Research Highlights

Deterministic Global Optimization

The top plot depicts the graph of a nonlinear function (blue) among with an associated function called a convex underestimator (yellow). Such underestimators are very important in global optimization, as they provide easy-to- work with surfaces for the generation of valid lower bounds to the global minimum. The tightness of the particular underestimator depicted here helps standard, branch-and-bound based, global optimization algorithms to converge to the final solution very quickly and with relatively few iterations.

Return to home page.