Computationally Efficient Suboptimal Control design for Impulsive Systems based on Model Predictive Static Programming

Abstract A new suboptimal control design approach for impulsive control system is proposed. It is extension of model predictive static programming for continuous control system. This approach is applicable to finite time problems with terminal constraints. Starting from initial guess control history, control is updated in iterative manner till convergence criteria is met. It is computationally efficient, hence, can be implemented online and it gives closed form control solution when control is unconstrained. Also, it does not require approximation of system dynamics. As an example problem, predator-prey (Lotka Volterra), which is a nonlinear model is considered, and simulation results are shown. System states are driven to its equilibrium point. Here fish and shark harvesting is represented in the form of impulse control. Harvesting is done in 10 equal intervals (9 times a year), whereas plots are shown for two years. It took 1-2 sec to run this algorithm.