The goal of this project is the development of an algorithm for real-time adaptive control of drug delivery based on sparse measurements of a patient's responses. It will involve recursive Bayesian parameter estimation, forecasting the expected performance of adaptive control, and the determination of an appropriate sampling schedule. The algorithm will be validated using Monte Carlo simulations of adaptive control based on sparse measurements for several representative clinical applications. This research is important since it could form the basis for new, innovative drug delivery technology that will offer significant advantages over current technology. This system should allow patient response to be controlled more accurately than with model based open-loop control.