Towards an implementation of the B-H algorithm for recognizing the unknot

Abstract

In their paper `A new algorithm for recognizing the unknot', in Geometry and Topology', 2 (1998) n. 9, 175-220, the first author and Michael Hirsch presented a then new algorithm for recognizing the unknot. The first part of the algorithm required the systematic enumeration of all discs which support a `braid foliation' and are embeddable in 3-space. The boundaries of these `foliated embeddable discs' (FED's) are the collection of all closed braid representatives of the unknot, up to conjugacy, and the second part of the algorithm produces a word in the generators of the braid group which represents the boundary of the previously listed FED's. The third part tests whether a given closed braid is conjugate to the boundary of a FED on the list. In this paper we describe implementations of the first and second parts of the algorithm. We also give some of the data which we obtained. The data suggest that FED's have unexplored and interesting structure. Open questions are interspersed throughout the manuscript. The third part of the algorithm was studied by the first author, H. K. Ko and S. J. Lee in their paper `A new approach to the word and conjugacy problems in the braid groups', in `Advances in Mathematics', 139 (1998), 322-353, and in their preprint `The infimum, supremum and geodesic length of a braid conjugacy class', arXiv:math.GT/0003125, and implemented by S. J. Lee in his preprint `Implementation of an algorithm for solving the conjugacy problem in the braid groups' (sangjin@knot.kaist.ac.kr). At this writing his algorithm is polynomial for n less than or equal to 4 and exponential for n greater than or equal to 5.