RHODES UNIVERSITY DEPARTMENT OF MATHEMATICS (Pure & Applied) EXAMINATION : JUNE 2009 APPLIED MATHEMATICS III Examiners : Dr C.C. Remsing Dr F.A.M. Frescura AVAILABLE MARKS : 110 FULL MARKS : 100 DURATION : 3 HOURS AM3.2 (LINEAR CONTROL) NB : All questions may be attempted. All steps must be clearly motivated. Marks will not be awarded if this is not done. Question 1. [18 marks] Let A ∈ Rn×n . (a) Define the transpose matrix A⊤ and then show that the linear map τ : Rn×n → Rn×n , A 7→ A⊤ satisfies the condition τ (AB) = τ (B) τ (A). Hence deduce that if the matrix A is invertible, then so is its transpose A⊤ , and −1 ⊤ A⊤ = A−1 . (b) Let φ denote the quadratic form x ∈ Rn 7→ x⊤ Ax. Explain what is meant by saying that φ is positive definite, and then give two sets of necessary and sufficient conditions for positive definiteness. For what values of α is φ(x) = x21 + 2x22 + 3x23 + 2αx1 x3 positive definite ? (c) Show that A can be written as A1 + A2 , where A1 is symmetric and A2 is skew-symmetric. Hence deduce that φ(x) = x⊤ A1 x for all x ∈ Rn . [5,7,6] Page 1 of 4 Appl. Maths III Linear Control June 2009 Question 2. [22 marks] (a) Given A ∈ Rm×m , B ∈ Rm×ℓ and x0 ∈ Rm , prove that the initialized linear control system ẋ = Ax + Bu x(0) = x0 has an unique solution curve x(·) : R → Rm . (b) Consider the linear control system ẋ1 = ω x2 ẋ2 = −ω x1 + u, ω ∈ R. Find the state transition matrix Φ(·, 0). (c) Determine the solution curve x(·) starting from the origin when u(·) = 1. [8,14] Question 3. [22 marks] Consider a linear control system (with outputs) Σ given by ẋ = Ax + Bu and y = Cx. (a) Define the terms complete controllability and complete observability (of Σ). (b) State and prove the Kalman Controllability Criterion. (c) Consider the control system (with outputs) 1 2 ẋ = x + b u and y = c x. 0 3 Find b ∈ R2×1 and c ∈ R1×2 such that the system is i. not completely controllable; ii. completely observable. When c = 1 1 , determine x(0) if y(t) = et − 2e3t . [4,10,8] Page 2 of 4 Appl. Maths III Linear Control June 2009 Question 4. [16 marks] (a) Consider a single-input linear control system Σ described by ẋ = Ax + bu A ∈ Rm×m , b ∈ Rm×1 . Given an arbitrary set Λ = {θ1 , . . . , θm } of complex numbers (appearing in conjugate pairs), prove that if Σ is completely controllable, then there exists a feedback matrix K ∈ R1×m such that the eigenvalues of A + bK are the set Λ. (b) For the control system ẋ1 = x2 ẋ2 = x3 + u ẋ3 = x1 − x2 − 2x3 find a linear feedback control which makes all the eigenvalues (of the system) equal to −1. [10,6] Question 5. [20 marks] Consider a linear dynamical system ẋ = Ax, x ∈ Rm . (a) Give necessary and sufficient conditions for i. asymptotic stability ii. neutral stability iii. instability of the system (at the origin). (b) Investigate the stability of the dynamical system ẋ1 = x2 ẋ2 = −bx1 − ax2 , a, b > 0. (c) Define the term Lyapunov function and then state the First Lyapunov Theorem. (d) State and prove the Lyapunov Linearization Theorem. [3,7,3,7] Page 3 of 4 Appl. Maths III Linear Control June 2009 Question 6. [12 marks] Consider the following optimal control problem : Minimize the cost functional J = Z T 0 u2 (t) + x2 (t) dt subject to ẋ = −x + u, x ∈ R, u ∈ R x(0) = 1 x(T ) = 2. The final time T > 0 is fixed. Find the optimal control u∗ (·) : [0, T ] → R as well as (an explicit expression for) the optimal trajectory x∗ (·) : [0, T ] → R. [12] END OF THE EXAMINATION PAPER Page 4 of 4