An Original Computational Framework for Optimal Maintenance Scheduling of Road Networks in Resource-Constrained African Contexts

Road network maintenance scheduling in resource-constrained environments such as Nigerian state road agencies must balance deteriorating infrastructure conditions against severely limited maintenance budgets, requiring computationally efficient optimisation frameworks that can identify maintenance priority sequences maximising network-level performance within budget constraints. This study develops an original computational framework for optimal maintenance scheduling of road networks under budget constraints, calibrated for African tropical road management conditions. The framework integrates a calibrated pavement performance prediction model (based on the empirical HDM-4 calibration for West African conditions), a network-level maintenance optimisation algorithm using dynamic programming with rolling horizon planning, and a risk-adjusted performance indicator that accounts for road segment criticality for network connectivity. The framework accommodates the discrete nature of maintenance interventions, uncertainty in deterioration rates, and the non-linear relationship between maintenance timing and life-cycle cost. The framework is applied to the Oyo State road network (4,200 km) using 8 years of network condition monitoring data. Comparison between the optimised maintenance programme and current state agency practice shows that the optimised schedule achieves 22 percent better network average PCI improvement for equivalent budget expenditure over a 10-year planning horizon. Budget sensitivity analysis identifies the minimum effective annual maintenance investment as 0.9 percent of network replacement value. The framework is implemented as an open-source Python tool applicable to any state road agency with network condition data.

Keywords: road maintenance optimisation, network management, resource constraints, African roads, dynamic programming