AGENT-AIDED PRELIMINARY AIRCRAFT DESIGN
Preliminary aircraft design is a multi-disciplinary optimization MDO problem. By exploiting interactions between disciplines, MDO enables to find superior design. However, including all disciplines simultaneously increases the complexity of the problem. In this paper, we show how a cooperative multi-agent framework can offer advantages to system decomposition and inter-disciplinary communications. Each agent represents a physical model, is designed separately, and has a local system view based on its neighbours relationships. Then with a cooperative reasoning, agents collaborate to find equilibrium between local optimal solutions. First experimentations find design solutions with a distributed and cooperative decision process. It offers promising perspectives by optimising simultaneously both model organisation and model parameters.