Paper detail

Recovering the Period in Shor's Algorithm with Gauss' Algorithm for Lattice Basis Reduction

Shor's algorithm contains a classical post-processing part for which we aim to create an efficient, understandable method aside from continued fractions. Let r be an unknown positive integer. Assume that with some constant probability we obtain random positive integers of the form x=[ N k/r ] where [.] is either the floor or ceiling of the rational number, k is selected uniformly at random from {0,1,...,r-1}, and N is a parameter that can be chosen. The problem of recovering r from such samples occurs precisely in the classical post-processing part of Shor's algorithm. The quantum part (quantum phase estimation) makes it possible to obtain such samples where r is the order of some element a of the unit group of Z_n and n is the number to be factored. Shor showed that the continued fraction algorithm can be used to efficiently recover r, since if N>2r^2 then k/r appears in lowest terms as one of the convergents of x/N due to a standard result on continued fractions. We present here an alternative method for recovering r based on the Gauss algorithm for lattice basis reduction, allowing us to efficiently find the shortest nonzero vector of a lattice generated by two vectors. Our method is about as efficient as the method based on continued fractions, yet it is much easier to understand all the details of why it works.