The increasing degree of connectivity between vehicles and infrastructure, and the impending deployment of autonomous vehicles (AV) in urban streets, presents unique opportunities and challenges regarding the on-street parking provision for AVs. This study develops a novel simulation- optimisation approach for intelligent curbside management, based on a metaheuristic technique. The hybrid method bal- ances curb lanes for driving or parking, aiming to minimise the average traffic delay. The model is tested using an idealised grid layout with a range of flow rates and parking policies. Results demonstrate delay decreased by 9%-27% from the benchmark case. Additionally, the traffic delay distribution shows the trade- offs between expanding road capacity and minimising traffic demand through curb management, indicating the interplay between curb parking and traffic management in the AV era.