4.6.14 · D4Ordinary Differential Equations

Exercises — Non-homogeneous — method of undetermined coefficients (annihilator method)

3,454 words16 min readBack to topic

Before we begin, one reminder of the four symbols we reuse everywhere:


Level 1 — Recognition

Goal: look at a function and instantly write the operator that kills it. No solving yet — just reading the function backwards to its characteristic root.

L1.1 State the annihilator of .

Recall Solution L1.1

An exponential is born from the single characteristic root . Here , so the root is , and the operator with that root is . Check: The constant is irrelevant — annihilators kill the shape, not the size.

L1.2 State the annihilator of .

Recall Solution L1.2

Powers all come from the root repeated times, because a root of multiplicity produces . For we have , so we need repeated times: . Check: differentiate three times:

L1.3 State the annihilator of .

Recall Solution L1.3

A pure sine or cosine of frequency comes from the pair of complex roots (see Characteristic equation and repeated roots). Building a polynomial with roots and : With that is . Check:

L1.4 State the annihilator of .

Recall Solution L1.4

The extra factor means the root must be repeated once more. A root of multiplicity gives and . For the rule is multiplicity ; here , so multiplicity : Trap avoided: alone kills but NOT — you need the square.


Level 2 — Application

Goal: run the full six-step algorithm on clean cases with NO overlap between and .

L2.1 Solve .

Recall Solution L2.1

Step 1 — homogeneous. Characteristic equation , roots . So Step 2 — annihilate. has root . It is NOT or , so no overlap. . Step 3–4 — bigger homogeneous. Roots become . The genuinely new root is , giving the trial Step 5 — determine. , . Substitute: Step 6.

L2.2 Solve .

Recall Solution L2.2

Step 1. , so . Step 2. is a degree-1 polynomial → root repeated twice → . (No overlap: is not .) Step 3–4. New roots give terms and , so Step 5. , . Then . Match: ; . Step 6.

L2.3 Solve .

Recall Solution L2.3

Step 1. . So Step 2. has roots (that is ). Annihilator . No overlap (). Step 3–4. New terms , so we MUST include both (differentiation mixes them): Step 5. , . Substitute into : Set equal to (so -coeff , -coeff ): From the first, . Sub: , . Step 6.


Level 3 — Analysis (overlap / resonance / sums)

Goal: detect when shares a root with (forcing an extra ), and handle sums by superposition.

L3.1 Solve .

Recall Solution L3.1

Step 1. , a double root . So . Step 2 — spot the overlap. has root , but is already a root of with multiplicity . Annihilator . Applied: Step 3–4. Root now has multiplicity , giving . The first two are exactly ; the genuinely new term is : (The multiplicity jumped from 2 to 3, so the new term carries — that is why we multiply by where is the old multiplicity, here .) Step 5. With : . , . Then So . Step 6. This is resonance in exponential clothing: forcing at a system's own root pumps energy in, and the response grows like .

L3.2 Solve .

Recall Solution L3.2

Step 1. , so . Step 2 — overlap. has roots , the same as ! Annihilator . Applied: Step 3–4. Roots now have multiplicity , giving . Drop the first two (); the new pair carries the : Step 5. Compute . Let . . Then (the -terms cancel — that is the point of resonance). Match to : ; . Step 6. The term is a growing oscillation — textbook resonance.

L3.3 Solve (a sum — use superposition).

Recall Solution L3.3

Step 1. , roots . So . Step 2 — split by superposition. Solve and separately, then add.

  • Piece 1: , root overlaps 's root (simple). So multiply by : . , . Then . Set . So .
  • Piece 2: (a constant → root , no overlap). Trial . Then . Step 6.

Level 4 — Synthesis

Goal: assemble everything — build annihilators for products, mix overlap with sums, and choose forms carefully.

L4.1 Give the trial form (unknown coefficients only, do NOT solve) for given that .

Recall Solution L4.1

: , double root , so . Piece : annihilator . But root already has multiplicity in , so applying it makes multiplicity , giving . Drop the first two (). New terms: . Trial: (Equivalently: base form times because of the double overlap.) Piece : roots , annihilator . No overlap with root . Trial: Combined trial: (Six unknowns — wait, five. Solving them is a separate exercise; here we only needed the correct shape.)

L4.2 Solve (overlap at the resonant frequency) and interpret.

Recall Solution L4.2

Step 1. , . Step 2 — overlap. has roots , identical to . , so , multiplicity 2. Step 3–4. New terms: . So . Step 5. With : . . Match : ; . Step 6. Interpretation. The amplitude grows without bound — pure resonance. See Resonance in forced oscillations. Physically, pushing a swing at exactly its natural frequency makes it swing ever higher.

L4.3 Solve (a repeated root at overlapping the constant forcing).

Recall Solution L4.3

Step 1. Characteristic: . Roots (double), . So Step 2 — overlap. is a constant → root . But already has multiplicity in ! Annihilator . Applied: , so root has multiplicity . Step 3–4. Terms from root : . First two are ; the new one is . Trial: Step 5. , . Then . Step 6.


Level 5 — Mastery

Goal: reason about the method itself — general statements, degenerate inputs, and why the machinery works.

L5.1 Without solving, state the order (degree of the operator) of the smallest homogeneous equation you would build for Then state how many unknown coefficients the trial has.

Recall Solution L5.1

has order (roots ).

  • : annihilator , order .
  • : annihilator , order . Combined annihilator , order . So has order . Overlap accounting for the trial:
  • Root : originally multiplicity in , forcing needs → in multiplicity . Terms ; drop (), new: → coefficients from base .
  • Root : originally multiplicity , forcing → multiplicity . Terms ; drop , new: . So : 3 unknown coefficients .

L5.2 Degenerate input. What is the annihilator of ? What does the method then produce, and is that consistent?

Recall Solution L5.2

The zero function is annihilated by the identity operator... trivially by anything, and its lowest-order annihilator is the "do nothing to reach zero" case: literally (the identity times a constant), since and no lower order exists. Applying to gives unchanged — no new roots. So has no new terms: . Consistency: the equation IS homogeneous; its full solution is just . The method correctly returns . Everything agrees. ✓

L5.3 Explain the mechanism. In one paragraph, argue why the terms coming from the roots of (inside the big equation ) can never help build .

Recall Solution L5.3

Any term generated by a root of satisfies by definition — that is what "root of " means (see Characteristic equation and repeated roots). So if we tried to use such a inside , plugging it into the original equation gives , which contributes nothing to the required right-hand side . Only the new roots that the annihilator adds produce terms with , and those are the only ones capable of matching . Hence is built purely from the "-part," and the "-part" is discarded because it merely reconstructs . This is also why overlap forces an : a shared root gains multiplicity in , and the extra multiplicity's term carries the higher power of that is genuinely new (not already in ).

L5.4 Solve completely with the initial conditions : (This is Worked Example 1 from the parent, now with an IVP — the mastery step is pinning the constants.)

Recall Solution L5.4

From the parent note, the general solution is Apply : . ; apply : . Subtract the first equation from the second: . Then . Verify: ✓; ✓.


Wrap-up

Recall One-line summary of the whole page

Read backwards to its roots → write → build keep only the roots added (multiply by for any shared root of multiplicity ) → solve coefficients in the real equation → add → then (if given) apply initial conditions to the total.