Optimizing actuator locations in active noise control systems using subset selection

This numerical simulation uses multiple linear regression with subset selection to optimize the locations of control actuators in a feedforward active noise control system. By formulating the feedforward control problem as a regression, subset selection can find the actuator locations that provide the best system performance. Subset selection provides benefits over continuous optimization approaches because of its computational efficiency; subset selection can examine systems for which the system response is expensive to compute, such as fluid–structure interaction problems, and can efficiently optimize large numbers of actuators. This paper describes a method for exhaustive-search subset selection, and provides numerical results for a simple active structural-acoustic control problem.