[GAP Forum] Computing a subgroup of G which does not map i to j

[Alastair Donaldson]

Dear GAP forum

I have the following problem:  I have a permutation group G acting on the 
set 1..n, given by a set of generators.

I've discovered that one of the generators of G maps a certain value i to 
a certain value j, and that this is not suitable for my purposes.

I want a subgroup of G which does not map i to j.  I could throw away all 
generators that map i to j, but I'd probably lose most of the group then.

Is there an efficient way to compute generators for a "large" subgroup of 
G which has no element that maps i to j?

Thanks

Alastair.