Riemann Sum Approximations — Approximating Areas With Riemann Sums

Historical Context & Motivation

The problem of computing areas enclosed by curves is one of the oldest in mathematics, predating modern calculus by millennia. Ancient mathematicians recognized that curved regions resist the straightforward length-times-width formula available for rectangles, and they devised increasingly clever strategies to approximate these elusive areas. The idea at the heart of every such strategy — exhaustion — is to fill the region with simpler shapes whose areas can be computed exactly, then refine those shapes until the approximation converges to the true value. This is precisely the intuition behind Riemann sums, the formal machinery that places integration on rigorous footing.

The central question that Riemann sums answer is deceptively simple: given a function f defined on a closed interval [a, b], how can we assign a precise numerical value to the signed area between the curve y = f(x) and the x-axis? Riemann's insight was to reduce this geometric question to an algebraic one — partition the interval into finitely many subintervals, erect a rectangle on each subinterval whose height is determined by sampling f, sum the areas, and then examine what happens as the partition becomes infinitely fine.

Core Principles & Definitions

Before diving into calculations, it is essential to understand the structural components of a Riemann sum. Every Riemann sum is built from three ingredients: a partition of the interval, a set of sample points, and the subinterval widths. Together, these elements determine the collection of rectangles whose areas sum to approximate the integral.

Visual Explanation — Left, Right, and Midpoint Sums

The diagram below illustrates the three most common Riemann sum types — left endpoint, right endpoint, and midpoint — applied to the function f(x) = x² + 1 on the interval [0, 4] with n = 4 subintervals. Notice how the choice of sample point shifts each rectangle's height, producing different approximations of the area under the curve.

In the left-endpoint diagram, each rectangle's height is determined by the function value at the left edge of its subinterval, so the first rectangle has height f(0) = 1, the second has height f(1) = 2, and so on. Because f is increasing on [0, 4], every left-endpoint rectangle sits entirely below the curve, producing an underestimate. Conversely, right-endpoint rectangles extend above the curve, yielding an overestimate. The midpoint rule — where each rectangle's height is f evaluated at the center of the subinterval — often provides a more balanced approximation because the over- and under-estimation regions partially cancel.

Mathematical Framework

We now formalize the Riemann sum construction. Let f be a bounded function on the closed interval [a, b]. A partition P of [a, b] is a finite set of points a = x₀ < x₁ < x₂ < … < xₙ = b. For each i = 1, 2, …, n, the width of the i-th subinterval is Δxᵢ = xᵢ − xᵢ₋₁, and a sample point cᵢ ∈ [xᵢ₋₁, xᵢ] is selected.

For most computational work in this course, we employ a regular partition, where every subinterval has the same width. This simplification allows us to factor Δx out of the summation.

The definite integral is defined as the common limit of all these sums (and indeed any valid choice of sample points) as n → ∞ (equivalently, as ‖P‖ → 0):

Types of Riemann Sums & Error Analysis

Understanding when a Riemann sum over- or under-approximates the integral is critical for bounding errors without computing the exact integral. The behavior depends on both the monotonicity and concavity of the function. The table below summarizes the over/under behavior for the three standard sum types when f is positive.

For a function with a continuous first derivative on [a, b], the error of the left (or right) Riemann sum with n subintervals is bounded by |Eₙ| ≤ M₁(b − a)²/(2n), where M₁ = max|f′(x)| on [a, b]. This shows that the error decreases like O(1/n) — doubling the number of rectangles roughly halves the error. For the midpoint sum applied to a function with a continuous second derivative, the error bound improves to |Eₙ| ≤ M₂(b − a)³/(24n²), where M₂ = max|f″(x)|, yielding O(1/n²) convergence. This quadratic improvement is one reason the midpoint rule is often preferred in numerical applications.

Worked Example

Let us compute the left, right, and midpoint Riemann sums for f(x) = x³ on the interval [1, 3] using n = 4 subintervals, then compare each approximation with the exact integral.

Strengths, Limitations, & Comparisons

Riemann sums are foundational but not the only approach to numerical integration. In practice, the choice between Riemann sum variants and more sophisticated rules depends on the desired accuracy, the smoothness of the integrand, and computational cost. The following table provides a concise comparison.

Connection to the Definite Integral & Beyond

The Riemann sum is not merely a computational tool — it is the very definition of the definite integral for Riemann-integrable functions. Understanding how the finite sum transitions to the integral is essential for grasping subsequent topics in Calculus 2, including improper integrals, numerical integration techniques, and applications such as arc length, surface area, and work. The table below highlights the progression from Riemann sums to more advanced integration concepts.

The Fundamental Theorem of Calculus (FTC) provides the crucial bridge: if F is an antiderivative of f on [a, b], then ∫ₐᵇ f(x) dx = F(b) − F(a). The FTC means we rarely need to compute a Riemann sum limit directly in practice — but the Riemann sum formulation remains indispensable for setting up integrals in applied contexts. Whenever you slice a solid into cross-sections, decompose a force over a displacement, or partition a probability distribution, you are implicitly constructing a Riemann sum whose limit yields the integral you need.

Practice Problems

Summary