small changes that may help your article.

Hi Kumseok Jung, Your article explains the “Routing Problem” really amazingly well and is very insightful. I don’t have many suggestions to you, except for some nitpicks;

• Typo in “interatively” [Asynchronously and interatively, the nodes exchange]
• You could hyperlink “Temporal Difference Learning”
• How often are the ants/mobile agents sent on their way? Or how often is the calculation of the shortest path done? As the nodes can keep failing, congestion can keep varying depending on how many messages travel through various links?

I also had two questions after reading the article. Maybe, the answers to them could be included in the write-up.

• If a forward ant at a node has already visited all the neighbors of, it has run into a circular path. The ant is destroyed if the time spent in the cycle is higher than the time incurred in coming to this point. However, what happens in the other case? Should the ant not be destroyed anyway? As in, how does the ant get out of the cycle? How does the ant reach the destination using this node?
• Why does the backward ant use the same route as the forward ant? Why does it not use updated information and use the shortest path that it finds through it?

Overall it’s a very thorough article!! Loved it!

Regards, Surbhi.