ABSTRACT
Scientific Officer, Computational Studies Section, Machine Dynamics Division, Bhabha Atomic Research Centre, Trombay, Mumbai, India-400 085, Tel.: +91-22-2559-3611; E-mail: gnsk@barc.gov.in
14.1 INTRODUCTION TO SHAPE OPTIMIZATION
Shape optimization problem is unlike other optimization problems involving in finding the optimal set of parameters. These sets of parameters, which represent the shape of a body should optimize the required objective functions. Thus, the parameters to be optimized are indirectly related to the objective functions, through parameterization and flow solution. The first challenge in shape optimization is to find appropriate parameterization, so that the feasible set P is best represented. The bounds for these parameters are to be specified by the designer, which is sometimes, a difficult task. Ant Colony Optimization (ACO) resolves this difficulty, as it has the capability to reach an optimum, which is outside the designer specified bounds. The shape optimization problems are generally governed by partial differential equation leading to multi-modal objective functions (OF). Multimodal objective functions have numerous local optima. The evolutionary algorithms with ability to search global optimum are best suited for shape optimization problems.
