Abstract
In this paper, we demonstrate the issues of feasibility, stability and performance of constrained finite receding horizon linear quadratic regulator (RHLQR) problems using primal-dual interior point (PDIP) method developed in FORTRAN. Instead of including path constraints, we have chosen sufficiently long horizon to achieve stability with finite horizon cost leading to Lyapunov function. We observed a significant improvement of stability of model predictive control using PDIP over active set method.