AI-Driven Risk Prediction and Allocation Optimization for Cost Transparency in Public Infrastructure Renovation Projects

Abstract

Due to inappropriate risk distribution among stakeholders, public infrastructure rehabilitation projects typically encounter cost overruns, timetable delays, and financial misallocation. Conventional methods of risk assessment lack the predictive intelligence necessary for clear financial planning and instead rely on qualitative judgment. An artificial intelligence-driven approach for integrated risk prediction, allocation analysis, and cost transparency in public infrastructure rehabilitation projects is presented in this study. Financial risk parameters, such as project budget, timing, contract type, contingency allocation, and stakeholder risk sharing, are modeled using publicly accessible statistics about construction delays and cost overruns. Random Forest, XGBoost, and Logistic Regression are examples of machine learning models that are used to forecast the likelihood of cost overruns and delays. To measure the impact of inappropriate risk distribution on total project costs, a Risk Allocation Impact Index is created. Cross-validation experimental assessment shows lower error and better prediction accuracy when compared to baseline models. In order to improve transparency, maximize risk sharing between public and private players, and minimize unforeseen financial losses in remodeling projects, the suggested framework offers a data-driven decision support tool. From a procurement strategy standpoint, this optimization structure can guide contract selection decisions such as Guaranteed Maximum Price (GMP), unit-price, design-build, or construction management-at-risk models. Aligning allocation weights with stakeholder capability improves financial resilience and reduces downstream claims.