A new intelligent strategy for optimal design of high dimensional systems

The present paper proposes a new intelligent strategy that is suitable for optimal design of large scale complex systems such as ships and aircrafts. The strategy relies on three key aspects. The first is choosing a suitable compact compressed invertible representation for the high dimensional design variables vector. The second is defining a goal-focused problem-specific objective function. The third is seeking the global optimal of this objective function using a gradual non-greedy optimization procedure. The effectiveness of the proposed case study is demonstrated through a ship design case study. Starting from an initial far from optimal design, the strategy successfully converged to a greatly improved design in only three iterations.