Reviewer #1: The authors proposed two simple schemes to maximize the benefit that is the difference between the utilization and queue delay cost. They are designed to choose the best control parameters. The point of novelty is that it is an online method and doesn't need the knowledge of the traffic conditions, specific delay-utilization models, and complex parameter estimation. This is a self-contained paper.The authors derives them detailedly and discuss them with thorough simulations. But it is better to compare the existed schemes to show the profits of their proposed schemes. Otherwise, they can not convince people very well. In some figures, I am confused with the plots. The authors should polish those figures and show them in a better fashion. Furthermore,it is weird for the page 6 line6-7. Also for the section summary, why not remove some parts to the section of simulations as a discussion part. The conclusion part is not clear. Reviewer #2: The paper investigates the relationship between queuing delays and link utilization, and making the optimal choice. This could be much better than the traditional system (AQM) that assumes that "low delay" is ideal. The approach followed in this paper, allows decision making based on the application requirement for delay.