Network Communication Efficiency¶
A shortest-path-based network measure that assigns pairwise efficiency as reciprocal distance and summarizes global or local potential exchange under a declared graph convention.
Core Idea¶
Network communication efficiency is a shortest-path-based measure of potential exchange in a graph. For distinct nodes i and j, pairwise efficiency is εᵢⱼ = 1/dᵢⱼ, where dᵢⱼ is their shortest-path distance under a declared edge-cost convention. Global efficiency averages these reciprocal distances over node pairs.
If no path exists, distance is treated as infinite and efficiency as zero, allowing disconnected networks to be compared. Local efficiency applies the same principle to neighborhoods—commonly the subgraph induced by a node’s neighbors after that node is removed—to describe local alternative connectivity.
Operationally, Graph and node set define the network being summarized. Distance convention turns edges or weights into path lengths or costs. Reciprocal transform converts shorter distance into greater pair efficiency. Disconnected-pair rule assigns zero when no finite route exists. Aggregation scale selects global pairs or local neighbor subgraphs. Communication assumption treats shortest paths as potential information routes.
Scope of Application¶
The measure is used in network science, neuroscience, transportation, infrastructure, and communication studies to compare integration and local fault tolerance. Normalization against reference graphs may aid comparison but introduces additional model choices.
It is not observed bandwidth, packet-delivery efficiency, latency, reliability, clustering coefficient, or average path length. A high topological value does not prove that real traffic uses shortest paths or that capacity and congestion are favorable. Weighted networks require an explicit conversion from strength or capacity to distance.
Illustrative cases include: An unweighted connected graph uses hop-count shortest paths and averages 1/d across all ordered distinct node pairs. A node’s local efficiency is computed among its neighbors after removal to characterize alternative local connectivity.
Clarity¶
The abstraction separates graph accessibility from operational performance and global from local scale. It makes disconnected pairs explicit rather than deleting them and reveals whether an edge weight is being treated as closeness or cost.
Manages Complexity¶
All-pairs shortest paths are compressed into one global mean, or many neighborhood relations into local summaries. The measure remains sensitive to graph construction, node sampling, weight transformation, and normalization, which must accompany the reported value.
Abstract Reasoning¶
Define nodes, edges, direction, and weights. Convert weights into nonnegative path costs, compute shortest distances, set unreachable-pair efficiency to zero, and average over the declared ordered or unordered pair convention. For local efficiency, construct each required neighbor subgraph consistently. Test sensitivity to thresholding and compare with actual communication data before making performance claims.
Knowledge Transfer¶
The formula transfers across networks, while the meaning of distance does not. A neural connection strength, road travel time, and communication cost require different transformations. Topological findings transfer only when graph construction and routing assumptions match.
Topological accessibility versus actual performance. Short paths need not match congestion, capacity, or protocols. Diagnostic: Is the claim about potential topology or measured traffic?
Relationships to Other Abstractions¶
Current abstraction Network Communication Efficiency Domain-specific
Parents (1) — more general patterns this builds on
-
Network Communication Efficiency is a kind of, conditional Measurement Scale Domain-specific
Supported where the node defines a metric or scale rather than the property alone.
Condition / exception Supported where the node defines a metric or scale rather than the property alone.
Hierarchy path (1) — routes to 1 parentless root
- Network Communication Efficiency → Measurement Scale → Measurement
Neighborhood in Abstraction Space¶
Network Communication Efficiency sits in a sparse region of the domain-specific corpus (88th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.
Family — Unclustered & Miscellaneous (2551 abstractions)
Nearest neighbors
- Diameter (graph theory) — 0.82
- Small-World Routing — 0.82
- Level structure — 0.81
- Hyper-Wiener Index — 0.80
- Graph Power — 0.79
Computed from structural-signature embeddings · 2026-10-08