SUMMARY
The objective of this paper is to establish a dichotomy - opposing analytical methods (such as Queue Theory) to experimental methods (such as Simulation) and discussing their adequateness to complex operations - set up in the matter of dimensioning the handling capacity of a large brazilian call centers company. The literature related to the application of such methods at call centers is reviewed, and the way the question is treated nowadays by the company is described. Then an experimental approach is suggested to be implemented as an alternative methodology to deal with the issue, instead of the analytical method in use. The results obtained are used to justify the adequacy of the experimental approach to the modern call centers operation, as long as it is possible to have the model closer to reality. The main implication points to a better understanding of the operation achieved with the new approach