English：

Under the background of urban renewal, the development of underground space is facingchallenges such as complex environments and cost constraints. The pile⁃anchor⁃bracing combined supportsystem has the advantages of strong deformation control ability of the pile⁃bracing system and fastconstruction speed, low cost, and flexible layout of the pile⁃anchor system. It is more widely used indeep and large foundation excavations. However, its safety and economic balance are restricted bychanges in geotechnical mechanics indicators, water level, live load, construction errors, etc. Thereliability design method was used to calculate the failure probability of the support system based on thedetermined parameter statistical law, which easily leads to design redundancy or risk underestimation. Itis urgent to consider the robustness design method of parameter uncertainty. And with the construction ofcollaborative control to achieve safety and cost balance. A phased robust design method and process of“quantifiable parameter optimization⁃non⁃quantifiable parameter iteration” was proposed. In the firststage, the variation coefficient of water level and live load was fixed. Based on the NSGA⁃II algorithm,the initial optimal support design parameters were obtained by considering the variation coefficient ofquantifiable geotechnical mechanics indices and construction error. In the second stage, the variationcoefficient of iterative water level and live load was analyzed by grey correlation analysis to obtain theoptimal solution that meets the correlation requirements, and the implementation of construction collaborative control measures was fed back to obtain the final optimal support design parameters. Thismethod has been applied to the project of the Zhijiang Cultural Center in Zhejiang Province. After theoptimization of the pile⁃anchor⁃bracing combined support system by the phased robust design method, theconstruction collaborative control measures, such as the joint control of water level and live loadmanagement by the waterproof curtain and precipitation, are implemented to meet the requirements of theoptimal solution of non⁃quantifiable parameters, and the support design parameters that meet the safety,robustness, and economy are obtained.