Global and Convex Optimization in Modeling Environments: Compiler-Based, Excel, and Mathematica Implementations

We present a review of several software products that serve to analyze and solve nonlinear (specifically including global) optimization problems across different hardware and software platforms. The implementations discussed are LGO, as a stand-alone, but compiler-dependent modeling and solver environment; its Excel platform implementation; and MathOptimizer, a native solver package for Mathematica users. The discussion advocates the use of hybrid solution strategies, in the context of practical global optimization. Several non-trivial illustrative applications, taken from the literature or received from users of our software products, are also reviewed.

Citation

Research Report, PCS Inc., August 2002.Submitted for publication. (Invited contribution to: Annals of Operations Research.)

Article

Download

View PDF