STATISTICAL METHOD TO EVALUATE CONVERGENCE OF NON-LINEAR OPTIMIZATION ALGORITHMS IN CALL CENTERS PROBLEMS

The optimization of Call Centers is not an easy problem to solve, due to the complexity of the mathematical models that derive from the Erlang formulas. This complexity is transferred to optimization models, which in most cases are made up of non-linear and non-differentiable objective functions. As in all areas of Operations Research, solving these problems demands efficient, fast and precise algorithms. Simulation as an experimental tool constitutes an essential environment for the validation of optimization algorithms, especially when there are no well-defined problem repositories with known results metrics that can be compared. This paper describes a strategy that uses stochastic simulation to study the statistical convergence of integer nonlinear optimization algorithms in the study of Call Center problems.