GRADIENTS AND FLOWS: CONTINUOUS OPTIMIZATION APPROACHES TO THE MAXIMUM FLOW PROBLEM

We use the lens of the maximum flow problem, one of the most fundamental problems in algorithmic graph theory, to describe a new framework for design of graph algorithms. At a high level, this framework casts the graph problem at hand as a convex optimization task and then applies to it an appropriate method from the continuous optimization toolkit. We survey how this new approach led to the first in decades progress on the maximum flow problem and then briefly sketch the challenges that still remain.