Exploring Direct Generation of Smart Pareto Sets
Neil Haddock, Christopher A. Mattson, Darren C. Knight · 12th AIAA/ISSMO Multidisciplinary Analysis and Optimization Conference · 2008
This paper explores the possibility of directly generating smart Pareto sets using a series of added constraints. The added constraints force the generated solutions to be significantly different from previously generated Pareto solutions, under the Normal Constraint Method. Two basic strategies are explored in the paper. The first strategy seeks to create a near exact representation of the smart Pareto filter’s region of Practically Insignificant Tradeoff (PIT). While worth considering, the first strategy brings practical difficulties that led to the development of a second method, which is also described in this paper. While both methods can be used to directly generate smart Pareto sets, neither method provides a compelling argument to use it over the originally proposed approach of smart Pareto filtering.