Engineering design is typically a complex process that involves finding a set of designs satisfying various performance criteria. As a result, optimisation algorithms dealing with only single-objective are not sufficient to deal with many real-life problems. Meanwhile, scientific workflows have been shown to be an effective technology for automating and encapsulating scientific processes. While optimisation algorithms have been integrated into workflow tools, they are generally single-objective. This paper first presents our latest development to incorporate multi-objective optimisation algorithms into scientific workflows. We demonstrate the efficacy of these capabilities with the formulation of a three-objective aerodynamics optimisation problem. We target to improve the aerodynamic characteristics of a typical 2D airfoil profile considering also the laminar-turbulent transition location for more accurate estimation of the total drag. We deploy two different heuristic optimisation algorithms and compare the preliminary results.
原文链接
分享
引用
