1: \begin{abstract}

2: The quantum Fourier transform (QFT) has emerged as the primary tool

3: in quantum algorithms which achieve exponential advantage over

4: classical computation and lies at the heart of the solution to

5: the abelian hidden subgroup problem, of which Shor's celebrated

6: factoring and discrete log algorithms are a special case.

7: We begin by addressing various computational issues surrounding the QFT

8: and give improved parallel circuits for both the QFT over a power

9: of $2$ and the QFT over an arbitrary cyclic group.  These circuits

10: are based on new insight into the relationship between the discrete Fourier

11: transform over different cyclic groups.  We then exploit this

12: insight to extend the class of hidden subgroup problems with

13: efficient quantum solutions.  First we relax the condition

14: that the underlying hidden subgroup function be 

15: distinct on distinct cosets

16: of the subgroup in question 

17: and show that this relaxation can be solved whenever $G$ is

18: a finitely-generated abelian group.  We then extend this reasoning

19: to the hidden cyclic subgroup problem over the reals, showing

20: how to efficiently generate the bits of the period of any

21: sufficiently piecewise-continuous function on $\Re$.   

22: Finally, we show that this problem of period-finding over $\Re$,

23: viewed as an oracle promise problem, is strictly harder than

24: its integral counterpart.  In particular, period-finding over

25: $\Re$ lies outside the complexity class $MA$, a class which contains

26: period-finding over the integers. 

27: 

28: \abstractsignature

29: \end{abstract}

30: