Optimization of Mechanical Ventilator Settings

Abstract

In this paper we investigate the use of global optimization algorithms coupled with a validated pulmonary physiology simulator, to determine optimal solutions to the tradeoffs involved in Mechanical Ventilation (MV) therapy. Formulating the problem as a multiobjective, multivariable constrained optimization problem, we use two dissimilar algorithms, one based on evolutionary computation, the other on a pattern search strategy, to compute optimal settings of tidal volume (Vtidal, ml), ventilation rate (breaths min⁻¹), inspiratory to expiratory duty cycle (IE) and positive end expiratory pressure (PEEP, kPa). The resulting settings minimize the values of a number of risk factors for VALI, while maintaining adequate levels of patient oxygenation, and provide valuable insights into the effects of variations in the different settings.