Evolutionary numerical methods applied to minimum weight structural design and cardiac mechanics

In this thesis numerical methods combining evolutionary optimization algorithms and finite element analysis are developed and applied to problems in cardiac mechanics and structural optimization. In the first application finite element analysis is combined with a real encoded genetic algorithm to determine unknown material parameters in constitutive model for ventricular myocardium using an inverse approach. As a second application suppression of global instabilities in minimum-weight Michell structures is investigated. A real encoded genetic algorithm is combined with a finite element eigenvalue buckling analysis for suppressing global buckling associated with minimum-weight structures. A novel finite element based evolutionary algorithm is developed to determine optimal structural topologies for a 2D continuum. Finally, the topology optimization algorithm described above is modified to determine optimal structural topology for design domains composed of materials having different strengths in tension and compression (dual strength materials). Optimal topologies evolve through successive removal and admission of material based on the current strain energy distribution. Illustrative examples demonstrate the ability to determine optimal structural configuration for a dual strength material domain. Optimal topologies obtained using current scheme are shown to agree with minimum-volume layouts generated using analytic methods based on dual strength material optimality criteria.