Real & Distinct Roots — Question 8

PDF ↗

Question 8

The positive response y(t)=e−t+9e−3t(t≥0)y(t)=e^{-t}+9e^{-3t}\qquad(t\ge 0) solves a homogeneous second-order equation with real distinct roots. Define its instantaneous decay rate by R(t)=−y′(t)/y(t)R(t)=-y'(t)/y(t).

Tasks

  1. Recover the monic differential equation and verify the initial value and slope.

  2. Express RR as a weighted average of the two decay rates. Find R(0)R(0), its limit, and the unique time when R=2R=2.

  3. Differentiate RR using the differential equation to prove that it is strictly decreasing. Sketch RR with its limiting level marked.

  4. A constant-rate extrapolation from the initial data predicts ŷ(t)=10e−(14/5)t\widehat y(t)=10e^{-(14/5)t}. Prove whether this estimate lies above or below the actual response for every t>0t>0, using the curvature of ln⁡y\ln y.

Original worksheet page 1: question and worked solution for 3-2-008
Show solutionHide solution

Question 8 – Solution

Strategy. A mixture of two exponential modes need not have a constant effective decay rate; use its logarithmic derivative to quantify the change.

Step 1: Recover and check the equation. The roots are −1,−3-1,-3, so the monic characteristic polynomial is r2+4r+3r^2+4r+3. Hence y″+4y′+3y=0\boxed{y''+4y'+3y=0}. Direct differentiation gives y′=−e−t−27e−3ty'=-e^{-t}-27e^{-3t}, so y(0)=10,y′(0)=−28.\boxed{y(0)=10,\qquad y'(0)=-28}. The response is positive because both mode terms are positive.

Step 2: Read the rate as a changing weighted average. Let z=9e−2t>0z=9e^{-2t}>0. Then R=e−t+27e−3te−t+9e−3t=1+3z1+z.R=\frac{e^{-t}+27e^{-3t}}{e^{-t}+9e^{-3t}} =\frac{1+3z}{1+z}. This averages 1 and 3 with positive weights 1,z1,z, so 1<R<31<R<3. In particular, R(0)=14/5,R→1,R=2⇔z=1⇔t=ln⁡3.\boxed{R(0)=14/5},\qquad \boxed{R\to 1},\qquad R=2\ \Longleftrightarrow\ z=1\ \Longleftrightarrow\ \boxed{t=\ln 3}.

Step 3: Derive the rate equation. Using y″=−4y′−3yy''=-4y'-3y and y′/y=−Ry'/y=-R, R′=−y″y+(y′/y)2=R2−4R+3=(R−1)(R−3)<0.R'=-\frac{y''}{y}+(y'/y)^2 =R^2-4R+3=(R-1)(R-3)<0. Thus RR is strictly decreasing; its long-time limit is the slower mode’s rate, rather than an unchanging average of the roots.

See the diagram in the original worksheet below.

Step 4: Compare with a constant-rate prediction. Set f=ln⁡yf=\ln y. Then f′=−Rf'=-R and f″=−R′>0f''=-R'>0. Since f′f' is strictly increasing, integration on [0,t][0,t] for t>0t>0 gives f(t)>f(0)+tf′(0)=ln⁡10−145t.f(t)>f(0)+t f'(0)=\ln 10-\frac{14}{5}t. Exponentiation preserves this inequality, so y(t)>10e−(14/5)t\boxed{y(t)>10e^{-(14/5)t}} for every t>0t>0. The extrapolation matches the initial value and slope but decays too quickly afterward. The changing mode weights, not an error in the initial data, cause the discrepancy.

Original worksheet page 2: question and worked solution for 3-2-008

Original worksheet layout. Use Enlarge or open the PDF for a closer view.