Design Procedures

Abstract

In establishing an optimal design procedure, the following steps should be taken:

• The design problem is formulated. The design variables are chosen, the constraints and the objective function are defined and an analysis model is introduced. This step is of crucial importance for the solution process. A poor problem formulation might lead to incorrect results and/or prohibitive computational cost. Various formulations have been discussed in Chap. 1.

• The optimization method is selected. One of the methods presented in Chap. 2 might be suitable for the solution process. In general, the reliability and ease of use of the method are more important than its computational efficiency. Since most of the cost of optimization is associated with the exact analysis and derivative calculations, efficiency of the method used to solve the problem is not a major consideration in choosing the method.

• Approximations are introduced. It has been noted that approximations are essential in most practical design problems. Using linking and basis reduction methods, it is possible to reduce the number of independent design variables. Scaling of variables, constraint normalization and constraint deletion techniques (Sect. 1.3.4) are all intended to improve the solution efficiency. Approximate behavior models, discussed in Chap. 3, are often necessary in order to reduce the number of exact analyses during the solution process.

• A design procedure is established. The problem formulation, the chosen optimization method and the approximation concepts are integrated to introduce an effective solution strategy. In this chapter, various design procedures demonstrate the solution methodology.