Perspective on complexity measures targetting read-once branching programs
Abstract
A model of computation for which reasonable yet still incomplete lower bounds are known is the read-once branching program. Here variants of complexity measures successful in the study of read-once branching programs are defined and studied. Some new or simpler proofs of known bounds are uncovered. Branching program resources and the new measures are compared extensively. The new variants are developed in part in the hope of tackling read-k branching programs for the tree evaluation problem studied in Cook et al. Other computation problems are studied as well. In particular, a common view of a function studied by Gal and a function studied by Bollig and Wegener leads to the general combinatorics of blocking sets. Technical combinatorial results of independent interest are obtained. New leads towards further progress are discussed. An exponential lower bound for non-deterministic read-k branching programs for the GEN function is also derived, independently from the new measures.
Cite
@article{arxiv.2305.11276,
title = {Perspective on complexity measures targetting read-once branching programs},
author = {Yaqiao Li and Pierre McKenzie},
journal= {arXiv preprint arXiv:2305.11276},
year = {2023}
}
Comments
38 pages, a version without the appendix is submitted to the journal Information and Computation