An evolutionary method is proposed for the constrained optimization of chemical engineering processes. Apart from the classical mutation, crossover and creep (small mutation), it makes use of several novel reproductive operators : shift, smoothing, extrapolation and swapping. An adaptive mutation rate is used to guard against stalling at local peaks. The method was able to solve dynamic optimization problems involving constrained time-dependent vectors, such as those arising in process control and inverse heat transfer. In addition, the method solves reputedly difficult test problems such as Shwefel＇s and Grierwangk＇s functions better than any known previous method.