The extractive distillation of ethanol using glycerol as entrainer is studied in order to find its optimal design and operating conditions. The optimization is formulated as a mixed integer nonlinear programming (MINLP) problem. The discrete variables determine the number of stages of the columns and their feed stage locations. The continuous variables include the variables of the equilibrium model and operating variables. The solution of the optimization problem is achieved through a two-level strategy that combines stochastic and deterministic algorithms. The result obtained establishes the process that maximizes an economic criterion for the industrial production of bioethanol satisfying the problem constraints.