Accurate travel demand forecasting is crucial before finalizing the decision to implement a public–private partnership (PPP). While government needs to determine the government payments (GPs) and level of user fee, private investors need to evaluate financial feasibility. However, travel demand forecasting involves a deep uncertain future, which relies on other exogenous forecasts. Plain reliance on the best available estimates often causes negative consequences. This study implements a many-objective robust decision-making (MORDM) framework to minimize the vulnerabilities and negative effects that deep uncertainty in travel demand forecasts might have on the PPP project. The MORDM framework is established to allow the government and private partner to pursue their objectives and identify the plausible minimum revenue guarantee and level of user fee conditions in the presence of deep uncertainties in travel demand forecasts. Results from a case study on a Korean railway PPP project illustrate that there exists a significant risk of GPs to private partners under the contracted demand incorporated within the current contract conditions. The GPs have been estimated under the current contract conditions, considerably higher than under robust contract conditions. Similarly, the net present value (NPV) under the robust contract conditions is lower than the NPV under the current contract conditions. Unlike previous approaches, the study provides an alternative to prediction-based analysis that could determine contract conditions by minimizing any negative effects caused by deep uncertainty in future outturn travel demand.

在最终决定实施公私合作伙伴关系 (PPP) 之前，准确的出行需求预测至关重要。虽然政府需要确定政府支付 (GP) 和用户费用水平，但私人投资者需要评估财务可行性。然而，出行需求预测涉及一个高度不确定的未来，它依赖于其他外生预测。单纯依赖现有的最佳估计通常会导致负面后果。本研究实施了一个多目标稳健决策 (MORDM) 框架，以最大限度地减少旅行需求预测中的深度不确定性可能对 PPP 项目产生的脆弱性和负面影响。MORDM 框架的建立是为了让政府和私人合作伙伴能够在旅行需求预测存在很大不确定性的情况下追求他们的目标并确定合理的最低收入保证和用户费用条件水平。韩国铁路 PPP 项目的案例研究结果表明，在当前合同条件下的合同需求下，GP 对私人合作伙伴存在重大风险。GP 是根据当前合同条件估算的，远高于稳健合同条件下的估算值。同样，稳健合同条件下的净现值 (NPV) 低于当前合同条件下的 NPV。与以前的方法不同，