First Year Engineering is coming soon

We're still writing and reviewing the notes for every subject, and we'll open it when every subject is ready. Meanwhile you can practise aptitude, ask Aria about any subject, or see the branches that are ready.

Taylor's and Maclaurin's series

Taylor’s and Maclaurin’s series: definitions, theorems (with remainder), standard expansions, convergence, error bounds, and exam-style worked examples.

Drafted with Aria, reviewed by the AiCanCode.org team · Updated . How we write and check our notes. Spotted an error? Use Give Feedback at the bottom of the page.

Why it matters

Taylor’s and Maclaurin’s series let us replace complicated functions by polynomials that are easier to compute, differentiate, integrate, and analyze. Engineers use these series to linearize systems around operating points, design control laws, approximate sensor models, compute with limited precision, and estimate errors in numerical methods.

Key ideas

  • Derivative: The instantaneous rate of change of f with respect to x at a point; f′(a) is the limit of the average rate of change near a.
  • n-th derivative: The derivative of order n, written f⁽ⁿ⁾(x), obtained by differentiating n times.
  • Polynomial approximation: Replacing f(x) near a point a by a polynomial Pₙ(x) that matches the function’s value and derivatives up to order n at x = a.
  • Taylor series about a: The infinite power series that, if it converges to f(x), represents f as a sum of all orders of derivatives at a.
  • Maclaurin series: The special case of the Taylor series expanded about a = 0.
  • Remainder (error) term: The difference Rₙ(x) = f(x) − Pₙ(x) between the function and its degree-n Taylor polynomial; used to bound or estimate error of approximation.
  • Radius/interval of convergence: The set of x values around the expansion point where the Taylor/Maclaurin series converges; outside this set, the series may diverge or converge to a different function.

Exact statements commonly required in university answers:

  1. Taylor’s theorem with Lagrange’s form of the remainder
  • Statement: Let f be (n+1) times continuously differentiable on an open interval containing a and x. Then f(x) = f(a) + f′(a)(x−a) + f″(a)(x−a)²/2! + … + f⁽ⁿ⁾(a)(x−a)ⁿ/n! + Rₙ(x), where the remainder Rₙ(x) is given by Rₙ(x) = f⁽ⁿ⁺¹⁾(ξ) (x−a)ⁿ⁺¹ / (n+1)! for some ξ between a and x (by the Mean Value Theorem).
  • Assumptions: f ∈ Cⁿ⁺¹ on an open interval containing a and x (that is, f has a continuous derivative up to order n+1). The proof uses repeated application of the Mean Value Theorem.
  • Typical derivation (outline): Define ϕ(t) = f(t) − Pₙ(t) − k(t−a)ⁿ⁺¹, choose k so that ϕ(x)=0 and apply Rolle’s theorem (a corollary of the Mean Value Theorem) repeatedly to show existence of ξ with Rₙ(x)=f⁽ⁿ⁺¹⁾(ξ)(x−a)ⁿ⁺¹/(n+1)!.
  1. Taylor’s theorem with Cauchy’s form of the remainder
  • Statement: Under the same assumptions, another form is Rₙ(x) = f⁽ⁿ⁺¹⁾(ξ) (x−a)ⁿ⁺¹ / n! · 1/(x−ξ)ⁿ, for some ξ between a and x.
  • Use: Sometimes sharper bounds; Lagrange’s form is more common in first-year courses.
  1. Maclaurin series definition
  • Statement: If f has derivatives of all orders at 0 and the infinite series converges to f(x) for |x| less than some radius R > 0, then f(x) = f(0) + f′(0)x + f″(0)x²/2! + … + f⁽ⁿ⁾(0)xⁿ/n! + … .
  • This is the Taylor series with a = 0.
  1. Convergence and equality to the function
  • A Taylor series may converge for some x and diverge for others. Even when it converges, it may not equal f(x) unless the remainder Rₙ(x) → 0 as n → ∞. Sufficient condition: If there exist constants M > 0 and R > 0 such that |f⁽ⁿ⁾(a)| ≤ M·n!·(1/R)ⁿ for all n, then the Taylor series has radius of convergence at least R and Rₙ(x) → 0 for |x−a| < R. Standard tests (ratio test, root test) determine convergence for power series.
  1. Standard Maclaurin expansions (with intervals of convergence)
  • exp: eˣ = 1 + x + x²/2! + x³/3! + …, convergent for all x ∈ ℝ.
  • sine: sin x = x − x³/3! + x⁵/5! − …, convergent for all x ∈ ℝ.
  • cosine: cos x = 1 − x²/2! + x⁴/4! − …, convergent for all x ∈ ℝ.
  • geometric-like: 1/(1−x) = 1 + x + x² + x³ + … for |x| < 1.
  • logarithm: ln(1+x) = x − x²/2 + x³/3 − x⁴/4 + … for −1 < x ≤ 1, with conditional convergence at x = 1 and divergence at x = −1.
  • arctan: arctan x = x − x³/3 + x⁵/5 − … for |x| ≤ 1, with conditional convergence at x = ±1.
  1. Error bounds using Lagrange remainder
  • If |f⁽ⁿ⁺¹⁾(t)| ≤ M on the interval between a and x, then |Rₙ(x)| ≤ M |x−a|ⁿ⁺¹ / (n+1)!. This lets you choose n to meet a required accuracy.

What to draw and label (when asked for a diagram):

  • Draw x–y axes. Plot a smooth curve y = f(x) around x = a. Plot the tangent line at x = a (label slope f′(a)). Optionally plot the quadratic parabola matching up to f″(a)/2!. Mark the point (a, f(a)). Show the horizontal distance (x−a) and vertical error Rₙ(x) as the gap between f(x) and Pₙ(x) at a nearby x. Label P₁(x) for linear and P₂(x) for quadratic approximations.

Formulas

Pₙ(x) = Σ_{k=0}^n f⁽ᵏ⁾(a) (x−a)ᵏ / k!

  • Pₙ(x): Taylor polynomial of degree n (dimension depends on f); f⁽ᵏ⁾(a): k-th derivative at a (same unit as f divided by xᵏ); x, a: real variables (typically dimensionless for pure math; otherwise units of the input quantity); k!: factorial of k (dimensionless). Applies when f is differentiable up to order n at a.

f(x) = Σ_{k=0}^∞ f⁽ᵏ⁾(a) (x−a)ᵏ / k!

  • Infinite Taylor series about a. Applies where the series converges and equals f(x).

Rₙ(x) = f(x) − Pₙ(x)

  • Remainder after degree n. Same units as f. Always defined.

Rₙ(x) = f⁽ⁿ⁺¹⁾(ξ) (x−a)ⁿ⁺¹ / (n+1)!

  • Lagrange’s remainder. ξ lies between a and x. Requires f ∈ Cⁿ⁺¹ on interval containing a and x.

|Rₙ(x)| ≤ M |x−a|ⁿ⁺¹ / (n+1)!

  • Error bound if |f⁽ⁿ⁺¹⁾(t)| ≤ M on interval between a and x.

eˣ = Σ_{k=0}^∞ xᵏ / k!

  • Maclaurin series for exponential. Converges ∀ x ∈ ℝ.

sin x = Σ_{k=0}^∞ (−1)ᵏ x²ᵏ⁺¹ / (2k+1)!

  • Maclaurin for sine. Converges ∀ x ∈ ℝ.

cos x = Σ_{k=0}^∞ (−1)ᵏ x²ᵏ / (2k)!

  • Maclaurin for cosine. Converges ∀ x ∈ ℝ.

1 / (1−x) = Σ_{k=0}^∞ xᵏ, |x| < 1 (1+x)^m = 1 + m x + m(m−1) x²/2! + m(m−1)(m−2) x³/3! + …, valid for real m with |x| < 1

  • Geometric series. Converges for |x| < 1.

ln(1+x) = Σ_{k=1}^∞ (−1)ᵏ⁺¹ xᵏ / k, −1 < x ≤ 1

  • Maclaurin for natural log around 0. Convergence as stated.

arctan x = Σ_{k=0}^∞ (−1)ᵏ x²ᵏ⁺¹ / (2k+1), |x| ≤ 1

  • Maclaurin for arctan with endpoint convergence.

Radius of convergence R = 1 / limsup_{k→∞} |a_k|^{1/k}

  • For a power series Σ a_k (x−a)ᵏ. Dimensionless x. Applies to determine convergence interval.

Ratio test: L = lim_{k→∞} |a_{k+1}/a_k|; converges if |x−a| < 1/L

  • Practical test for radius when the limit exists.

Linearization: f(x) ≈ f(a) + f′(a)(x−a)

  • First-order Taylor approximation near x = a.

Worked examples

Example 1: Maclaurin expansion and error bound for eˣ at x = 0.5 up to x³ term Given: Approximate eˣ at x = 0.5 using the Maclaurin polynomial up to degree 3. Estimate the magnitude of the remainder using Lagrange’s bound.

Steps:

  1. Write the formula up to degree 3: f(x) = eˣ, so f⁽ᵏ⁾(0) = 1 for all k. Formula: P₃(x) = 1 + x + x²/2! + x³/3!.
  2. Substitute x = 0.5: P₃(0.5) = 1 + 0.5 + (0.5)²/2 + (0.5)³/6.
  3. Compute numerically with units (dimensionless here): (0.5)² = 0.25; 0.25/2 = 0.125. (0.5)³ = 0.125; 0.125/6 ≈ 0.0208333. Sum: 1 + 0.5 + 0.125 + 0.0208333 = 1.6458333. Final approximation: e^{0.5} ≈ 1.6458333.
  4. Remainder bound: Lagrange remainder R₃(x) = f⁽⁴⁾(ξ) x⁴/4! for some ξ between 0 and 0.5. Here f⁽⁴⁾(ξ) = e^{ξ} ≤ e^{0.5} (monotonic). Use e^{0.5} < 1.65 for a bound (conservative). |R₃(0.5)| ≤ 1.65 × (0.5)⁴ / 24 = 1.65 × 0.0625 / 24 = 0.103125 / 24 ≈ 0.0042969.
  5. Reference value (for checking): e^{0.5} ≈ 1.64872 (dimensionless). Error of our polynomial ≈ 1.64872 − 1.6458333 ≈ 0.0028867, which is below the bound.

Answer: e^{0.5} ≈ 1.6458333 with estimated error magnitude ≤ 0.00430. Final value in bold: 1.6458333 ± 0.00430.

Example 2: Taylor series of ln x about a = 1 up to (x−1)³ and interval of convergence of the full series Given: Find the Taylor polynomial of ln x about x = 1 up to degree 3, and write the general series with its interval of convergence.

Steps:

  1. Compute derivatives at a = 1: f(x) = ln x f′(x) = 1/x ⇒ f′(1) = 1 f″(x) = −1/x² ⇒ f″(1) = −1 f‴(x) = 2/x³ ⇒ f‴(1) = 2
  2. Write P₃(x): P₃(x) = f(1) + f′(1)(x−1) + f″(1)(x−1)²/2! + f‴(1)(x−1)³/3! = 0 + 1·(x−1) + (−1)·(x−1)²/2 + 2·(x−1)³/6 = (x−1) − (x−1)²/2 + (x−1)³/3.
  3. General Taylor series using substitution u = x−1: ln x = ln(1+u) = Σ_{k=1}^∞ (−1)^{k+1} u^k / k, for −1 < u ≤ 1 (with conditional convergence at u = 1). Therefore, ln x = Σ_{k=1}^∞ (−1)^{k+1} (x−1)^k / k.
  4. Interval of convergence in x: −1 < x−1 ≤ 1 ⇒ 0 < x ≤ 2. Note ln x is defined for x > 0. At x = 0 (left end), the series diverges with the function being undefined; at x = 2, the series converges conditionally to ln 2.

Answer: P₃(x) = (x−1) − (x−1)²/2 + (x−1)³/3; full series ln x = Σ_{k=1}^∞ (−1)^{k+1} (x−1)^k/k converges for 0 < x ≤ 2.

Example 3: Determine the Maclaurin series of sin x · eˣ up to x⁵ and estimate the error term order Given: Expand f(x) = eˣ sin x as a Maclaurin series up to x⁵. Indicate the order of the neglected terms.

Steps:

  1. Known Maclaurin series: eˣ = 1 + x + x²/2! + x³/3! + x⁴/4! + x⁵/5! + O(x⁶) sin x = x − x³/3! + x⁵/5! + O(x⁷)
  2. Multiply and collect up to x⁵: f(x) = (1 + x + x²/2 + x³/6 + x⁴/24 + x⁵/120)(x − x³/6 + x⁵/120) + O(x⁶) Keep products whose total power ≤ 5:
    • x term: 1·x = x
    • x² term: x·x = x²
    • x³ terms: (x²/2)·x + 1·(−x³/6) = (1/2 − 1/6)x³ = (1/3)x³
    • x⁴ terms: (x³/6)·x + x·(−x³/6) = (1/6 − 1/6)x⁴ = 0·x⁴ = 0
    • x⁵ terms: (x⁴/24)·x + (x²/2)·(−x³/6) + 1·(x⁵/120) = (1/24 − 1/12 + 1/120) x⁵ = (5/120 − 10/120 + 1/120) x⁵ = (−4/120) x⁵ = (−1/30) x⁵
  3. Write result with error notation and interpretation: eˣ sin x = x + x² + (1/3)x³ − (1/30)x⁵ + O(x⁶). The neglected terms are of order x⁶ and higher. If desired, the Lagrange form of remainder after degree 5 involves f⁽⁶⁾(ξ) x⁶/6! for some ξ between 0 and x.

Answer: eˣ sin x ≈ x + x² + (1/3)x³ − (1/30)x⁵ with neglected terms of order O(x⁶). Final polynomial in bold: x + x² + (1/3)x³ − (1/30)x⁵.

Common mistakes

  • Mixing up signs in alternating series like sin x or ln(1+x).
  • Using the Maclaurin series outside its interval of convergence, especially for ln(1+x) at x = −1 or beyond.
  • Assuming convergence implies equality to the function; forgetting to justify Rₙ(x) → 0.
  • Dropping higher-order terms incorrectly when multiplying two series; not tracking to the requested order.
  • Confusing the expansion point a; substituting x directly into coefficients f⁽ᵏ⁾(a) when a ≠ 0.
  • Not bounding derivatives correctly when using the Lagrange error bound; using M from outside the interval between a and x.
  • Writing units where not applicable, or ignoring dimensions when approximating functions of physical variables; remember polynomial terms must have the same physical dimensions as f(x).

For your exam

  • Short notes: Radius of convergence; use the ratio test to find it, and mention behavior at endpoints separately.
  • Derivations: Taylor’s theorem with Lagrange remainder (outline acceptable unless explicitly asked for full proof). Include Rolle’s/Mean Value Theorem steps and conditions.
  • Expansions: Compute Maclaurin/Taylor expansions up to a specified order for standard functions (eˣ, sin x, cos x, ln(1+x), (1+x)^m for real m via binomial series). Show derivatives at the point, substitute, and simplify. State interval of convergence when asked.
  • Error estimation: Use |Rₙ(x)| ≤ M |x−a|ⁿ⁺¹/(n+1)!. Identify M by bounding |f⁽ⁿ⁺¹⁾(t)| on the interval between a and x.
  • Combined series: Multiply or divide two known series to required order; justify truncation order and collect like powers carefully.
  • Applications: Linearization about an operating point, or approximating a function value numerically to given accuracy. State the approximation and error bound clearly.

Quick check

  1. Write the Maclaurin series of (1−x)^{−1} and state its interval of convergence.
  2. Find the Taylor polynomial of degree 2 for f(x) = cos x about a = π/3.
  3. Using Lagrange’s remainder, give an upper bound for the error when approximating sin 0.2 by 0.2 − 0.2³/6.

Answers: (1) 1 + x + x² + x³ + … for |x| < 1. (2) P₂(x) = cos(π/3) + (−sin(π/3))(x−π/3) + (−cos(π/3))(x−π/3)²/2 = 1/2 − (√3/2)(x−π/3) − (1/4)(x−π/3)². (3) |R₃(0.2)| ≤ max_{t∈[0,0.2]}|sin t|·(0.2)⁴/4! ≤ 1·0.0016/24 ≈ 0.0000667.

Finished this topic? Mark it so your progress, study plan and readiness keep up.

Stuck on something here?