An integral formulation of the ɛ-problem and a new computational approach to control function optimization

A version of Balakrishnan's ε-technique is developed using an integrated description of the system equations. The associated necessary conditions for optimality of theintegral ε-technique appear as integral equations and lead to computational algorithms which are largely independent of the penalty function characteristics of the ε-technique. The merits and shortcomings of this approach are discussed, alternative solution methods are proposed, and some computational results are presented.