ABSTRACT
The MINLP optimization problems of structures are in most cases comprehensive, non-convex and highly non-linear. This paper reports the experience in solving this type of problems by using the Modified Outer-Approximation/ Equality-Relaxation algorithm by Kravanja & Grossmann (1994), see also Kravanja et al. (1998a). The Linked Multilevel Hierarchical Strategy (LMHS) has been developed to accelerate the convergence of the mentioned algorithm. Since the number of discrete alternatives and defined binary 0-1 variables are usually too high for normal solution of the MINLP, a special prescreening procedure is developed to automatically reduce the number of binary variables on a reasonable level.
