A knot recognition algorithm (Q4249737)

A number of researchers have applied Monte Carlo techniques to questions of random knotting. These studies seek to understand the knotting statistics of randomly grown knots. Their relevance to the physical sciences is primarily in the field of polymer or biopolymer configurations. These computations typically require generation of random knots followed by characterization of the knot state of each generated configuration. So, obviously, knot recognition algorithms are essential to such computations. This paper discusses a new approach to the problem of knot recognition. A set of Monte Carlo algorithms will be presented that appear capable of recognizing the unknot, of recognizing and factoring composite knots, and finally of establishing the equivalence of any two prime knots. Mansfield (1996) gives a more detailed description. Being Monte Carlo techniques, these algorithms are non-rigorous. However, they are based on the premise that over a sufficiently long Monte Carlo run, the knot in question will, at least once, be found in a recognizable configuration. Therefore, the failure probability can be made arbitrarily small by extending the duration of the computation. For the entire collection see [Zbl 0899.00044].