Optimal control of autonomous aircraft with modeling uncertainties is a challenging problem, especially when onboard computational resources are limited, and in presence of modeling uncertainty. A concurrent learning based adaptive-optimal control architecture is presented that is suitable for implementation on resource constrained platforms. Exponential parameter convergence properties of concurrent learning adaptive controllers are leveraged to reduce modeling uncertainty through adaptation. A multiparametric quadratic optimization basedmodel predictive control approach is used to optimally shape the reference command. Since the reference model is preselected in our approach, the optimal solutions for several flight conditions can be generated a-priori. Hence, the optimal control problem does not need to be solved online, significantly reducing the computational burden. Exponentially convergent stability bounds are presented for the entire adaptiveoptimal control architecture. Numerical simulations show significant increase in controller performance under input and state constraints.