Real-time Rail Traffic Management Problem (rtRTMP) consists in adjusting train timetables in the presence of disturbances, where pre-defined train timetables become infeasible. It does so by employing appropriate control measures, such as re-timing, re-ordering, and re-routing. To address this issue and assist dispatchers, several optimization strategies have been proposed in the literature [1]. Several connecting trains run throughout the day, allowing passengers to transfer from one train to another. Trains may get delayed when there is a perturbation in the network. A minimum amount of time is allotted for connections called minimum connection time. Reducing the minimum connection time between trains is one option to limit delay propagation. This is often possible thanks to buffers existing in the timetable, provided a minimum connection time is always ensured.

We propose two enhancements to the existing rtRTMP model RECIFE-MILP\cite{pellegrini2014optimal} to allow passenger transfers and limit delay propagation exploiting connections. We aim to minimize the total delay of all the trains in the system. We also examine the impact of the two enhancements on the solution time of the model.