Seyedashraf Omid, Bottacin-Busolin Andrea, Harou Julien J
Department of Mechanical, Aerospace and Civil Engineering, The University of Manchester, Sackville Street, M13 9PL, Manchester, UK; Department of Civil Engineering, Kermanshah University of Technology, Kermanshah, Iran.
Department of Mechanical, Aerospace and Civil Engineering, The University of Manchester, Sackville Street, M13 9PL, Manchester, UK; Department of Industrial Engineering, University of Padova, Via Venezia 1, 35121, Padova, Italy.
J Environ Manage. 2023 Jun 15;336:117689. doi: 10.1016/j.jenvman.2023.117689. Epub 2023 Mar 14.
Multi-objective design approaches can help identify future infrastructure system designs that appropriately balance different engineering, environmental, and other societal goals. Planners benefit from assessing the trade-offs implied by the best-performing infrastructure system solutions. However, a large number of possible efficient system designs, obtained when using multi-objective optimization, can be overwhelming to interpret. This study attempts to aid decision-making in multi-criteria infrastructure system design by reducing the complexity of the identified set of efficient infrastructure designs, i.e., the Pareto-front. A soft clustering algorithm is applied, which identifies similarities between solutions, partitions the front accordingly, and selects a set of representative solutions while preserving the multi-dimensional structure of the solutions on the efficiency frontier. Three post-optimization decision-making metrics are introduced to help quantify the overall performance of the Pareto-optimal designs to further summarize design process outputs for decision-makers. We apply the method to an illustrious urban drainage network case study. Results show how the approach can simplify Pareto-fronts with thousands of solutions into sets of highlighted designs that aid interpreting the trade-offs implied by the best-performing simulated systems.
多目标设计方法有助于确定未来的基础设施系统设计，这些设计能在不同的工程、环境及其他社会目标之间实现恰当的平衡。规划者可通过评估最佳性能基础设施系统解决方案所隐含的权衡来受益。然而，使用多目标优化时获得的大量可能的高效系统设计可能会让人难以解读。本研究试图通过降低已识别的高效基础设施设计集（即帕累托前沿）的复杂性，来辅助多标准基础设施系统设计中的决策。应用了一种软聚类算法，该算法识别解决方案之间的相似性，相应地对前沿进行划分，并选择一组代表性解决方案，同时保留效率前沿上解决方案的多维结构。引入了三个优化后决策指标，以帮助量化帕累托最优设计的整体性能，从而进一步为决策者总结设计过程输出。我们将该方法应用于一个著名的城市排水管网案例研究。结果表明，该方法如何能将具有数千个解决方案的帕累托前沿简化为突出设计集，有助于解读最佳性能模拟系统所隐含的权衡。
