This thesis focuses on research and development of inverse identification methods of material parameters. A particular attention is attributed to the viscoelastic response of collagen-reinforced soft tissues (artery, intervertebral disc, skin, tendon, ligament, etc) submitted to large displacements and large deformations (hyperelasticity). Highly non-linear and anisotropic, biomechanical constitutive laws account for a large number of material parameters. The inverse problem that allows their identification is of high non-linearity and of large dimension. By reason of numerical difficulties related to its resolution with gradient-based methods, we developed two new identification methods labelled GAO (Genetic algorithms & Analytical Optimization) and MMIM (Maximum-Minimum Identification Method).GAO advantageously combines deterministic methods of gradient type with genetic algorithms. Its originality consists in introducing analytical computations for the deterministic part leading to a gain in the speed up and in the convergence of genetic algorithms. This strategy is used in the context of anisotropic hyperelasticity.Regarding MMIM method, it operates according to an identification criterion that is expressed with the infinite norm and uses genetic algorithms. MMIM method identifies parameters of quasi-linear viscoelastic laws. It guarantees a constant viscous response that characterises the insensitivity of soft tissues to strain rate. GAO and MMIM methods successfully identified parameters of arterial wall and intervertebral disc tissues. The properties of these tissues are described in a more general context that exhibits the anatomy, histology and deformation mechanism at different hierarchical levels (nano-scale to milli-scale) of collagen-reinforced soft tissues. This gives understanding of the role of forces in relating structure to function in biology.
