The so-called generative design methods are beginning to emerge in various fields of mechanics. They allow, through an iterative process, to come to solutions that meet an optimal design. The central brick of the method is carried by the calculation of the behavior of the structure under predefined loading scenarios. This brick thus links the parameters describing the structure (input) to the quantities of interest to be optimized (output). It is placed at the center of the iterative device, the progress of which is controlled by an "agent designer". With regard to the objectives to be minimized, the latter orients the state of the input parameters from an analysis of the outputs of interest. The efficiency of the system depends directly on the choice and implementation of optimization methods that must be adapted to the physical problem studied.
It is in this context that the proposed thesis is placed. The aim is to develop and implement a generative design methodology applicable to the optimization of reinforcement in reinforced concrete structures. Considering the strong conservatisms imposed by the current design rules, it will propose an innovative design approach, based on an analysis of non-intrusive multi-criteria and multi-objective optimization techniques and their adaptation and implementation on concrete application cases of reinforced concrete structures.