Optimal control in wide-sense stationary continuous-time stochastic models