📝 SummarXiv

Abstract

Real-world networks, whether shaped by evolution or intelligent design, are typically optimized for some functionality while adhering to physical, geometric, or budget constraints. Yet, theoretical and computational tools for identifying such optimal structures remain limited. We develop a gradient-based optimization framework to identify synchrony-optimal weighted networks under a constrained coupling budget. The resulting networks exhibit a combination of counterintuitive features: they are sparse, bipartite, elongated, and extremely monophilic (i.e., the neighbors of any given node are highly similar to one another while differing from the node itself). These computational findings are matched by a "constructive" theory for optimal networks; a nonlinear differential equation identifies the optimal pairing of nodes for edges, whereas a variational principle prescribes optimal node-strength allocation. The resulting dynamics provably lack the traditional synchronization threshold; instead, as the budget exceeds a calculable critical value, the system becomes globally phase-locked, exhibiting critical scaling at the transition. Our results have broad implications for real-world systems including power grids, neuromorphic computing, and coupled oscillator technologies.