Aircraft Design Optimization as a Geometric Program Pdf, Recent improvements in convex optimization make it feasible to address certain types of constrained optimization problems reliably and economically. These techniques provide substantial benefits over general nonlinear optimization procedures. Within this informative article, conceptual-stage aircraft layout issues are devised as geometric applications (GPs), which can be a certain sort of convex optimization issue. Modern GP solvers are very fast, even on big difficulties, demand no first suspects or pruning of solver parameters, and assure global optimal answers.
Additionally they yield best dual variables, which encode sensitivity data that’s particularly important in an aircraft layout circumstance. These advantages come at a cost: all objective and constraint functions — both the mathematical models that explain aircraft layout connections — needs to be expressed inside the limited functional kinds of GP. Maybe surprisingly, this limited set of practical forms appears again and again in existing physics-based versions for aircraft programs. Moreover, for versions that can’t be manipulated algebraically to the types required by GP, an individual may use techniques developed in this thesis to match compact GP models that correctly approximate the initial versions.