Approaches Based on Genetic Algorithms for Multiobjective Optimization Problems

In engineering, it is often necessary to formulate problems in which there are several criteria or objectives. It is unlikely that a solution that optimizes one of the objectives will be optimal for any of the others. Compromise solutions are therefore sought such that no other solutions are better in any one objective while remaining no worse in the others. These types of problems are known as either multiobjective, multicriteria, or vector optimization problems. The problem addressed in this paper concerns the proposition of different approaches based on Genetic Algorithms to solve multiobjective optimization problems. We use notions about population manipulation and Pareto theory to develop our approaches, and study the Left Ventricle 3D Reconstruction problem from two Angiographics Views to test them.