One of the longest-standing open problems in computational geometry is to bound the lower envelope of n univariate functions, each pair of which crosses at most s times, for some fixed s. This problem is known to be equivalent to bounding the length of an order-s Davenport-Schinzel sequence, namely a sequence over an n-letter alphabet that avoids alternating subsequences of the form a · · · b · · · a · · · b · · · with length s + 2. These sequences were introduced by Davenport and Schinzel in 1965 to model a certain problem in differential equations and have since been applied to bounding the running times of geometric algorithms, data structures, and the combinatorial complexity of geometric arrangements.Let λ s (n) be the maximum length of an order-s DS sequence over n letters. What is λ s asymptotically? This question has been answered satisfactorily (by Hart and Sharir, Agarwal, Sharir, and Shor, Klazar, and Nivasch) when s is even or s ≤ 3. However, since the work of Agarwal, Sharir, and Shor in the mid-1980s there has been a persistent gap in our understanding of the odd orders.In this work we effectively close the problem by establishing sharp bounds on Davenport-Schinzel sequences of every order s. Our results reveal that, contrary to one's intuition, λ s (n) behaves essentially like λ s−1 (n) when s is odd. This refutes conjectures due to Alon et al. (JACM, 2008) and Nivasch (JACM, 2010). If the sequence corresponding to the lower envelope contained an alternating subsequence abab · · · with length s + 2 then the functions fa and f b must have crossed at least s + 1 times, a contradiction. applications, with a growing number [72,59,9,21,48,65] that are not overtly geometric. 2 In each of these applications some quantity (e.g., running time, combinatorial complexity) is expressed in terms of λ s (n), the maximum length of an order-s DS sequence over an n-letter alphabet. To improve bounds on λ s is, therefore, to improve our understanding of numerous problems in algorithms, data structures, and discrete geometry. Davenport and Schinzel [27] established n 1+o(1) upper bounds on λ s (n) for every order s. In order to properly survey the improvements that followed [26,79,41,73,74,56,4,51, 63] we must define some notation for forbidden sequences and their extremal functions.
Sequence Notation and TerminologyLet |σ| be the length of a sequence σ = (σ(i)) 1≤i≤|σ| and let σ be the size of its alphabet Σ(σ) = {σ(i)}. Two equal length sequences are isomorphic if they are the same up to a renaming of their alphabets. We say σ is a subsequence of σ , written σ ≺ σ , if σ can be obtained by deleting symbols from σ . The predicate σ ≺ σ asserts that σ is isomorphic to a subsequence of σ . If σ ⊀ σ we say σ is σ-free. If P is a set of sequences, P ⊀ σ holds if σ ⊀ σ for every σ ∈ P . The assertion that σ appears in or occurs in or is contained in σ means either σ ≺ σ or σ ≺ σ , which one being clear from context. The projection of a sequence σ onto G ⊆ Σ(σ) is obtained by deleting all non-G symbols from σ. A sequence...