Genetic Algorithm based PID controller for blood pressure and Cardiac Output regulation

The paper presents a study on the control scheme of two physiological states: Cardiac Output (CO) and Mean Arterial Pressure (MAP) using Proportional-Integral-Derivative (PID) controller. A control strategy for CO and MAP regulation is developed by simultaneous infusion of two drugs: Sodium Nitroprusside (SNP) and Dopamine (DPM) as inputs. Since drug infusion is interacting type and sensitivity to drugs differs from patient to patient; so the appropriate controller design is challenging task. Therefore suitable controller is required to improve the patient care and also to minimize clinical expenses. Firstly; for tuning of PID controller gains conventional Ziegler Nichols (ZN) method is used. Later PID gains are optimized using Genetic Algorithm (GA) with aggregation of Settling Time; Overshoot and Integral Square Error (ISE) as an objective function. Simulations are carried out to examine the performance of the controller under sensor noise (band limited white noise) and model parameter variations. The result indicates superior performance of the GA for optimization of the controller parameters under sensor noise and parameter variations as compared to the ZN tuning method.