6 Exponential of a matrix Definition 6.1. Given a square matrix A, we define its exponential eA as the series ∞ X Ak A2 A3 e =I +A+ + + ... = . 2! 3! k! k=0 A It is easy to check that this series converges for any square matrix A whatsoever. Theorem 6.2 (Constant matrices). If A is a constant square matrix, then Φ(t) = eAt is a fundamental matrix for the system y 0 (t) = Ay. Moreover, the initial value problem y 0 (t) = Ay, y(0) = y0 has a unique solution which is defined for all times, namely y(t) = eAt y0 . Example 6.3 (Computation of eAt ). When A is a diagonal matrix, we have λ1 eλ1 t At .. .. A= =⇒ e = . . . λn t λn e When A is a 2 × 2 Jordan block, we have ¸ · λ 1 =⇒ A= λ When A is a 3 × 3 Jordan block, we have λ 1 A = λ 1 =⇒ λ · At e ¸ eλt teλt = . eλt eAt eλt teλt = eλt eλt teλt . eλt t2 2! Using these facts, one can compute the exponential of any square matrix A. Namely, one may determine the Jordan form P −1 AP and then use the formula eAt = P · etP −1 AP · P −1 to relate the exponential of the given matrix A to that of its Jordan form. Lemma 6.4 (Product rule). If A, B are square matrices, then (AB)0 = A0 B + AB 0 . Lemma 6.5. If A, B are square matrices that commute, then eA+B = eA eB . (6.1)