Design Optimization

Aim

The course aims at integrating traditional design methodologies with concepts and techniques of modern optimisation theory and practice. In the course the student will learn to create an appropriate mathematical description (a simulation model) of the design problem, to formulate the optimisation problem and finally to use numerical optimisation techniques and computer support tools in order to solve the problem. After the course the student:

• Should be able to describe the function of a set of iterative optimization methods and discuss their suitability to solve different types of problems. Examples are the Complex Method, Genetic Algorithms and and gradient methods (Newton methods).
• Should be able to formulate design problems as mathematical optimization problems, be able to implement them into MATLAB and solve them using numerical optimization methods.
• Should be able to describe how optimization of simulation models can be performed and demonstrate it using MATLAB and Simulink.
• Should be able to analyze the sensitivity of the obtained optimal solutions.
• Should be able to discuss the plausibility of results retained both by yourselves and others.

Content

Optimization in the design process Optimization methods: Genetic Algorithms, the Complex method, gradient based methods Formulation of design problems as optimization problems Formulation of objective functions Optimization based on simulation models Multi-objective optimization Constraints and penalty functions Post optimal analysis Sensitivity analysis to analyze the robustness of the solutions Metamodels to create numerical models from experiments