OPTIMIZATION OF PROCESS PLANT LAYOUT USING GENETIC ALGORITHMS

The layout of chemical facilities is an activity that is largely carri ed out by the human designer. Few methods exist for optimising layout. Difficulties in formulating the problem as a mathematical program ste m from non-convexities as well as from non-differentiable cast functio ns. Recent attempts (Penteado and Chic, 1996) have succeeded in only i mplicitly including land costs, resulting in layouts which can be exce ssively conservative from the point of view of safety. The use of stoc hastic optimisation techniques, though not guaranteed to find the glob al optimum, has proved to be effective in obtaining good, practical so lutions and permits the incorporation of more realistic cost functions and constraints. In this research, experiments were carried out using various genetic algorithm formulations, and the resulting program, wh ich includes a useful graphical interactive component, is presented he re. (C) 1998 Elsevier Science Ltd. All rights reserved.