Global extrema — the absolute maximum and minimum values of a function on a specified domain — are tested repeatedly on the AP Calculus AB and BC exams, both in the multiple-choice section and as free-response questions. The phrase test candidates for global extrema refers to a specific workflow: identify every place the function could reach its highest or lowest value, evaluate the function at each, and compare the results. On a closed interval, the workflow is mechanical. On an open interval or the entire real line, it requires a slightly different argument. Mastering this workflow is one of the highest-leverage skills in Unit 5 of the AP Calculus curriculum, and it reappears as a building block in optimisation, related rates, and accumulation FRQs in later units.
The closed-interval test for absolute extrema: the standard AP workflow
When a continuous function f is defined on a closed interval [a, b], the Extreme Value Theorem guarantees that f attains both an absolute maximum and an absolute minimum somewhere on that interval. The College Board expects candidates to apply a single, repeatable procedure whenever they see a closed interval in the problem statement. The procedure has three moves, and missing any one of them is the most common source of point loss on absolute-value FRQs.
Move one is to take the derivative and solve f′(x) = 0 inside the open interval (a, b). These are the interior critical points. On a typical AP problem, a polynomial of degree three or four will yield one or two such points, and a trigonometric or exponential expression will yield points that you can express in radians or as natural logs. Move two is to find points where f′ is undefined — vertical tangents, cusps, or corners — that still sit inside the closed interval. Move three is to evaluate f at the endpoints a and b, and at every number collected in the first two moves. The largest value is the absolute maximum; the smallest is the absolute minimum.
A worked example: polynomial on a closed interval
Consider f(x) = x³ − 3x² − 9x + 5 on [−2, 6]. The derivative f′(x) = 3x² − 6x − 9 = 3(x² − 2x − 3) = 3(x − 3)(x + 1). Setting f′(x) = 0 gives x = 3 and x = −1. Both sit inside [−2, 6], so both are critical points. The derivative exists everywhere, so there are no points where f′ is undefined. The four evaluation points are x = −2, −1, 3, and 6. Compute f(−2) = 7, f(−1) = 10, f(3) = −22, and f(6) = 59. The absolute maximum is 59 at x = 6; the absolute minimum is −22 at x = 3. Notice that the absolute maximum occurred at an endpoint, not a critical point. A candidate who records only x = −1 and x = 3 in the calculator and never types f(−2) and f(6) will mistakenly declare the maximum to be 10.
A worked example: trigonometric function on a closed interval
Now consider g(x) = 2 sin(x) + x on [0, 2π]. Here g′(x) = 2 cos(x) + 1. Setting g′(x) = 0 gives cos(x) = −½, so x = 2π/3 and x = 4π/3 inside the interval. Both derivatives exist, so the only evaluation points are the two critical points plus the two endpoints. The values g(0) = 0, g(2π/3) ≈ 3.83, g(4π/3) ≈ −0.48, g(2π) ≈ 6.28. The absolute maximum is 2π at x = 2π; the absolute minimum is 2 sin(4π/3) + 4π/3 ≈ −0.48 at x = 4π/3. A graphing calculator in radian mode confirms both answers. The lesson: with trig expressions, always confirm the calculator is in radians, and always state the critical points to the nearest 0.001 rather than rounding prematurely.
Critical points: how to find them and how to avoid the textbook traps
A critical point of f is a value x = c in the domain of f where f′(c) = 0 or f′(c) does not exist. Candidates often misread this definition in two ways. First, x = c must belong to the domain. A point where the original function is undefined is never a critical point of f, even if the simplified derivative accidentally equals zero there. Second, “f′(c) does not exist” must be visible in the original derivative expression — a vertical tangent, a corner, or a cusp. Simply failing to factor a polynomial does not make the derivative undefined; it makes it unwritten. In an AP setting, the College Board distinguishes these cases in the FRQ rubric, and an unsupported claim that “f′ is undefined at x = 0” usually loses a point.
Where critical points hide in AP problems
- Inside absolute values and piecewise definitions. For f(x) = |x² − 4|, the derivative fails to exist at x = ±2 because of the corner. These are critical points even though f is continuous and differentiable almost everywhere.
- At endpoints of piecewise functions. A piecewise function defined on [0, 5] with a different formula on [0, 3] and [3, 5] may have a corner at x = 3 worth checking, even when the formula change is signposted in the problem.
- At endpoints of the closed interval itself. Endpoints are not critical points in the strict sense, but they are evaluation points. Conflating the two is a scoring error in rubric language.
- Where natural log arguments vanish. For f(x) = ln(x² − 1), the domain excludes x = ±1. These domain holes are not critical points of f, and including them as evaluation candidates will cost a point.
The difference between critical points and inflection points
Candidates preparing for the AP exam sometimes confuse critical points (where f′ = 0 or undefined) with inflection points (where f″ = 0 and concavity changes). The two are unrelated in general. A critical point tests for local or global extrema. An inflection point tests for concavity. Several FRQs in the released problem sets put both on the same page, and the rubric allocates the points separately. A clean, table-style answer that lists the x-values, the function value, the derivative value, and the second derivative value side by side will make the reader's job easier and the candidate's score higher.
Open intervals and the entire real line: when the closed-interval test does not apply
The Extreme Value Theorem has two hypotheses: f must be continuous, and the domain must be a closed and bounded interval. Drop either hypothesis, and the theorem is no longer a guarantee. The function f(x) = x³ has no absolute maximum or minimum on (−∞, ∞) because the values grow without bound in both directions. The function f(x) = 1/x on (0, 1] is continuous but never attains a maximum because the supremum is reached only in the limit as x approaches 0, and 0 is not in the domain. These are the situations where AP candidates must argue from the shape of the function rather than plug in numbers.
Argument patterns for unbounded domains
When the domain is open or unbounded, the exam expects an argument — not just a calculator read-out. Three argument patterns cover most AP scenarios. The first is end behaviour: if lim x→∞ f(x) = ∞ and lim x→−∞ f(x) = ∞, the function has no absolute maximum; if the limits differ, the function has no global extremum in the unbounded direction. The second is monotonicity: if f′(x) > 0 on the entire domain, f is strictly increasing and has no absolute maximum or minimum, only a supremum or infimum at the boundary of the domain. The third is limit comparison: as x approaches a vertical asymptote from one side, f(x) can grow without bound, ruling out a global maximum on that side.
Worked example: function on an open interval
Consider h(x) = x · e^(−x) on (0, ∞). The derivative h′(x) = e^(−x) − x · e^(−x) = e^(−x)(1 − x). Setting h′ = 0 gives x = 1. The function is positive, increasing on (0, 1), decreasing on (1, ∞), and tends to 0 as x → ∞. So h has an absolute maximum of 1/e at x = 1, but it has no absolute minimum — the infimum is 0, attained only in the limit. The AP-style answer would state both findings explicitly: maximum 1/e at x = 1; no absolute minimum because h(x) > 0 for all x in the domain. A candidate who reports only the maximum and ignores the minimum is half-right and loses a point.
First and Second Derivative Tests as global tools
The First Derivative Test classifies a critical point as a local maximum, local minimum, or neither by tracking the sign of f′ on either side. The Second Derivative Test classifies a critical point as a local maximum or local minimum by inspecting the sign of f″ at the critical point, provided f″ exists and is nonzero. Both tests are local, but on a closed interval with only a few critical points, they often pinpoint the global extremum without exhaustive comparison. The trade-off is between algebraic effort and tabulating effort. For most AP problems, the safest global workflow is still to evaluate f at every candidate point and compare. The derivative tests are best used to confirm the answer, or as a fallback when the evaluation table is long.