In this paper, a novel evolutionary computation technique, the Particle Swarm Optimisation (PSO) is applied to optimal control problems of hybrid autonomous switched systems. In particular, the developed algorithms amount to the computation of the optimal switching instants which are the optimization parameters. The objective is that to minimize a performance index, depending on these instants, over a finite time horizon. We assume that a pre-assigned modal sequence is given and that at each switching instant, a jump in the state space variable may occur and that an additional cost is then associated with it. We demonstrate via numerical examples, the effectiveness of the PSO-based algorithms for such a problem known to be NP-hard and constrained nonlinear optimization one, like many scheduling problems.When compared to results obtained by based-gradient methods, PSO results seem to be very promising without requiring any regularity of the objective function to be minimized.