We consider a generalised impulse control model for controlling a proc
ess governed by a stochastic differential equation. The controller can
only choose a parameter of the probability distribution of the conseq
uence of his control action which is therefore random. We state optima
lity results relating the value function to quasi-variational inequali
ties and a formal optimal stopping problem. We also remark that the va
lue function is a viscosity solution of the quasivariational inequalit
ies which could lead to developments and convergence proofs of numeric
al schemes. Further, we give some explicit examples and an application
in financial mathematics, the optimal control of the exchange rate.