AI-Enabled Robust Performance-Based Budgeting: A Hybrid Neural–Optimization Approach for Smart Governance

Performance-based budgeting (PBB) remains one of the most critical challenges in modern public expenditure management, particularly as governments transition toward data-driven and intelligent governance systems. Advances in artificial intelligence now allow budgeting processes to move beyond subjective estimates and toward analytical, transparent, and evidence-based decision-making. This study introduces an AI-enabled hybrid framework for PBB that integrates Artificial Neural Networks (ANN) with robust linear optimization to support smart governance and resilient resource allocation. The ANN component learns complex cost–activity relationships from historical organizational data and generates predictive estimates for future costs and cost drivers, which are among the primary sources of uncertainty in budgeting. These ANN-derived uncertainty intervals are then incorporated into a robust optimization model following Soyster and Bertsimas–Sim formulations to ensure both feasibility and optimality under uncertain conditions. The proposed hybrid model provides a transparent and intelligence-driven mechanism for allocating resources in alignment with organizational goals. The framework is validated within a major Iranian bank. Empirical results demonstrate that integrating ANN-based cost estimation with robust optimization significantly improves budget deviation indicators and enhances the accuracy, stability, and resilience of the budgeting process. The study contributes to the evolving field of smart governance by presenting a mathematically rigorous, AI-enhanced, and uncertainty-aware budgeting methodology.