# Dp 512 en

In model predictive control optimal control problem (OCP) based on measurement/estimation of current state is solved each sampling time then the the result, inputs control is applied to the plant. The OCP can be efficiently represented via quadratic programming (QP). In general, there are two types of QP formulation of MPC. The first is called \textit{dense} where QP has smaller number of optimized variables and no zero entries. The second is called \textit{sparse}, here QP is formulated with larger number of optimized variables but with many zero entries and sparsity structure occurs.In order to control of real-time process applications solution for the QP solved each time should be found very quickly (let say in $ms$ or even $\mu s$). Thus in order to reduce the computational complexity of the QP related to box constrained linear model predictive control (MPC) two novel approximations of MPC are introduced in this work.