Vivo

Solution

Vivo's Fractal Project is an in-house AI system for planning network expansion and creating transport-network capacity. It ingests geolocation context, site-density patterns, traffic-demand factors, existing hub and centroid locations, fiber distances, topology rules, IP-layer robustness requirements, and reuse opportunities from the network inventory and installed optical infrastructure. Fractal first applies DBSCAN to segment macro-regions with similar node density, then uses the Coral Reef Algorithm (CRA) to organize site clusters, preserve viable existing hubs, create new hubs only when thresholds are exceeded, and define capacity-bounded sub-polygons. Vivo's House of Cats (HOC) algorithm then builds logical adjacency graphs that minimize connections while preserving resilience constraints, and Dijkstra's algorithm selects efficient routes that balance new fiber builds with reuse of existing optical paths. The system serves engineers through intent-based, closed-loop automation that converts high-level requirements into network designs and actions.

Data flow

Fractal takes network node density, geolocation, traffic demand, existing hub locations, fiber-distance data, topology constraints, and current optical infrastructure context from Vivo's network environment. DBSCAN segments regions, CRA forms and optimizes clusters and hubs, HOC builds resilient adjacency graphs, and Dijkstra selects routes. The resulting intent-driven network design is delivered to engineering teams and used to automate creation and planning decisions.