Trace and local volume change Created 2026-10-14 Updated 2026-10-20
For any fixed real -by- matrix , as the real parameter tends to zero,
The big O notation means the remainder is bounded in absolute value by for some constant near zero. The determinant stays positive for sufficiently small , so its oriented volume factor is also its ordinary volume factor. Thus the trace is the first-order fractional volume change, measured from the identity matrix.
A proof follows directly from multilinearity of the determinant. Each column of is . The constant term chooses all the standard basis columns and equals one. A term containing exactly one factor replaces just column by ; expanding that vector in the standard basis leaves only its component, , because every other choice repeats a column. Summing gives . Terms with two or more replacements have order or higher. This also proves the derivative of the determinant at the identity matrix.
This coefficient has a direct visual meaning. At the point , the linear map has component along direction . In a small step , these are the first-order changes of the square's or box's sides in their own directions. Summing them gives its fractional area or volume change; the off-diagonal entries initially change shape rather than contributing to this first-order total.
For the constant-coefficient linear system of differential equations , the exact linear flow is , using the matrix exponential. Its volume factor is
This is the matrix exponential determinant identity. One can derive it without diagonalization: set , whose matrix inverse is , and use in Jacobi's formula to get . The scalar differential equation with initial value one gives the displayed exponential function. Thus positive trace causes volume expansion, negative trace causes contraction, and zero trace preserves volume, even when individual directions behave differently.
In two dimensions, compare four generators at :
The first and third diagonal matrices expand or contract both directions. The second expands horizontally and contracts vertically by reciprocal factors, making a rectangle of unchanged area. For the last nilpotent matrix, , so its matrix exponential is exactly ; it sends to , a shear mapping. Zero trace allows substantial shape change and does not mean the flow map is stationary.
Figure 1.
Four examples of a linear system of differential equations: trace controls area change while a saddle and a shear mapping preserve area
. Details.
The first-order statement and the exact flow map should be distinguished. For , , whereas for all real . Zero trace eliminates the first-order change in the approximate step, and it preserves volume exactly in the matrix exponential flow. The examples use an abstract parameter and coordinate vectors; no physical units are assumed.
Triangularity of an inverse matrix Created 2026-10-31 Updated 2026-11-05
An invertible lower triangular matrix has a lower triangular matrix inverse. To see this, solve for column of the inverse. The nonzero diagonal and forward substitution force for . Applying the result to the transpose proves the upper-triangular case as well. Invertibility is essential; a singular triangular matrix has no ordinary matrix inverse.
Given Pivoted LU decomposition , the linear system is equivalent to . Introduce the intermediate vector : first use forward substitution to solve , then back substitution to solve . With unpivoted LU decomposition, is the identity matrix, so the first right-hand side is simply .[ref]
For the factors in Computed three-by-three LU decomposition and , forward substitution gives
Then back substitution gives
Direct matrix multiplication verifies . More generally, the back substitution recurrence is
The direction matters: the last row contains only , whereas the first row generally involves every unknown. No dense matrix inverse is needed for either triangular solve.
Differentiating and multiplying by the inverse gives the displayed variation. It is the matrix inverse derivative, expressed with the covariant and contravariant components of a metric tensor.