On adapting the Super-Efficient Global Optimization solver to handle mixed-variables, with applications in aircraft design

Recent advances in aircraft design and multidisciplinary formulation have led to new workflows involving many disciplines, ranging up to high levels of fidelity codes. However, there is still a lack of information about the different available optimization methods and their efficiency depending on the problem properties. Aeronautics design is especially multidisciplinary and, in this context, some integer or categorical variables could arises such as the number of engine, the shape of the tail of the type of the wingtip devices. To tackle this class of problem, some mixed integer multidisciplinary design optimization algorithms have recently been developed. The framework of this internship is limited to low-dimension (less than 50) mixed integer optimization and its application to aeronautic design. The problem will be treated as an expensive black-box that we try to optimize with as few evaluations as possible.Firstly, we present an overview of the different types of methods that exists to solve these problems. Then, to illustrate this optimization framework, a method based on continuous optimization is presented and was implemented in an open-source library. Finally, a benchmark of functions was proposed to validate the method and a preliminary result was then treated on a reference model of an A320 aircraft design problem.