Performance of near-optimal protocols in weak processes

A natural criticism of the optimal protocol of the irreversible work found for weakly driven processes is its experimental difficulty in being implementable due to its singular part. In this work, I explore the possibility of taking its continuous linear part as an acceptable near-optimal protocol. First, I prove that such a solution is the optimal protocol for non-singular admissible functions. I corroborate this result by observing successful comparisons with test protocols on six reasonable examples. Also, extending such analysis, I conclude that the error committed on this near-optimal protocol is considerable compared to the first-order singular approximation solution, except for sudden and slowly-varying processes. A conjecture is made about a general structure of a near-optimal protocol for systems under arbitrarily strong perturbations.