- Lorente, S., Two-objective optimization and robustness,
Institute of Physics Conference Series, vol. 180
pp. 359 - 368 .
(last updated on 2007/04/06)
In this paper, the issue of optimization with one constraint or several competing or non-competing constraints is addressed through examples. In the design of heat exchanger structures the objectives are minimal thermal resistance and minimal pumping power. These objectives can be met at the same time by improving the flow architecture, for example, by using tree networks. Trees between a circle and its center have been optimized by two methods: the minimization of global flow resistance, and the local minimization of every duct length. The two optimized architectures are similar. Their complexity is an optimization result. The tree-shaped design is robust, and continues to perform at a high level even when it has not been optimized in every detail. More complex and a lot closer to reality are problems where the constraints imposed lead to competing results of the optimization. This statement is illustrated through the design of vertical wall cavities undergoing two kinds of constraints: mechanical and thermal. It is shown that in spite of the existence of competing objectives, an optimal structure can be designed so that both objectives are met.
Machine design;Heat resistance;Mass transfer;Robustness (control systems);Approximation theory;Optimization;