Method of undetermined coefficients: A technique to find a particular solution to non-homogeneous linear differential equations when the forcing function is a polynomial, exponential, sine, cosine, or a combination thereof.
- Solve the homogeneous equation to find the complementary solution \(y_c\)
- Guess the form of the particular solution \(y_p\) based on the forcing function
- Adjust the guess if any terms in \(y_p\) overlap with \(y_c\)
- Substitute \(y_p\) into the original equation to determine coefficients
- Form the general solution: \(y = y_c + y_p\)
For \(y'' - 3y' + 2y = 0\), the characteristic equation is:
\(r^2 - 3r + 2 = 0\)
Factoring: \((r-1)(r-2) = 0\)
Therefore: \(r_1 = 1\), \(r_2 = 2\)
Complementary solution: \(y_c = c_1e^x + c_2e^{2x}\)
The forcing function is \(x^2 + 1\), which is a polynomial of degree 2.
For a polynomial of degree \(n\), we guess a polynomial of the same degree.
Since neither \(e^x\) nor \(e^{2x}\) (from \(y_c\)) appears in the polynomial forcing function, we don't need to adjust our guess.
Guess: \(y_p = Ax^2 + Bx + C\)
\(y_p = Ax^2 + Bx + C\)
\(y_p' = 2Ax + B\)
\(y_p'' = 2A\)
Substituting into \(y'' - 3y' + 2y = x^2 + 1\):
\(2A - 3(2Ax + B) + 2(Ax^2 + Bx + C) = x^2 + 1\)
\(2A - 6Ax - 3B + 2Ax^2 + 2Bx + 2C = x^2 + 1\)
\(2Ax^2 + (-6A + 2B)x + (2A - 3B + 2C) = x^2 + 1\)
Comparing coefficients on both sides:
For \(x^2\): \(2A = 1 \Rightarrow A = \frac{1}{2}\)
For \(x^1\): \(-6A + 2B = 0 \Rightarrow -6 \cdot \frac{1}{2} + 2B = 0 \Rightarrow -3 + 2B = 0 \Rightarrow B = \frac{3}{2}\)
For \(x^0\): \(2A - 3B + 2C = 1 \Rightarrow 2 \cdot \frac{1}{2} - 3 \cdot \frac{3}{2} + 2C = 1 \Rightarrow 1 - \frac{9}{2} + 2C = 1 \Rightarrow -\frac{7}{2} + 2C = 1 \Rightarrow 2C = \frac{9}{2} \Rightarrow C = \frac{9}{4}\)
\(y_p = \frac{1}{2}x^2 + \frac{3}{2}x + \frac{9}{4}\)
\(y = y_c + y_p = c_1e^x + c_2e^{2x} + \frac{1}{2}x^2 + \frac{3}{2}x + \frac{9}{4}\)
The general solution is \(y = c_1e^x + c_2e^{2x} + \frac{1}{2}x^2 + \frac{3}{2}x + \frac{9}{4}\) where \(c_1\) and \(c_2\) are arbitrary constants
• Polynomial guess: For polynomial forcing of degree n, guess polynomial of degree n
• Coefficient matching: Equate coefficients of like terms on both sides
• Superposition: General solution is sum of homogeneous and particular solutions
Resonance: Occurs when the forcing function has the same form as one of the terms in the complementary solution, requiring modification of the particular solution guess.
For \(y'' - 4y' + 4y = 0\), the characteristic equation is:
\(r^2 - 4r + 4 = 0\)
Factoring: \((r-2)^2 = 0\)
Therefore: \(r = 2\) (repeated root)
Complementary solution: \(y_c = (c_1 + c_2x)e^{2x}\)
The forcing function is \(e^{2x}\), and \(e^{2x}\) appears in the complementary solution.
This is a resonance case because the forcing function matches a term in the homogeneous solution.
When resonance occurs, we multiply our initial guess by \(x\) raised to the power of the multiplicity of the root.
Since \(e^{2x}\) appears with multiplicity 1 in \(y_c\), we multiply by \(x^1\).
But wait: the homogeneous solution contains both \(e^{2x}\) and \(xe^{2x}\), so we need to multiply by \(x^2\).
Since \(e^{2x}\) and \(xe^{2x}\) are both in \(y_c\), we multiply our guess by \(x^2\).
Initial guess would be \(Ae^{2x}\), but since \(e^{2x}\) appears in \(y_c\) with multiplicity 2, we use:
Guess: \(y_p = Ax^2e^{2x}\)
\(y_p = Ax^2e^{2x}\)
Using the product rule: \(y_p' = 2Axe^{2x} + 2Ax^2e^{2x} = 2Axe^{2x}(1 + x)\)
For the second derivative: \(y_p'' = 2Ae^{2x}(1 + x) + 2Axe^{2x} \cdot 2(1 + x) + 2Axe^{2x} \cdot 1\)
\(y_p'' = 2Ae^{2x}(1 + x) + 4Axe^{2x}(1 + x) + 2Axe^{2x}\)
\(y_p'' = 2Ae^{2x}[(1 + x) + 2x(1 + x) + x] = 2Ae^{2x}[1 + x + 2x + 2x^2 + x] = 2Ae^{2x}[1 + 4x + 2x^2]\)
\(y_p'' = 2Ae^{2x}(1 + 4x + 2x^2)\)
Substituting into \(y'' - 4y' + 4y = e^{2x}\):
\(2Ae^{2x}(1 + 4x + 2x^2) - 4 \cdot 2Axe^{2x}(1 + x) + 4Ax^2e^{2x} = e^{2x}\)
\(2Ae^{2x}(1 + 4x + 2x^2) - 8Axe^{2x}(1 + x) + 4Ax^2e^{2x} = e^{2x}\)
\(2Ae^{2x}(1 + 4x + 2x^2) - 8Axe^{2x} - 8Ax^2e^{2x} + 4Ax^2e^{2x} = e^{2x}\)
\(2Ae^{2x}(1 + 4x + 2x^2 - 4x - 4x^2 + 2x^2) = e^{2x}\)
\(2Ae^{2x}(1 + 0x + 0x^2) = e^{2x}\)
\(2Ae^{2x} = e^{2x}\)
Comparing coefficients: \(2A = 1\)
Therefore: \(A = \frac{1}{2}\)
\(y_p = \frac{1}{2}x^2e^{2x}\)
\(y = y_c + y_p = (c_1 + c_2x)e^{2x} + \frac{1}{2}x^2e^{2x}\)
The general solution is \(y = (c_1 + c_2x)e^{2x} + \frac{1}{2}x^2e^{2x}\) where \(c_1\) and \(c_2\) are arbitrary constants
• Resonance rule: When forcing function matches homogeneous solution, multiply by \(x^k\) where k is multiplicity
• Product rule: \((uv)' = u'v + uv'\) for differentiation of \(x^ne^{rx}\)
• Modification: Adjust guess when terms overlap with complementary solution
Trigonometric forcing: When the forcing function is a sine or cosine function, the particular solution typically has the same form but with undetermined coefficients.
For \(y'' + 4y = 0\), the characteristic equation is:
\(r^2 + 4 = 0\)
Therefore: \(r = \pm 2i\)
Complementary solution: \(y_c = c_1\cos(2x) + c_2\sin(2x)\)
The forcing function is \(\sin(2x)\), and both \(\cos(2x)\) and \(\sin(2x)\) appear in the complementary solution.
This is a resonance case because the forcing function matches terms in the homogeneous solution.
When resonance occurs with trigonometric functions, we multiply our initial guess by \(x\).
For a forcing function of the form \(\sin(2x)\) or \(\cos(2x)\), we normally guess:
\(A\cos(2x) + B\sin(2x)\)
But since both \(\cos(2x)\) and \(\sin(2x)\) appear in \(y_c\), we multiply by \(x\):
Guess: \(y_p = Ax\cos(2x) + Bx\sin(2x)\)
\(y_p = Ax\cos(2x) + Bx\sin(2x)\)
Using the product rule: \(y_p' = A\cos(2x) - 2Ax\sin(2x) + B\sin(2x) + 2Bx\cos(2x)\)
\(y_p' = (A + 2Bx)\cos(2x) + (B - 2Ax)\sin(2x)\)
For the second derivative: \(y_p'' = 2B\cos(2x) - 2(A + 2Bx)\sin(2x) - 2A\sin(2x) + 2(B - 2Ax)\cos(2x)\)
\(y_p'' = [2B - 4Ax]\cos(2x) + [-2A - 4Bx - 2A]\sin(2x)\)
\(y_p'' = (2B - 4Ax)\cos(2x) + (-4A - 4Bx)\sin(2x)\)
Substituting into \(y'' + 4y = \sin(2x)\):
\((2B - 4Ax)\cos(2x) + (-4A - 4Bx)\sin(2x) + 4(Ax\cos(2x) + Bx\sin(2x)) = \sin(2x)\)
\((2B - 4Ax)\cos(2x) + (-4A - 4Bx)\sin(2x) + 4Ax\cos(2x) + 4Bx\sin(2x) = \sin(2x)\)
\(2B\cos(2x) - 4Ax\cos(2x) - 4A\sin(2x) - 4Bx\sin(2x) + 4Ax\cos(2x) + 4Bx\sin(2x) = \sin(2x)\)
\(2B\cos(2x) - 4A\sin(2x) = \sin(2x)\)
Comparing coefficients on both sides:
For \(\cos(2x)\): \(2B = 0 \Rightarrow B = 0\)
For \(\sin(2x)\): \(-4A = 1 \Rightarrow A = -\frac{1}{4}\)
\(y_p = -\frac{1}{4}x\cos(2x) + 0 \cdot x\sin(2x) = -\frac{1}{4}x\cos(2x)\)
\(y = y_c + y_p = c_1\cos(2x) + c_2\sin(2x) - \frac{1}{4}x\cos(2x)\)
The general solution is \(y = c_1\cos(2x) + c_2\sin(2x) - \frac{1}{4}x\cos(2x)\) where \(c_1\) and \(c_2\) are arbitrary constants
• Trigonometric guess: For \(\sin(ax)\) or \(\cos(ax)\), guess \(A\cos(ax) + B\sin(ax)\)
• Resonance adjustment: When frequency matches natural frequency, multiply by x
• Trigonometric derivatives: \(\frac{d}{dx}[\cos(ax)] = -a\sin(ax)\), \(\frac{d}{dx}[\sin(ax)] = a\cos(ax)\)
Method of undetermined coefficients: A technique to find a particular solution to non-homogeneous linear differential equations by guessing the form of the solution and determining coefficients.
Complementary solution (\(y_c\)): The general solution to the associated homogeneous equation.
Particular solution (\(y_p\)): Any specific solution to the non-homogeneous equation.
Resonance: Occurs when the forcing function has the same form as one of the terms in the complementary solution.
- Solve homogeneous equation: Find \(y_c\) by solving \(ay'' + by' + cy = 0\)
- Identify forcing function: Analyze the form of \(g(x)\) on the right side
- Check for resonance: See if any terms in \(g(x)\) appear in \(y_c\)
- Guess particular solution: Based on the table of standard forms
- Adjust for resonance: Multiply by \(x^k\) if resonance occurs
- Substitute and solve: Plug \(y_p\) into the original equation to find coefficients
- Form general solution: \(y = y_c + y_p\)
• General form: \(y = y_c + y_p\)
• For \(g(x) = P_n(x)\) (polynomial): Guess \(y_p = Q_n(x)\) where \(Q_n\) is polynomial of same degree
• For \(g(x) = ae^{bx}\): Guess \(y_p = Ae^{bx}\) (unless \(e^{bx}\) in \(y_c\), then multiply by \(x^k\))
• For \(g(x) = a\cos(cx) + b\sin(cx)\): Guess \(y_p = A\cos(cx) + B\sin(cx)\) (adjust for resonance)
• For resonance: Multiply guess by \(x^k\) where k is the multiplicity of the root
Exponential resonance: Occurs when the exponential forcing function has the same base as a term in the complementary solution.
For \(y'' - 2y' + y = 0\), the characteristic equation is:
\(r^2 - 2r + 1 = 0\)
Factoring: \((r-1)^2 = 0\)
Therefore: \(r = 1\) (repeated root)
Complementary solution: \(y_c = (c_1 + c_2x)e^x\)
The forcing function is \(3e^x\), and \(e^x\) appears in the complementary solution.
Actually, both \(e^x\) and \(xe^x\) appear in \(y_c\) (due to the repeated root).
This is a resonance case with multiplicity 2, so we multiply our initial guess by \(x^2\).
For the forcing function \(3e^x\), we would normally guess \(Ae^x\).
But since \(e^x\) and \(xe^x\) are both in \(y_c\), we multiply by \(x^2\):
Guess: \(y_p = Ax^2e^x\)
\(y_p = Ax^2e^x\)
Using the product rule: \(y_p' = 2Axe^x + Ax^2e^x = Axe^x(2 + x)\)
For the second derivative: \(y_p'' = Ae^x(2 + x) + Ax \cdot e^x(2 + x) + Ax \cdot e^x \cdot 1\)
\(y_p'' = Ae^x(2 + x) + Ax \cdot e^x(2 + x) + Axe^x = Ae^x[(2 + x) + x(2 + x) + x]\)
\(y_p'' = Ae^x[2 + x + 2x + x^2 + x] = Ae^x[2 + 4x + x^2]\)
Substituting into \(y'' - 2y' + y = 3e^x\):
\(Ae^x(2 + 4x + x^2) - 2Axe^x(2 + x) + Ax^2e^x = 3e^x\)
\(Ae^x(2 + 4x + x^2) - 2Axe^x(2 + x) + Ax^2e^x = 3e^x\)
\(Ae^x(2 + 4x + x^2 - 4x - 2x^2 + x^2) = 3e^x\)
\(Ae^x(2 + 0x + 0x^2) = 3e^x\)
\(2Ae^x = 3e^x\)
Comparing coefficients: \(2A = 3\)
Therefore: \(A = \frac{3}{2}\)
\(y_p = \frac{3}{2}x^2e^x\)
\(y = y_c + y_p = (c_1 + c_2x)e^x + \frac{3}{2}x^2e^x\)
The general solution is \(y = (c_1 + c_2x)e^x + \frac{3}{2}x^2e^x\) where \(c_1\) and \(c_2\) are arbitrary constants
• Exponential resonance: When \(e^{rx}\) appears in \(y_c\) with multiplicity k, multiply guess by \(x^k\)
• Product rule: \((uv)' = u'v + uv'\) for differentiation of \(x^ne^{rx}\)
• Repeated root solution: For root r with multiplicity 2, solutions include \(e^{rx}\) and \(xe^{rx}\)
Mixed forcing: When the forcing function is a sum of different types of functions, we find particular solutions for each term separately and add them together.
For \(y'' + y = 0\), the characteristic equation is:
\(r^2 + 1 = 0\)
Therefore: \(r = \pm i\)
Complementary solution: \(y_c = c_1\cos(x) + c_2\sin(x)\)
The forcing function is \(2x + \cos(x)\), which is a sum of two different types of functions.
By linearity, we can find particular solutions for each term separately and add them.
For \(2x\): Find \(y_{p1}\)
For \(\cos(x)\): Find \(y_{p2}\)
Then: \(y_p = y_{p1} + y_{p2}\)
Forcing function: \(2x\) (polynomial of degree 1)
Guess: \(y_{p1} = Ax + B\)
\(y_{p1}' = A\), \(y_{p1}'' = 0\)
Substituting into \(y'' + y = 2x\): \(0 + (Ax + B) = 2x\)
\(Ax + B = 2x\)
Comparing coefficients: \(A = 2\), \(B = 0\)
Therefore: \(y_{p1} = 2x\)
Forcing function: \(\cos(x)\)
Notice that \(\cos(x)\) appears in \(y_c\), so we have resonance.
Normally we'd guess \(A\cos(x) + B\sin(x)\), but since both appear in \(y_c\), we multiply by \(x\).
Guess: \(y_{p2} = Ax\cos(x) + Bx\sin(x)\)
\(y_{p2}' = A\cos(x) - Ax\sin(x) + B\sin(x) + Bx\cos(x)\)
\(y_{p2}' = (A + Bx)\cos(x) + (B - Ax)\sin(x)\)
\(y_{p2}'' = B\cos(x) - (A + Bx)\sin(x) - A\sin(x) + (B - Ax)\cos(x)\)
\(y_{p2}'' = (2B - Ax)\cos(x) + (-2A - Bx)\sin(x)\)
Substituting into \(y'' + y = \cos(x)\):
\((2B - Ax)\cos(x) + (-2A - Bx)\sin(x) + Ax\cos(x) + Bx\sin(x) = \cos(x)\)
\(2B\cos(x) - 2A\sin(x) = \cos(x)\)
Comparing coefficients: \(2B = 1\), \(-2A = 0\)
Therefore: \(B = \frac{1}{2}\), \(A = 0\)
So: \(y_{p2} = \frac{1}{2}x\sin(x)\)
\(y_p = y_{p1} + y_{p2} = 2x + \frac{1}{2}x\sin(x)\)
\(y = y_c + y_p = c_1\cos(x) + c_2\sin(x) + 2x + \frac{1}{2}x\sin(x)\)
The general solution is \(y = c_1\cos(x) + c_2\sin(x) + 2x + \frac{1}{2}x\sin(x)\) where \(c_1\) and \(c_2\) are arbitrary constants
• Linearity principle: For \(g(x) = g_1(x) + g_2(x)\), \(y_p = y_{p1} + y_{p2}\)
• Superposition: Particular solutions for individual terms can be added
• Resonance handling: Adjust guesses when forcing functions match homogeneous solutions
Superposition principle: The general solution to a non-homogeneous linear equation is the sum of the homogeneous solution and a particular solution.
Linear operator: For \(L[y] = ay'' + by' + cy\), the equation becomes \(L[y] = g(x)\), and \(L[y_c] = 0\), \(L[y_p] = g(x)\).
Resonance phenomenon: When the forcing frequency matches the natural frequency of the system, the response grows unbounded.
- Variation of parameters: More general method that works for any forcing function
- Annihilator method: Uses differential operators to convert non-homogeneous equations to homogeneous ones
- Laplace transform: Particularly effective for equations with discontinuous forcing
- Fourier transform: Useful for periodic forcing functions
• General solution: \(y = y_c + y_p\) where \(L[y_c] = 0\) and \(L[y_p] = g(x)\)
• For resonance with exponential: If \(e^{rx}\) has multiplicity k in \(y_c\), multiply guess by \(x^k\)
• For resonance with trigonometric: If \(\cos(\omega x)\) or \(\sin(\omega x)\) in \(y_c\), multiply by x
• Linear operator: \(L[c_1y_1 + c_2y_2] = c_1L[y_1] + c_2L[y_2]\)
• Uniqueness: The general solution contains all possible solutions to the equation
\(m\frac{d^2x}{dt^2} + c\frac{dx}{dt} + kx = F_0\cos(\omega t)\) where m is mass, c is damping, k is spring constant, and F₀cos(ωt) is the forcing function.
This equation exhibits resonance when ω matches the natural frequency.
Analysis: The solution shows how the system responds to external forcing, with resonance occurring at specific frequencies.
- Below resonance: Small amplitude response
- At resonance: Maximum amplitude (potentially destructive)
- Above resonance: Decreasing amplitude response