In this paper, we propose a mixed-integer linear programming formulation for tackling a classic problem in railway traffic management. In particular, we deal with perturbed traffic in a complex junction. Here, multiple lines cross allowing train routing and scheduling modifications with respect to the planned timetable. Our formulation exploits alternative routing and scheduling to minimize delay propagation. This formulation allows modelling the infrastructure in terms of track-circuits, that is, with a higher granularity than what is typically done in the literature (where only block sections are modeled). In a thorough experimental analysis, we assess the impact of this granularity. The results indicate that, through the finer granularity, it is possible to achieve a much higher exploitation of the existing capacity.