A Multistart Scatter Search Heuristic for Smooth NLP and MINLP Problems

摘要: The algorithm described here, called OptQuest/NLP or OQNLP, is a heuristic designed to find global optima for pure and mixed integer nonlinear problems with many constraints variables, where all problem functions are differentiable respect the continuous variables. It uses OptQuest, commercial implementation of scatter search developed by OptTek Systems, Inc., provide starting points gradient-based local NLP solver. This solver seeks solution from subset these points, holding discrete variables fixed. procedure motivated our desire combine superior accuracy feasibility-seeking behavior solvers optimization abilities OptQuest. Computational results include 144 smooth MINLP due Floudas et al, most both linear constraints, coded in GAMS modeling language. Some quite large optimization, over 100 constraints. Global solutions almost found small number calls, often one two.