This technique can be extremely helpful in designing solutions to multivariable control problems. It recommends the best "pairing" of controlled variables with manipulated variables and will highlight any problems of interaction. While RGA has fallen into disuse with the advent of multivariable control packages it can still be of benefit. It can identify which pairings can be removed from the controller and replaced with single-input single-output schemes. The condition number of the relative gain matrix will highlight any problems with parallel control variables.