ON THE GLOBAL AND EFFICIENT SOLUTION OF STOCHASTIC BATCH PLANT-DESIGNPROBLEMS

The presence of uncertainty in product demands of batch plant design f ormulations with fixed structure and continuous equipment sizes transf orms them into large-scale nonconvex nonlinear programs. This paper de scribes recent developments towards the efficient solution of such mat hematical models. Two global optimization algorithms, a specialized GO P algorithm and a reduced space branch and bound algorithm, are presen ted and applied to this class of batch plant design models. It is show n that, by taking advantage of the special structure of the resulting mathematical formulations, encouraging computational results can be ob tained from both algorithms for problem sizes that would otherwise be practically unsolvable with conventional global optimization technique s. An efficient, specialized Gaussian quadrature technique is also des cribed for the case of product demands following normal probability di stribution functions with which reduced model size and improved estima tion of the expected profit integral are achieved. These developments are tested on example problems from the literature covering single bat ch plant configuration with various scheduling policies and flexible c onfigurations with alternative production sequences. (C) 1997 Elsevier Science Ltd.