The air conditioning systems used in transport vehicles (railway passenger cars) are constrained by two issues: the consumed energy and the occupied space. In order to assess how both restrictions can be fulfilled a parametric analysis and an optimization have been carried out by means of Design of Experiments (DoE) techniques applied to a mathematical model of a real AC system. To evaluate the energy efficiency of the system four parameters have been used: the COP (Coefficient of Performance), the refrigeration power of the system and its dimensionless expression and the effectiveness of the evaporator. The occupied space of the AC system has been characterized through the dimensionless volume of the evaporator. The same compressor has been used in the analysis and the input parameters varied have been the dimensionless evaporator volume and four operating conditions: the evaporator inlet air temperature and mass flow rate, the condenser inlet air temperature and the air temperature increment in the condenser. Through a Central Composite Surface Response design, results show that the dimensionless refrigeration power is the best parameter to assess the energy efficiency because is independent of the operating conditions and only depends on the dimensionless volume of the evaporator: the lower the latter, the higher the former. An optimal dimensionless volume has been identified for the simultaneous optimization of the four output variables of the system that provides an increase of 26% in the dimensionless refrigeration power with respect to the baseline system.