next up previous
Next: 11.3 Pohlig-Helman factor method Up: 11 Algorithms for groups Previous: 11.1 Pollard's

11.2 Shanks' Baby step-Giant step

Now suppose that we know that we are given bounds L1, L2, ... Lr with the assurance that there is a relation $ \prod_{i}^{}$qiai with 0 $ \leq$ ai < Li. (We will see below how such a collection of bounds can be constructed inductively given a bound L on the order of G).

Let h(x) be a collision free function for x $ \in$ G (for example h is a ``hash function''). Let (a1,..., ar) be a sequence with ai an integer not larger than ni = $ \lceil$$ \sqrt{L_i}$$ \rceil$. We compute the terms

($\displaystyle \prod_{i}^{}$giai;a1,..., ar;h($\displaystyle \prod_{i}^{}$giai))

for all ai in this range and store them sorted according the final entry. This allows us to perform a search operation in a time proportional to the logarithm of the length of the list. Now, for each sequence (b1,..., br) where bi is an integer not larger than $ \lfloor$$ \sqrt{L_i}$$ \rfloor$ + 1, we compute the expression h($ \prod_{i}^{}$(gi-ni)bi) and try to find it in the given sorted list. Clearly, if it is found then we have a relation of the form

$\displaystyle \prod_{i}^{}$giai + nibi = 1

By the given assertion on Li, such relation will eventually be found.

One way to approach this algorithm given only a bound on the size of the group G is to go inductively. First find a relation involving g1 alone (in other words find the order of g1) by using L1 = L. For the remainder of the algorithm we put L1 = o(g1) as found in this step. In the second iteration we work with g1 and g2 with L2 = L/L1. Iteratively, we would have computed Li which is the order of gi in the group G/ < g1,..., gi - 1 >. We now work with g1, ..., gi + 1 and take Li + 1 = L/($ \prod_{k=1}^{i}$Li). Using the above algorithm we obtain the order of gi + 1 in the group G/ < g1,..., gi >. For the remainder of the iterations we take Li + 1 to be this order. Thus, at the end we would have found a minimal relation among the gi's rather than just one relation.


next up previous
Next: 11.3 Pohlig-Helman factor method Up: 11 Algorithms for groups Previous: 11.1 Pollard's
Kapil Hari Paranjape 2002-10-20