Purpose <p>An artificial respiratory system provides support to critically ill patients. Optimizing the control of airway pressure in an artificial respiratory system is difficult due to its non-linear characteristics. A PID controller is most widely used for artificial respiratory systems. This paper presents the optimization of a PID controller for controlling airway pressure in an artificial respiratory system based on a Teaching Learning-Based Optimization Algorithm.</p> Methods <p>The artificial respiratory system is modeled mathematically, and a transfer function is derived to design the proposed controller. A comparative study of the proposed TLBO-based controller is done with a Particle Swarm Optimization, PSO-based controller, and a conventional Zeigler-Nichols tuned controller in terms of performance indices. The proposed controller was also tested for robustness by varying lung compliance and leakage resistance from - 20% to + 20% of the nominal value.</p> Results <p>It is observed that the proposed controllers based on the TLBO Algorithm showed better response, in terms of rise time, overshoot, and settling time. The overshoot was reduced to 0% in the case of the TLBO-tuned PID controller, as compared to 10.56% in the PSO-tuned PID controller. Also, in the case of the proposed TLBO-tuned controller rise time and settling time were reduced to 2.30 and 2.98 seconds, from 4.65 and 5.26 seconds in the conventionally tuned Zeigler Nichols-tuned PID controller.</p> Conclusions <p>In this study, simulation results of Particle swarm optimization tuned and teaching learning optimization-based PID controllers on an artificial respiratory system are presented. The results showed that the poor performance of the conventionally tuned PID controller can be improved significantly by the application of TLBO algorithm-tuned controllers.</p>