A Memetic Approach to the Solution of Constrained Min-Max Problems

This paper proposes a novel memetic algorithm for the solution of constrained min-max problems that derive from the optimal design of complex systems under worst-case conditions. In this context the maximisation of a quantity of interest over the space of uncertain variables is required to identify the worst-case scenario (or worst-case solution under uncertainty). An optimal design vector is then identified such that the worst-case value of the quantity of interest is minimised. In the most general case, both maximisation and minimisation are subject to strict feasibility constraints. The ultimate goal of the minimisation problem is to identify the design solution that is feasible for all possible values of the uncertain parameters.