510 CHAPTER 27. DETERMINANTS

Definition 27.2.6 Let A be n×n. Then trace(A)≡ ∑ni=1 Aii.

Proposition 27.2.7 Let A be m×n and let B be n×m. Then trace(AB) = trace(BA) . Alsofor square matrices A,B, if A = S−1BS, then trace(A) = trace(B). Also det(A) = det(B).

Proof: trace(AB) ≡ ∑i ∑ j Ai jB ji = ∑ j ∑i B jiAi j ≡ trace(BA) . Now let A,B be as de-scribed. Then

trace(A) = traceS−1BS = trace((BS)S−1)

= trace(B(SS−1))= trace(B)

As to the claim about the determinant, it follows from the properties of the determinant that

det(A) = det(S−1BS

)= det

(S−1)det(B)det(S)

= det(B)det(S−1S

)= det(B) ■

These two, the trace and the determinant are two of the so called principal invariants of a3×3 matrix. The reason these are called invariants is that they are the same for A and B ifthese two are related as described in the above proposition. In this case, the other principalinvariant is

12(trace(A))2− 1

2trace

(A2)

It turns out these are related to the coefficients of the characteristic polynomial defined as

det(A−λ I)

and discussed below.To see this last is also an invariant, the above proposition implies

12(trace(A))2− 1

2trace

(A2) =

12(trace

(S−1BS

))2− 12

trace((

S−1BS)2)

=12(trace(B))2− 1

2trace

((S−1BS

)(S−1BS

))=

12(trace(B))2− 1

2trace

(S−1B2S

)=

12(trace(B))2− 1

2trace

(B2)

The physical reason these are important is that their invariance implies they do not changewhen one uses a different coordinate system to describe points. That which is physi-cally meaningful cannot depend on coordinate system because such coordinate systemsare purely artificial constructions used to identify points. Therefore, the principal invari-ants are good for formulating physical laws. This is as far as we go here. To see much moreon these ideas, you should take a course on continuum mechanics. However, the trace anddeterminant also have a very interesting relation to eigenvalues.

Theorem 27.2.8 The trace of a matrix is the sum of its eigenvalues listed according tomultiplicity as a root of the characteristic polynomial. Also, the determinant of the matrixequals the product of its eigenvalues.

510 CHAPTER 27. DETERMINANTSDefinition 27.2.6 Let A be nxn. Then trace (A) = V7", Ai.Proposition 27.2.7 Let A be m xn and let B be n x m. Then trace (AB) = trace (BA) . Alsofor square matrices A,B, if A= S~'BS, then trace (A) = trace (B). Also det (A) = det (B).Proof: trace (AB) =); AijBji = Lj Li ByiAij = trace (BA). Now let A,B be as de-scribed. Thentrace(A) = traceS~'BS = trace ((BS) s')= trace(B (ss~')) = trace (B)As to the claim about the determinant, it follows from the properties of the determinant thatdet (A) det (S~'BS) = det (S~') det (B) det (S)det (B) det (S~'S) = det (B) IThese two, the trace and the determinant are two of the so called principal invariants of a3 x 3 matrix. The reason these are called invariants is that they are the same for A and B ifthese two are related as described in the above proposition. In this case, the other principalinvariant is 1 15 (trace (A))* — 5 trace (A*)It turns out these are related to the coefficients of the characteristic polynomial defined asdet (A — AJ)and discussed below.To see this last is also an invariant, the above proposition implies(trace (A))* — 5 trace (A?) = ; (trace (s-'Bs))” — 5 trace ((s'Bs)”)= 5 (trace (B))? — 5 trace ((S~'BS) (S~'BS))= 5 (trace (B))? — 5 trace (S~'B’s)= 5 (trace (B))? — 5 trace (B’)The physical reason these are important is that their invariance implies they do not changewhen one uses a different coordinate system to describe points. That which is physi-cally meaningful cannot depend on coordinate system because such coordinate systemsare purely artificial constructions used to identify points. Therefore, the principal invari-ants are good for formulating physical laws. This is as far as we go here. To see much moreon these ideas, you should take a course on continuum mechanics. However, the trace anddeterminant also have a very interesting relation to eigenvalues.Theorem 27.2.8 The trace of a matrix is the sum of its eigenvalues listed according tomultiplicity as a root of the characteristic polynomial. Also, the determinant of the matrixequals the product of its eigenvalues.