Skip to content

Hyper-Wiener Index

Compress a connected molecular graph's shortest-path distance distribution into a scalar by summing both distance and squared distance over unordered vertex pairs.

Version
v3 · 2026-09-19 · History
Domain-specific #
2029
Domain group
Natural Sciences
Origin domain
Chemistry & Materials Science
Subdomain
Chemical Graph Theory → Chemistry & Materials Science
Aliases
Hyper-Wiener number

Core Idea

The Hyper-Wiener Index is a distance-based graph invariant developed in chemical graph theory. For a finite connected simple graph G=(V,E), let d_G(u,v) be the shortest-path distance between vertices u and v. Its standard connected-graph definition is.

WW(G) = (1/2) sum_{{u,v} subset V, u!=v} [d_G(u,v) + d_G(u,v)^2],

where the summation is over unordered pairs of distinct vertices. This convention is load-bearing. The equivalent ordered-pair formula is.

WW(G) = (1/4) sum_{u in V} sum_{v in V} [d_G(u,v) + d_G(u,v)^2].

How would you explain it like I'm…

Far Dots Count Extra

Draw a molecule as dots joined by lines. For every pair of dots, count how many hops it takes to get from one to the other along the lines. Far-apart pairs count extra: 1 hop scores 1, 2 hops score 1+2=3, and 3 hops score 1+2+3=6. Add up all the scores, and you get one number that tells you how spread out the shape is.

Molecule Spread-Out Score

Chemists sometimes draw a molecule as a graph: dots for atoms and lines for bonds. The Hyper-Wiener Index is a single number made from that picture. For every pair of dots, you find the shortest number of steps between them. A pair k steps apart adds 1+2+...+k, so far pairs add a lot more than close ones. A long, stretched-out molecule gets a bigger number than a compact one, but two different molecules can still end up with the same number.

Squared-Distance Graph Index

The Hyper-Wiener Index is a number computed from a molecule's graph, where atoms are vertices and bonds are edges (usually ignoring hydrogens). For each unordered pair of vertices you take their shortest-path distance d and add (d + d²)/2, then sum over all pairs. Its simpler cousin, the Wiener index, just adds up d; the squared term in the Hyper-Wiener Index makes distant pairs count much more, so a pair at distance k contributes k(k+1)/2. The result measures how compact or extended the molecular graph is. It stays the same if you relabel the graph, but different graphs can share the same value, and it says nothing directly about 3D shape or bond energies.

 

The Hyper-Wiener Index is a distance-based graph invariant from chemical graph theory. For a finite connected simple graph G, WW(G) = ½ Σ over unordered pairs {u,v} of [d(u,v) + d(u,v)²], with d the shortest-path distance. Equivalently, over ordered pairs it is ¼ ΣΣ [d + d²]; using ½ with an ordered double sum would double the value, and zero diagonal terms do not fix that. Randić introduced it for molecular trees as a generalization of the Wiener index, and Klein, Lukovits and Gutman extended the definition to cyclic structures. A pair at distance k contributes k(k+1)/2 versus k for the Wiener index, weighting long distances disproportionately. It is isomorphism-invariant because isomorphisms preserve distances, but it is not a complete invariant. Applied to hydrogen-suppressed molecular graphs, it summarizes topological compactness, not conformation, energetics or electronic structure.

Scope of Application

The Hyper-Wiener Index recurs primarily in mathematical chemistry and the graph theory that supports it.

  • Molecular descriptors. A molecular structure is converted to a graph and WW becomes one candidate feature in quantitative structure–property or structure–activity analysis. The model must specify atom and bond treatment; the scalar alone does not carry those choices.
  • Chemical graph families. Closed forms, recurrences, and extremal bounds compare benzenoid, phenylene, polyazulene, tree, and unicyclic families. These studies expose how branching, cycles, and long paths affect the descriptor.
  • Algorithm design. All-pairs shortest paths give a direct computation.

Clarity

Three equivalent representations prevent convention drift. Let N_k be the number of unordered vertex pairs at distance k. Then

W(G) = sum_{k>=1} N_k k

and

WW(G) = (1/2) sum_{k>=1} N_k(k+k^2).

The same quantity may be written (W(G) + sum_{{u,v}} d_G(u,v)^2)/2. Because the pair set is unordered, no further division by two is needed.

Manages Complexity

The index reduces an O(|V|^2) family of pair relations to one scalar. That compression supports sorting graph families, testing formulas, adding a fixed-size feature to a statistical model, and comparing how structural modifications alter long-range separation.

The quadratic term makes a useful diagnostic distinction. Changing a remote branch or extending a chain affects WW more strongly than a linear-only distance sum, while a local edit may have widespread effects because it changes shortest paths for many pairs.

Abstract Reasoning

Normalize the pair convention first. Decide whether an implementation enumerates unordered pairs or an ordered distance matrix. Align the leading coefficient before comparing results.

Move from graph to distance distribution. Compute d_G(u,v), collect counts N_k, and evaluate the polynomial weight k(k+1)/2. This separates shortest-path correctness from final aggregation.

Knowledge Transfer

The mathematical mechanism transfers literally to any finite connected simple graph: shortest-path distance, unordered pairs, and the same normalized polynomial aggregation require no chemical substrate. Network graphs, combinatorial constructions, and software test graphs can therefore use WW as the same graph invariant.

The interpretation does not transfer automatically. In a molecular graph, path distance counts bond steps in a declared structural representation; in a communication graph it counts hops; in a social graph it may encode an uncertain relationship path.

Relationships to Other Abstractions

Local relationship map for Hyper-Wiener IndexParents appear above the current abstraction, mutual partners to the right, and children below. Node labels state whether each abstraction is prime or domain-specific; colors identify relation types.Hyper-Wiener IndexDOMAINPrime abstraction: Metric — presupposesMetricPRIMEPrime abstraction: Aggregation — is a kind ofAggregationPRIME

Current abstraction Hyper-Wiener Index Domain-specific

Parents (2) — outgoing typed relations

  • Hyper-Wiener Index is a kind of Aggregation Prime

    Hyper-Wiener Index instantiates prime:aggregation by strict subsumption.

  • Hyper-Wiener Index presupposes Metric Prime

    Hyper-Wiener Index instantiates prime:aggregation by strict subsumption.

Hierarchy paths (2) — routes to 1 parentless root

Neighborhood in Abstraction Space

Hyper-Wiener Index sits in a sparse region of the domain-specific corpus (72nd percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Unclustered & Miscellaneous (2769 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-25