Keyboard shortcuts

Press ← or → to navigate between chapters

Press ? to show this help

Press Esc to hide this help

Rectangular weak two-metric dimension below the threshold

Abstract

For 4 <= n < m < 2n - 2, the weak two-metric dimension of the rectangular Hamming graph is min(ceil(2(n+m)/3), 2n - 2).

The graph is the Cartesian product of complete graphs on n and m vertices, with Hamming distance on its row and column coordinates. A landmark set is weak two-resolving when the sum of absolute distance differences is at least two for every distinct vertex pair. The rectangular range below m = 2n - 2 is identified in Aryan Kumar, arXiv:2609.32161v1, Section 7. Throughout the displayed statements, natDiv(a,b) denotes natural number division; natDiv(2(n+m)+2,3) equals ceil(2(n+m)/3).

Definition 1.1 (Generic left degree).

Lean statement: D5/S3/Combinatorics/Graph/HammingWeakTwoMetricDimension.leftCount

Formalization. D5/S3/Combinatorics/Graph/HammingWeakTwoMetricDimension.leftCount (✓ std3).

Source. Repository-derived.

Commentary.

For arbitrary coordinate types with decidable equality, leftCount S a counts the edges of a finite incidence set S whose first coordinate is a.

Definition 1.2 (Generic right degree).

Lean statement: D5/S3/Combinatorics/Graph/HammingWeakTwoMetricDimension.rightCount

Formalization. D5/S3/Combinatorics/Graph/HammingWeakTwoMetricDimension.rightCount (✓ std3).

Source. Repository-derived.

Commentary.

For arbitrary coordinate types with decidable equality, rightCount S b counts the edges of S whose second coordinate is b.

Definition 1.3 (Disjoint incidence union).

Lean statement: D5/S3/Combinatorics/Graph/HammingWeakTwoMetricDimension.join

Formalization. D5/S3/Combinatorics/Graph/HammingWeakTwoMetricDimension.join (✓ std3).

Source. Repository-derived.

Commentary.

For two finite incidence sets with separate coordinate types, join embeds both parts into disjoint sums and takes the union of their diagonal edge images.

Theorem 1.4 (Disjoint edge images).

Lean statement: D5/S3/Combinatorics/Graph/HammingWeakTwoMetricDimension.join_disjoint

Proof. Machine-checked in Lean as D5/S3/Combinatorics/Graph/HammingWeakTwoMetricDimension.join_disjoint (✓ std3). ∎

Source. Repository-derived.

Commentary.

For arbitrary finite incidence sets S,T, the left and right diagonal images used in join are disjoint because their first coordinates have different sum tags.

Theorem 1.5 (Cardinality of incidence union).

Lean statement: D5/S3/Combinatorics/Graph/HammingWeakTwoMetricDimension.join_card

Proof. Machine-checked in Lean as D5/S3/Combinatorics/Graph/HammingWeakTwoMetricDimension.join_card (✓ std3). ∎

Source. Repository-derived.

Commentary.

For arbitrary finite incidence sets S,T, the cardinality of join S T is the sum of their cardinalities.

Theorem 1.6 (Left degree in the first block).

Lean statement: D5/S3/Combinatorics/Graph/HammingWeakTwoMetricDimension.join_left_inl

Proof. Machine-checked in Lean as D5/S3/Combinatorics/Graph/HammingWeakTwoMetricDimension.join_left_inl (✓ std3). ∎

Source. Repository-derived.

Commentary.

At a left coordinate tagged inl a, the leftCount of join S T is exactly leftCount S a.

Theorem 1.7 (Left degree in the second block).

Lean statement: D5/S3/Combinatorics/Graph/HammingWeakTwoMetricDimension.join_left_inr

Proof. Machine-checked in Lean as D5/S3/Combinatorics/Graph/HammingWeakTwoMetricDimension.join_left_inr (✓ std3). ∎

Source. Repository-derived.

Commentary.

At a left coordinate tagged inr a, the leftCount of join S T is exactly leftCount T a.

Theorem 1.8 (Right degree in the first block).

Lean statement: D5/S3/Combinatorics/Graph/HammingWeakTwoMetricDimension.join_right_inl

Proof. Machine-checked in Lean as D5/S3/Combinatorics/Graph/HammingWeakTwoMetricDimension.join_right_inl (✓ std3). ∎

Source. Repository-derived.

Commentary.

At a right coordinate tagged inl b, the rightCount of join S T is exactly rightCount S b.

Theorem 1.9 (Right degree in the second block).

Lean statement: D5/S3/Combinatorics/Graph/HammingWeakTwoMetricDimension.join_right_inr

Proof. Machine-checked in Lean as D5/S3/Combinatorics/Graph/HammingWeakTwoMetricDimension.join_right_inr (✓ std3). ∎

Source. Repository-derived.

Commentary.

At a right coordinate tagged inr b, the rightCount of join S T is exactly rightCount T b.

Theorem 1.10 (Left degree under relabeling).

Lean statement: D5/S3/Combinatorics/Graph/HammingWeakTwoMetricDimension.leftCount_map

Proof. Machine-checked in Lean as D5/S3/Combinatorics/Graph/HammingWeakTwoMetricDimension.leftCount_map (✓ std3). ∎

Source. Repository-derived.

Commentary.

For equivalences f,g of the coordinate types, mapping a finite incidence set through their product preserves its left degree at f a.

Theorem 1.11 (Right degree under relabeling).

Lean statement: D5/S3/Combinatorics/Graph/HammingWeakTwoMetricDimension.rightCount_map

Proof. Machine-checked in Lean as D5/S3/Combinatorics/Graph/HammingWeakTwoMetricDimension.rightCount_map (✓ std3). ∎

Source. Repository-derived.

Commentary.

For equivalences f,g of the coordinate types, mapping a finite incidence set through their product preserves its right degree at g b.

Theorem 1.12 (The two-thirds upper bound).

Proof. Machine-checked in Lean as D5/S3/Combinatorics/Graph/HammingWeakTwoMetricDimension.ceiling_upper (✓ std3). ∎

Source. Repository-derived.

Commentary.

For every 4 <= n < m < 2n - 2, there is a weak two-resolving landmark set of cardinality at most ceil(2(n+m)/3). View landmarks as edges between row and column vertices. Remove one row and two columns when rows are no more numerous than columns, or two rows and one column otherwise, and place a two-edge star on the removed vertices. These reductions preserve positivity and the inequalities that each part has at most twice the size of the other. At a total of three to five vertices, explicit stars or paths complete the construction. Every vertex has positive degree and each edge has endpoint degree sum at least three, so the resulting set is weak two-resolving. Each three-vertex extension adds two edges.

Theorem 1.13 (The empty-row upper bound).

Proof. Machine-checked in Lean as D5/S3/Combinatorics/Graph/HammingWeakTwoMetricDimension.empty_row_upper (✓ std3). ∎

Source. Repository-derived.

Commentary.

For every 4 <= n < m < 2n - 2, there is a weak two-resolving landmark set of cardinality at most 2n - 2. Leave row zero empty. In each row i from one to n - 1, place landmarks in columns 2(i-1) mod m and (2(i-1)+1) mod m. These columns are distinct, and the consecutive residues cover all columns because m <= 2(n-1). Every occupied row has degree two, every column has positive degree, and any disjoint landmarks have total endpoint degree at least six. The pair degree conditions therefore give weak two-resolution.

Definition 1.14 (The exact rectangular formula).

Formalization. D5/S3/Combinatorics/Graph/HammingWeakTwoMetricDimension.claim (✓ std3).

Source. Repository-derived.

Commentary.

The formula asserts, for all natural n and m satisfying 4 <= n < m < 2n - 2, equality of the weak two-metric dimension with the smaller of the two-thirds ceiling and 2n - 2.

Theorem 1.15 (Exact weak two-metric dimension).

Proof. Machine-checked in Lean as D5/S3/Combinatorics/Graph/HammingWeakTwoMetricDimension.result (✓ std3). ∎

Resolves. Problems/kumar-2026-rectangular-weak-two-metric-dimension (proved) by D5/S3/Combinatorics/Graph/HammingWeakTwoMetricDimension.result.

Source. Repository-derived.

Commentary.

The formula holds throughout the stated rectangular range. The degree constraints give the minimum of the two quantities as a lower bound for every resolving set. The two constructions give each quantity as an upper bound, and hence give their minimum as the upper bound as well.

References