In pharmacokinetic (PK) studies one of the main goals is to compare the effects of several drugs that are applied to several individuals. Measurements are taken from each subject, and model parameters are estimated from the collected data. To reduce subject to subject variation cross over design is used and the period effect on the responses are compared to check bioavailability of the treatments. In such pharmacokinetic studies the underlying mechanism produces nonlinear responses. The drug effect mechanism is modeled using a compartmental model very popularly. One usual technique to analyze such data is non-parametric designs that are performed using summary statistics. We have used a parametric design to optimize the number of time points of the design. As the drugs are applied to individuals randomly, so the parameters for each individual are estimated and an individual design for time points have been suggested. Then a nonlinear mixed effects model is fitted using the individual parameter estimates and a sampling design have been suggested for the random subjects using the first order expansion. We demonstrate how to optimize the precision of parameter estimates by finding the best number and allocation of sampling times of the design space by maximizing the determinant of the Fisher Information matrix of the population parameters using D-optimum criteria. It is shown that a reduced number of samples may be taken without significant loss of precision of parameter estimates. It is also shown that the two treatments are bioequivalent.