BUSINESS CALCULUS • INTEGRATION

Trapezoidal Rule

Approximating definite integrals by summing trapezoidal areas when antiderivatives are impractical or impossible to find.

Historical Context & Motivation

The problem of computing areas bounded by curves predates modern calculus by millennia. Ancient mathematicians recognized that curved regions could be approximated by subdividing them into simpler geometric shapes whose areas were already known. The Trapezoidal Rule represents one of the most intuitive and enduring strategies in this tradition: replace a curve with a series of straight-line segments, then sum the areas of the resulting trapezoids. Although the Fundamental Theorem of Calculus provides an elegant exact-answer framework, many functions encountered in business analytics—demand curves fitted to empirical data, probability density functions without closed-form antiderivatives, or revenue streams sampled at discrete time intervals—resist symbolic integration entirely. In these situations, numerical approximation is not merely a convenience but an operational necessity.

~250 BCE
Archimedes' Method of Exhaustion
Archimedes approximated areas under parabolic arcs by inscribing and circumscribing polygons, establishing the conceptual foundation for all numerical integration methods.
1670s
Newton & Leibniz Formalize Calculus
The independent development of calculus by Newton and Leibniz provided the antiderivative framework, but also revealed that many integrals cannot be evaluated in closed form.
1743
Thomas Simpson Publishes Quadrature Methods
Simpson systematized several numerical quadrature rules, including the trapezoidal method, and compared their accuracy—helping establish numerical analysis as a distinct discipline.
1950s–Present
Digital Computing Era
Electronic computers made trapezoidal and more advanced quadrature rules practical for large-scale business modeling, financial simulation, and real-time data integration.

The central question the Trapezoidal Rule addresses is deceptively simple: given a continuous function f(x) on a closed interval [a, b], how can we estimate the value of ∫ f(x) dx when we either lack a closed-form antiderivative or only possess the function's values at a finite set of sample points? This question arises routinely in business contexts—estimating total revenue from a time-varying revenue rate, approximating consumer surplus from an empirical demand curve, or computing the present value of a cash-flow stream sampled at quarterly intervals.

Core Principles & Definitions

Before diving into the formula, it is essential to understand the geometric intuition and foundational assumptions that underpin the Trapezoidal Rule. The method rests on a straightforward idea: between any two adjacent sample points, the function's graph is approximated by a straight line segment rather than a curve. The region between that line segment and the x-axis forms a trapezoid whose area can be computed with elementary geometry. Summing these trapezoidal areas across the entire interval yields an estimate of the definite integral.

1

Partition the Interval

Divide [a, b] into n equal subintervals, each of width Δx = (b − a) / n. The partition points are x₀ = a, x₁, x₂, …, xₙ = b.
2

Connect Adjacent Points

At each partition point, evaluate f(xᵢ). Connect consecutive points (xᵢ, f(xᵢ)) and (xᵢ₊₁, f(xᵢ₊₁)) with a straight line, forming the top edge of a trapezoid.
3

Compute Trapezoidal Areas

The area of a trapezoid with parallel sides h₁ and h₂ and width w is ½(h₁ + h₂)w. Here h₁ = f(xᵢ), h₂ = f(xᵢ₊₁), and w = Δx.
4

Sum All Trapezoids

Add the areas of all n trapezoids to obtain the numerical estimate Tₙ ≈ ∫ₐᵇ f(x) dx. The endpoints contribute once; interior points contribute twice.
5

Refine for Accuracy

Increasing n (more subintervals) reduces the approximation error. The error decreases proportionally to 1/n², so doubling n roughly quadruples the accuracy.
KEY TAKEAWAY
Think of the Trapezoidal Rule like estimating the total area of an irregularly shaped parking lot by laying a series of rectangular tarps across it, each tilted to match the lot's edges at two adjacent survey stakes. No single tarp fits perfectly, but together they approximate the lot's total area. The more stakes you drive—the more subintervals you use—the more closely the tarps conform to the actual boundary, and the better your estimate becomes.

Visual Explanation

The diagram below illustrates the Trapezoidal Rule applied to a smooth curve using four subintervals (n = 4). The shaded trapezoids approximate the area between the curve and the x-axis. Notice how each trapezoid's top edge is a straight line connecting two consecutive function values, and the approximation error corresponds to the thin gaps or overlaps between the trapezoid tops and the actual curve.

Each colored trapezoid spans one subinterval Δx. The green curve shows f(x); the straight top edges of the trapezoids are linear interpolations between successive sample points. The sum of all four trapezoidal areas approximates ∫ₐᵇ f(x) dx.

Observe that where the curve is concave up (bowing upward), the chord connecting two points lies above the curve, so the trapezoid overestimates the area, and where the curve is concave down (bowing downward), the chord lies below the curve, so the trapezoid underestimates it. This interplay between concavity and error direction is a critical insight for understanding when the Trapezoidal Rule will overestimate or underestimate a given integral, and it motivates the error-bound formula we will derive in Section 5.

Mathematical Framework

The derivation of the Trapezoidal Rule begins with the area formula for a single trapezoid and extends it across n subintervals. Consider a single subinterval [xᵢ, xᵢ₊₁] of width Δx. The area of the trapezoid formed by the x-axis, the vertical lines x = xᵢ and x = xᵢ₊₁, and the line segment connecting (xᵢ, f(xᵢ)) to (xᵢ₊₁, f(xᵢ₊₁)) is given by ½ · Δx · [f(xᵢ) + f(xᵢ₊₁)]. Summing over all n subintervals and factoring out the common Δx/2 yields the complete formula.

SUBINTERVAL WIDTH
Δx = (b − a) / n
where a and b are the integration limits, and n is the number of equal subintervals.
TRAPEZOIDAL RULE FORMULA
Tₙ = (Δx / 2) × [f(x₀) + 2f(x₁) + 2f(x₂) + ⋯ + 2f(xₙ₋₁) + f(xₙ)]
The first and last function values carry a coefficient of 1; all interior values carry a coefficient of 2 because each interior point is shared by two adjacent trapezoids.
COMPACT SUMMATION FORM
Tₙ = (Δx / 2) × [f(x₀) + f(xₙ) + 2 × Σᵢ₌₁ⁿ⁻¹ f(xᵢ)]
This equivalent form is often more convenient for programming or spreadsheet implementation, as the endpoints and interior points are handled separately.

To understand why interior points are doubled, consider two adjacent trapezoids sharing the boundary at xᵢ. The right side of the left trapezoid contributes f(xᵢ), and the left side of the right trapezoid also contributes f(xᵢ). When the individual trapezoidal areas are summed, these contributions combine into 2f(xᵢ). Only the very first point x₀ and the very last point xₙ appear in exactly one trapezoid, so they carry coefficient 1. This 1-2-2-…-2-1 coefficient pattern is the hallmark of the Trapezoidal Rule and distinguishes it from other quadrature methods.

Error Bound & Accuracy

An approximation is only useful if we can quantify how far it might deviate from the true answer. The error bound for the Trapezoidal Rule is derived via Taylor series expansion and depends on the second derivative of f, since the method exactly integrates linear functions (whose second derivatives vanish) and incurs error only from the curvature of f.

TRAPEZOIDAL RULE ERROR BOUND
|Eₜ| ≤ (b − a)³ / (12n²) × max |f″(x)| for x ∈ [a, b]
Here Eₜ is the error Tₙ − ∫ₐᵇ f(x) dx, and max |f″(x)| is the largest absolute value of the second derivative on [a, b]. The error shrinks as O(1/n²), meaning doubling n reduces the maximum error by a factor of four.
As the number of subintervals n increases, the maximum error decreases proportionally to 1/n². Each doubling of n reduces the error bound by approximately a factor of four, illustrating the quadratic convergence rate of the Trapezoidal Rule.
Concavity & Error Direction
If f″(x) > 0 on [a, b] (the function is concave up), the trapezoids lie above the curve and the Trapezoidal Rule overestimates the integral. If f″(x) < 0 (concave down), the rule underestimates. When concavity changes sign across the interval, errors partially cancel.

Worked Example

A company's marginal revenue function is given by R′(x) = 0.5x² + 2x + 10 (in dollars per unit), and management wants to estimate the total revenue earned between producing 0 and 4 units. Use the Trapezoidal Rule with n = 4 subintervals to approximate ∫₀⁴ (0.5x² + 2x + 10) dx.

Estimating Total Revenue with the Trapezoidal Rule
1
Step 1 — Identify Given ValuesWe have a = 0, b = 4, n = 4, and f(x) = 0.5x² + 2x + 10. Our partition points will be x₀, x₁, x₂, x₃, x₄.
2
Step 2 — Compute ΔxΔx = (b − a) / n = (4 − 0) / 4 = 1. So the partition points are x₀ = 0, x₁ = 1, x₂ = 2, x₃ = 3, x₄ = 4.
Δx = 1
3
Step 3 — Evaluate f at Each Partition Pointf(0) = 0.5(0)² + 2(0) + 10 = 10. f(1) = 0.5(1) + 2 + 10 = 12.5. f(2) = 0.5(4) + 4 + 10 = 16. f(3) = 0.5(9) + 6 + 10 = 20.5. f(4) = 0.5(16) + 8 + 10 = 26.
f-values: 10, 12.5, 16, 20.5, 26
4
Step 4 — Apply the Trapezoidal Rule FormulaT₄ = (Δx / 2) × [f(x₀) + 2f(x₁) + 2f(x₂) + 2f(x₃) + f(x₄)] = (1/2) × [10 + 2(12.5) + 2(16) + 2(20.5) + 26] = (1/2) × [10 + 25 + 32 + 41 + 26] = (1/2) × 134 = 67.
T₄ = 67 dollars
5
Step 5 — Compare with the Exact IntegralThe exact value is ∫₀⁴ (0.5x² + 2x + 10) dx = [x³/6 + x² + 10x]₀⁴ = (64/6 + 16 + 40) − 0 = 10.667 + 16 + 40 = 66.667 (rounded). The Trapezoidal Rule yields 67, giving an error of approximately 0.333, which is only about 0.50% relative error.
Exact value ≈ 66.667; Error ≈ 0.333
💡 Why the Overestimate?
Since f″(x) = 1 > 0 for all x, the function is concave up on [0, 4], so the trapezoidal approximation lies above the curve everywhere. Consistent with the error bound, T₄ overestimates the integral. Using n = 8 would reduce this overestimate by roughly a factor of four.

Strengths, Limitations & Comparisons

The Trapezoidal Rule sits between the simpler Riemann sum methods and the more accurate Simpson's Rule in the hierarchy of numerical integration techniques. Understanding its relative strengths and weaknesses helps practitioners choose the right tool for a given business application.

Comparison of common numerical integration methods
CriterionLeft/Right Riemann SumsTrapezoidal RuleSimpson's Rule
Geometric shape usedRectanglesTrapezoidsParabolic arcs
Error orderO(1/n)O(1/n²)O(1/n⁴)
Exact for polynomials up to degree0 (constants)1 (linear)3 (cubic)
Requires n to be even?NoNoYes
Ease of implementationVery simpleSimpleSlightly more complex
Best use caseQuick estimates, conceptual understandingDiscrete data, moderate accuracy needsSmooth functions, high accuracy needs
KEY TAKEAWAY
The Trapezoidal Rule occupies a practical sweet spot: it is nearly as simple to implement as Riemann sums but converges much faster (quadratically vs. linearly). In business contexts where you are working with tabular data—quarterly earnings, monthly sales figures, or hourly production rates—the Trapezoidal Rule is often the default choice because it requires no assumptions about the underlying function beyond continuity. Simpson's Rule is more accurate for smooth analytical functions, but it demands an even number of subintervals and assumes the data follows parabolic arcs between points, which may not be appropriate for noisy empirical data.

Connection to Advanced Methods

The Trapezoidal Rule is not merely a standalone technique; it is the foundation from which several more powerful methods are built. Understanding these connections provides a roadmap for future study in numerical analysis and computational mathematics, areas that are increasingly important in quantitative business fields such as operations research, financial engineering, and actuarial science.

Trapezoidal Rule vs. Romberg Integration
FeatureTrapezoidal RuleRomberg Integration
Basic ideaSum trapezoidal areas for a fixed nCombine trapezoidal estimates for different n values using Richardson extrapolation
Convergence rateO(1/n²)O(1/n²ᵏ) for k levels of refinement
Computational costn + 1 function evaluationsReuses previous evaluations; minimal additional cost per level
When to useQuick estimates, discrete data tablesHigh precision needed from smooth functions

Additionally, the Trapezoidal Rule relates directly to Simpson's Rule through a weighted average. If T₁ is the trapezoidal estimate with n subintervals and T₂ is the estimate with 2n subintervals, then Simpson's Rule equals (4T₂ − T₁)/3. This elegant relationship means that anyone who understands the Trapezoidal Rule can bootstrap their way to Simpson's Rule without memorizing a new formula. In adaptive quadrature algorithms used in professional software (Excel's numerical solvers, Python's SciPy library, MATLAB), the trapezoidal estimate is computed repeatedly at finer and finer scales, and the results are intelligently combined until a user-specified error tolerance is met.

Practice Problems

PROBLEM 1CONCEPTUAL
Explain why the Trapezoidal Rule with any number of subintervals will give the exact value of ∫₁⁵ (3x + 7) dx. What property of the integrand makes this possible?
PROBLEM 2BASIC CALCULATION
Use the Trapezoidal Rule with n = 3 to approximate ∫₀⁶ x² dx. Then compute the exact value and find the absolute error.
PROBLEM 3INTERMEDIATE
A company measures its rate of daily profit (in thousands of dollars) at the start of each quarter: P′(0) = 5, P′(3) = 8, P′(6) = 12, P′(9) = 9, P′(12) = 6, where the argument is measured in months. Use the Trapezoidal Rule with these five data points to estimate total profit over the 12-month period.
PROBLEM 4APPLIED
A financial analyst needs to estimate ∫₀² e^(−x²) dx (a function with no elementary antiderivative) to within an error of at most 0.01. Using the error bound |Eₜ| ≤ (b − a)³/(12n²) × max|f″(x)|, determine the minimum number of subintervals n required. Note: f″(x) = (4x² − 2)e^(−x²), and |f″(x)| ≤ 2 on [0, 2].
PROBLEM 5CRITICAL THINKING
Suppose you apply the Trapezoidal Rule to ∫ₐᵇ f(x) dx with n subintervals and obtain T_n, then repeat with 2n subintervals and obtain T_{2n}. Show that Simpson's Rule applied with 2n subintervals can be recovered as S_{2n} = (4T_{2n} − T_n) / 3. Why does this weighted combination achieve O(1/n⁴) convergence instead of O(1/n²)?

Lesson Summary

The Trapezoidal Rule approximates a definite integral ∫ₐᵇ f(x) dx by partitioning [a, b] into n equal subintervals of width Δx = (b − a)/n, then summing the areas of the resulting trapezoids using the formula Tₙ = (Δx/2)[f(x₀) + 2f(x₁) + ⋯ + 2f(xₙ₋₁) + f(xₙ)]. The characteristic 1-2-2-…-2-1 coefficient pattern arises because interior partition points are shared by two adjacent trapezoids. The method is exact for linear functions and has an error that decreases as O(1/n²), with the maximum error bounded by (b − a)³/(12n²) × max|f″(x)|.

In business contexts, the Trapezoidal Rule is especially valuable because it works directly from tabular or discrete data—quarterly revenue figures, monthly cost rates, or empirical demand observations—without requiring an explicit formula for the integrand. Its concavity-error relationship (overestimate when concave up, underestimate when concave down) provides useful qualitative insight into the direction of the approximation error. For higher accuracy, the Trapezoidal Rule can be extended to Simpson's Rule via Richardson extrapolation, or iterated systematically in Romberg integration.

Varsity Tutors • Business Calculus • Trapezoidal Rule