Questions: Delta-Complexes

Question 1

The torus can be represented as a Delta-complex with just 2 triangles, 3 edges, and 1 vertex (via the standard square identification). How many simplices would a simplicial complex triangulation of the torus require at minimum?

• A. 2 triangles, 3 edges, 1 vertex (the same as the Delta-complex)
• B. 7 vertices, 21 edges, 14 triangles
• C. 4 vertices, 6 edges, 4 triangles
• D. 6 vertices, 12 edges, 8 triangles

Answer: (B)

A simplicial complex requires that all vertices of a simplex be distinct and that two simplices intersect only in a common face. The minimal simplicial triangulation of the torus needs 7 vertices, 21 edges, and 14 triangles (this is a theorem — the torus cannot be simplicially triangulated with fewer than 7 vertices). The Delta-complex achieves 2 triangles by allowing vertex and edge identifications that a simplicial complex forbids, demonstrating the enormous efficiency gain.

Question 2

In a Delta-complex, a 1-simplex (edge) can have both endpoints identified to the same vertex, forming a loop.

• T. True
• F. False

Answer: True

This is one of the key flexibilities of Delta-complexes over simplicial complexes. In a simplicial complex, the two endpoints of an edge must be distinct vertices. In a Delta-complex, the attaching map for a 1-simplex maps the standard 1-simplex [v_0, v_1] into the space, and v_0 and v_1 can map to the same point. For example, the circle S^1 can be built as a Delta-complex with a single vertex and a single edge whose endpoints are both identified to that vertex.

Question 3

What is the main advantage of Delta-complexes over simplicial complexes for computing homology?

• A. Delta-complexes produce different homology groups that are more informative
• B. Delta-complexes require far fewer simplices, making the boundary matrices smaller and computation more efficient
• C. Delta-complexes work for non-compact spaces while simplicial complexes do not
• D. Delta-complexes have simpler boundary operators

Answer: (B)

The homology groups computed from a Delta-complex are isomorphic to those from any simplicial complex triangulation of the same space. The advantage is purely computational: fewer simplices mean smaller chain groups and smaller boundary matrices. The torus example is striking — 2 triangles versus 14. The boundary formula is the same alternating sum, but the attaching maps may identify faces, which must be tracked in the boundary computation.

Question 4

Explain how the boundary operator works in a Delta-complex when faces of a simplex are identified.

Model answer: The boundary operator still uses the alternating sum formula d_n(sigma) = sum(-1)^i sigma|[v_0,...,v̂_i,...,v_n], but the face maps sigma|[v_0,...,v̂_i,...,v_n] may now send different faces to the same lower-dimensional simplex. When two faces are identified, the same (n-1)-simplex appears multiple times (possibly with different signs) in the boundary, and these terms add algebraically in the chain group. For example, if a triangle has two edges identified, the boundary might have a term where that edge appears with coefficient +1 and -1, canceling to 0, or with coefficient +2.

This is the key computational subtlety of Delta-complexes. In a simplicial complex, all faces of a simplex are distinct, so the boundary is always a sum of distinct simplices. In a Delta-complex, the identifications can cause coefficients other than +1 and -1 to appear, which is reflected in the boundary matrix. Despite this added complexity, the fundamental property d ∘ d = 0 still holds, and homology is well-defined.