Abstract The problem of determining the optimal strategy for complex batch processes is a difficult one. We have developed a method which exploits accumulated knowledge regarding the process. Our approach involves the use of local optimization criteria, derived at present from the known heuristics, or in the future, automatically through a pattern recognition algorithm. The problem can be cast in a more tractable form and the global optimum is obtained by determining the values of the few physical parameters which enter into the definition of the local criteria, employing any one of the available traditional methods. The local criteria employed in our case is an evaluation of the process characteristics for different control scenarios projected forward from that point in time. Results from the application of this methodology are presented for two examples. The first is the well known brachistochrone problem, which serves as a control problem, since the solution is known. The second is an autoclave molding process applied to the manufacture of thick composite laminates.