Subway systems play an important role in passenger transportation in large urban centers due to their high transport capacity, low pollutant content and for being relatively free of traffic congestion. The headway - intervals between trains – is an important variable in service planning because it influences the capacity offered and cost of operation. This work develops a discrete event simulation model applied in the analysis of the structure, capacity and operation of Line 4 of the Rio de Janeiro subway system. Experiments with different headway times are performed in order to measure and evaluate the system performance in meeting the forecasted demand. A mixed integer linear programming model (MIP) is developed in order to find optimal headway values for its operation. The models consider the train capacity constraints, travel times contrainsts and the dynamic passenger demand at each station. Data from the Demand Study for Line 4 (FGV, 2011) are used with the demand forecast for the morning peak period of 2016. The work allows to conclude that the dimensioning of the capacity of the subway Line 4 efficiently meets the projected demand for the year 2016 and that the simulation and optimization models developed can be adapted and used to support the planning and operation of other metro systems.