In the process of sustainability assessment in civil and urban engineering, traditional linear evaluation methods take a static and oversimplified role. The aim of this study was to evaluate the validity of curve-based adaptive modeling method in comparison with multi-criteria decision-making frameworks and hybrid decision-support systems on detecting variability of sustainability performance indicators on urban infrastructure projects. A novel approach was proposed relying on computational forecasting technology, which builds framework-like decision-support scaffolds by using (multi-criteria) structural equation modelling (SEM-AHP) reconstructions of sustainability indices as benchmarks for the optimization process. Regarding the implementation, the study uses GIS-based visualization software to transfer the policy and environmental performance information to the decision-support platform to further realize the display of the regional sustainability map. The accuracy of the predictive framework was verified by Monte Carlo robustness testing of different policy zones. We find that urban policy models have more attention to computational efficiency and predictive accuracy in the planning process than traditional regression models, indicating that curve-based modeling has effect on the extent of forecast reliability. By this result, the conclusion is that adaptive curve-based forecasting technology and multi- criteria evaluation can effectively monitor the indicators of environmental resilience problems during the design and operation of sustainable civil engineering systems and can further reduce potential policy misalignments. The established hybrid curve-based method provided an effective decision-support protocol for monitoring the sustainability index and reducing the forecasting errors in the urban engineering domain. Through quantitative validation, the development of policy-adaptive frameworks of civil infrastructure design and the in-depth study of curve-based mathematical modelling has practical application value.
Akramovich et al. (Wed,) studied this question.