Mathematicians and computer scientists have been challenged for decades by finding the most efficient way to move items across a network, but when the network has grown exponentially like the Internet, traditional methods prove problematic. A team at MIT has developed a new algorithm that reduces the time to solve these problems.
Maximum-flow algorithms use a representation of a network as a graph with a series of nodes and connecting lines. The MIT team presented a paper that describes a new algorithm that would dramatically reduce the number of processes required to solve max-flow problems. This new algorithm divides each graph into a cluster of well-connected nodes and the paths between the nodes create bottlenecks, resulting in an almost linear algorithm.
