What is the lure of optimization broadcasting?
Optimization broadcasting is a simple way of exploring pathological hyper-surfaces, the kind of mathematical terrain that requires multiple search methods and multiple starting points. In the early years of nonlinear optimization research, models were crafted to be somewhat regular (convex or concave). But with the blending of arbitrary modeling formulas, box canyons, narrow ridges, singular precipices, saddle points and various other complex topographies emerge. Because of the limitations of single strategy search algorithms, these kinds of surfaces can be daunting. Historically, optimization models have been of far lower fidelity (realism) than operational simulation models. But with the ORE mining approach, higher fidelity optimization models will quickly emerge in all scientific fields. Optimization broadcasting is a simple way of simultaneously unleashing multiple search strategies, like having an army of searchers covering the same terrain (the same model).
