A fuzzy generalized predictive controller to optimal drug dosage therapy of mathematical modeling of HIV

This paper proposes a fuzzy-GPC based on a mathematical model of human immunodeficiency virus (HIV) to determine the drug dosage and control the progression of the illness. For this purpose, a Takagi-Sugeno (TS) fuzzy model is generated to identify the nonlinear behavior of HIV. The parameters of HIV are estimated by the least square error (LSE) estimation method. Moreover, three scenarios are proposed to control HIV. In scenario 1, according to TS fuzzy model, generalized Predictive Control (GPC) is designed for a daily base drug therapy. Scenario 2 and 3 are more practical. In scenario 2, since the biological behavior of patients are different, the variation in the patients biology is taken into account by generating data according to a group of patients with varying parameters in their mathematical model. In senario3, since daily diagnosis of patient’s health is costly, it is assumed that a patient information is available every month, and drug dosage is determined each month. As a result of which, the sample time of the measurement increases to 30 make it a multi-rate system. The result shows that the TS fuzzy models  the mathematical model of HIV very well, and in all scenarios, the proposed controller has a good performance and the number of healthy cells are controlled in acceptable amount.