Bound-constrained global optimization helps answer many practical questions in chemistry, molecular biology, economics. Most of algorithms for solution of global optimization problems are a combination of interval methods and exhuastive search. The efficiency of such algorithms is characterized by their ability to detect and eliminate sub-optimal feasible regions. This ability is increased by availability of a good upper bound on the global minimum. In this paper, we present a symbolic-interval algorithm for calculation of upper bounds in bound-constrained global minimization problems and report the results of some experiments.
Citation
In proceedings of the 1st International Workshop on Global Constrained Optimization and Constraint Satisfaction (Cocos'02) Valbonne - Sophia Antipolis, France, October 2-4, 2002