sequences (Q1119587)

For the longest length of general Davenport-Schinzel sequences, i.e. sequences of n letters with no immediate repetitions and no subsequence of type a..b..a..b..a.. of length \(s+2\), the estimate \(\Omega (n\alpha^ s(n))\) is established, where \(\alpha\) (n) is the inverse of the Ackermann function. This improves earlier bounds. The proof uses generalized path compression schemes.