Multi-criteria optimisation problems occur naturally in many engineering practices. Pareto analysis has proven to be a powerful tool to characterise potentially interesting realisations of a particular engineering problem for design-space exploration. Depending on the optimisation goals, one of the Paretooptimal alternatives will be the optimal realisation. It occurs however, that partial design decisions have to be taken, leaving other aspects of the optimisation problem to be decided at a later stage, and that Pareto-optimal configurations have to be composed (dynamically) from Pareto-optimal configurations of components. Both aspects are not supported by current analysis methods. This paper introduces a novel, algebraic approach to Pareto analysis. The approach allows for describing incremental design decisions and composing sets of Pareto-optimal configurations. The algebra can be used to study the operations on Pareto sets and the efficient computation of Pareto sets and their compositions.
(pdf version of the complete report)
Back to the list of technical reports.