In this work we integrate a macroscopic and microscopic model to evaluate the evacuation dynamics within a university campus in emergency scenarios. While the macroscopic approach is derived from an optimization model with an underlying network structure, the microscopic model is based on a discrete event simulation. In this way, we propose an iterative process, in which the output from one model is the input of the other one: the movement of the pedestrians through the optimization paths are simulated, resulting in new information about the transit time within the campus, which is used to execute the optimization model again. This process is iterated until a convergence of the optimization is achieved, producing a mesoscopic model that considers bottlenecks, flows direction, pedestrian speed variations, and congestion. The model performance is evaluated in a real instance of the university campus, resulting on an evacuation time that considers pedestrian interactions and more realistic evacuation dynamics in the studied network.