Generation of finite classical groups by pairs of elements with large fixed point spaces. (Q468691)

The authors study `good elements' in finite \(2n\)-dimensional classical groups \(G\): namely \(t\) is a `good element' if \(o(t)\) is divisible by a primitive prime divisor of \(q^n-1\) for the relevant field order \(q\), and \(t\) fixes pointwise an \(n\)-space. The group \(\mathrm{SL}_{2n}(q)\) contains such elements, and they are present in \(\mathrm{SU}_{2n}(q)\), \(\mathrm{Sp}_{2n}(q)\), \(\mathrm{SO}_{2n}^\varepsilon(q)\), only if \(n\) is odd, even, even, respectively. They prove that there is an absolute positive constant \(c\) such that two random conjugates of \(t\) generate \(G\) with probability at least \(c\), if \(G\neq\mathrm{SO}_{2n}^\varepsilon(2)\) and \(G\neq\mathrm{Sp}_{2n}(q)\) with \(q\) even. In the exceptional case \(G=\mathrm{Sp}_{2n}(q)\) with \(q\) even, two conjugates of \(t\) never generate \(G\): in this case they prove that two random conjugates of \(t\) generate a subgroup \(\mathrm{SO}_{2n}^\varepsilon(q)\) with probability at least \(c\). The results underpin analysis of new constructive recognition algorithms for classical groups in even characteristic, which succeed where methods utilising involution centralisers are not available.