Basics: Goal

From OptiYummy

At the state of the art for information processing today, a considerable effort is still required to find optimal solutions of technical and scientific problems by computers.

Only a systematically adapted approach for problem processing of the computer aided solution process leads to aimed optimal results with a satisfying probability.

The chapter of theoretical basics shows, how to process the problem solution based on the design task to get the optimal solution by numerical optimization automatically.

• The starting point is a generalized design process. This will be explained at the view point of designers. It is valid for all creatively constitutive processes
• The center point of the thereby required knowledge process is the model. Simulation within a systematically scheduled experiment builds a base for attaining the optimal solution stepwise autoamtically based on the model.
• The actual focus at a simultion is the probabilistics considering variability, uncertainty and randomness. Classical simultion calculates only with exact model parameters as nominal values. But practically, all parameters scatter however in the reality with a probability function. Modern methods for design of experiment afford the consideration of this variability using the classical model calculation.
• As highlight, numerical optimization with a adapted configuration can automate the goal-oriented experiment with the model. Regard to finding optimal solutions, there are different goals. These pass from an ideal nominal optimization over a reliability based optimization till a robust design optimization. The main focus is to consider several objective functions concurrently, which always leads to compromise solutions in practise.