Pointer Chasing with Unlimited Interaction

Abstract

Pointer-chasing is a central problem in two-party communication complexity: given input size $n$ and a parameter $k$, the two players Alice and Bob are given functions $N_A, N_B: [n] \rightarrow [n]$, respectively, and their goal is to compute the value of $p_k$, where $p_0 = 1$, $p_1 = N_A(p_0)$, $p_2 = N_B(p_1) = N_B(N_A(p_0))$, $p_3 = N_A(p_2) = N_A(N_B(N_A(p_0)))$ and so on, applying $N_A$ in even steps and $N_B$ in odd steps, for a total of $k$ steps. It is trivial to solve the problem using $k$ communication rounds, with Alice speaking first, by simply ``chasing the function'' for $k$ steps. Many works have studied the communication complexity of pointer chasing, although the focus has always been on protocols with $k-1$ communication rounds, or with $k$ rounds where Bob (the ``wrong player'') speaks first. Many works have studied this setting giving sometimes tight or near-tight results. In this paper we study the communication complexity of the pointer chasing problem when the interaction between the two players is unlimited, i.e., without any restriction on the number of rounds. Perhaps surprisingly, this question was not studied before, to the best of our knowledge. Our main result is that the trivial $k$-round protocol is nearly tight (even) when the number of rounds is not restricted: we give a lower bound of $\Omega(k \log (n/k))$ on the randomized communication complexity of the pointer chasing problem with unlimited interaction, and a somewhat stronger lower bound of $\Omega(k \log \log{k})$ for protocols with zero error. When combined with prior work, our results also give a nearly-tight bound on the communication complexity of protocols using at most $k-1$ rounds, across all regimes of $k$; for $k > \sqrt{n}$ there was previously a significant gap between the upper and lower bound.