Northwestern Physicists Show Some Disorder Can Make Power Grids and Other Networks More Stable
A new framework published in Science overturns the assumption that networks work best when every part is identical. A moderate amount of variation helps, and too much hurts.
Physicists at Northwestern University have found that a carefully balanced amount of disorder can make networks more stable, reversing a long-standing assumption that systems work best when all their parts are identical.
The study, published in the journal Science, was led by Adilson Motter, the Charles E. and Emma H. Morrison Professor of Physics. The co-first authors are postdoctoral researcher Arthur Montanari and graduate student Pietro Zanin.
The team built a mathematical framework for systems sitting near a stable state. It asks a simple question: after a small disturbance, does the system settle back down or does the disturbance grow? They then compared uniform networks with ones whose parts differ, testing power grids, flocking models, materials and ecological networks.
The result is that variation can help. Disorder can improve stability through differences among the network's nodes or among the links connecting them. Even random differences sometimes beat a completely uniform system. But the effect has limits: a moderate degree of disorder can enhance stability, while too much can destabilize the same system.
"Real systems are rarely uniform," Montanari said. "Birds differ in personalities, neurons vary in shape and even our social relationships can be asymmetric." The paper's point is that those differences may not be noise to be engineered away. In some cases they are part of why the system holds together.
The work builds on earlier findings from Motter's group that diversity and asymmetry can prevent failures in large power grids. The new framework generalizes that idea to many kinds of networks and gives a way to work out when variation helps and how much is the right amount.
The practical payoff is in design. Engineers building power grids or architected materials could tune variation on purpose rather than aiming for perfect uniformity, and ecologists could use the same tools to examine why real ecosystems stay stable despite being made of very different species.
The research was funded by the Army Research Office, the National Science Foundation and the NSF-Simons National Institute for Theory and Mathematics in Biology. The authors say the next step is to test how the framework holds up in specific real-world systems, where the right dose of disorder will depend on the details.
The finding also cuts against an engineering instinct. Grid operators and materials designers have long worked to make components as identical as possible, on the theory that uniformity is predictable. The Northwestern framework suggests that in some cases a deliberately mixed design can be sturdier than a perfectly matched one.
