Opening subject page...
Loading your content
A systematic method for finding the largest and smallest values a continuous function attains on a closed interval.
Optimization—the quest to find maximum or minimum values of quantities—has been a driving force in mathematics for centuries. Long before calculus was formalized, ancient Greek geometers like Euclid and Apollonius investigated problems of maximizing areas and minimizing distances, but they lacked a general analytical framework. The development of differential calculus in the seventeenth century by Newton and Leibniz finally provided the tools needed to address optimization rigorously, yet it took further theoretical refinement to guarantee that solutions actually exist. The Candidates Test (sometimes called the Closed Interval Method) emerged from this tradition as a direct, elegant procedure for locating absolute (global) extrema of continuous functions on closed intervals, relying on the foundational Extreme Value Theorem.
With the Extreme Value Theorem established, a natural question arose: given that a continuous function on [a, b] must attain a global maximum and minimum, how do we find them efficiently? The Candidates Test answers precisely this question by narrowing the search to a finite, manageable set of candidate points—the critical numbers and the endpoints of the interval.
The Candidates Test rests on a small but powerful chain of ideas. Before applying the method, it is essential to understand each component precisely, since a missing hypothesis—such as a function that is not continuous or an interval that is not closed—would invalidate the entire procedure. The following foundational concepts form the logical scaffolding of the test.
The following diagram illustrates a continuous function on the closed interval [a, b] with three critical numbers labeled c₁, c₂, and c₃. Observe that the absolute maximum and absolute minimum must occur at one of the candidate points—the endpoints or the critical numbers. By evaluating f at each of these five points and comparing the resulting values, we can identify the global extrema without needing to examine any other points on the curve.
Notice that the absolute maximum in this example occurs at an interior critical number c₃ where f′(c₃) = 0, while the absolute minimum occurs at another interior critical number c₂. Neither extremum occurs at an endpoint—this is perfectly valid. In other functions, the absolute extrema might both occur at endpoints, or one at an endpoint and one at a critical number. The power of the Candidates Test is that it does not require you to predict where the extrema fall; you simply evaluate all candidates and let the largest and smallest values reveal themselves.
The Candidates Test can be expressed as a precise algorithmic procedure. Given a function f that is continuous on [a, b], the method proceeds in three stages: finding all critical numbers, evaluating the function at every candidate, and comparing values. The theoretical justification comes from combining the Extreme Value Theorem (which guarantees existence) with the fact that any interior absolute extremum must also be a local extremum, which in turn must occur at a critical number by Fermat's Theorem.
The following flowchart summarizes the decision process of the Candidates Test, from verifying the hypotheses through to stating the final answer. Each node represents a checkpoint or computation step. Before beginning the test, you must confirm the two prerequisites: (1) f is continuous on [a, b], and (2) the interval is closed and bounded. If either condition fails, the Extreme Value Theorem does not apply, and the Candidates Test cannot guarantee an answer.
| Candidate Type | How to Identify | Why It's a Candidate |
|---|---|---|
| Endpoint x = a | Given by the interval | Boundary of domain; not subject to Fermat's Theorem |
| Endpoint x = b | Given by the interval | Boundary of domain; not subject to Fermat's Theorem |
| Stationary point | Solve f′(x) = 0 in (a, b) | Horizontal tangent; derivative changes sign or has zero slope |
| Singular point | Find where f′(x) DNE in (a, b), but f(x) exists | Cusp, corner, or vertical tangent; derivative undefined but function value exists |
Let us apply the Candidates Test to find the absolute maximum and absolute minimum of f(x) = 2x³ − 3x² − 12x + 1 on the interval [−2, 3]. This polynomial is continuous everywhere, so the hypotheses of the Extreme Value Theorem are satisfied on any closed interval.
Calculus provides several tools for identifying extrema, and students sometimes conflate the Candidates Test with derivative tests for local extrema. It is important to understand when each method is appropriate and what conclusions each one supports. The Candidates Test is specifically designed for absolute (global) extrema on closed intervals, while the First and Second Derivative Tests address local (relative) extrema at individual critical numbers.
| Feature | Candidates Test | First Derivative Test | Second Derivative Test |
|---|---|---|---|
| Purpose | Find absolute max and min on [a, b] | Classify critical numbers as local max, local min, or neither | Classify critical numbers where f′(c) = 0 and f″(c) ≠ 0 |
| Requires closed interval? | Yes—essential | No—works on any interval | No—works at any interior critical number |
| Uses endpoints? | Yes—always evaluates f(a) and f(b) | No—focuses on sign changes of f′ | No—uses f″(c) only |
| Conclusion type | Global (absolute) extrema | Local (relative) extrema only | Local (relative) extrema only |
| Limitations | Only applies on closed intervals with continuous functions | Requires analyzing sign of f′ on intervals around each critical number | Inconclusive when f″(c) = 0; requires f″ to exist |
The Candidates Test provides a complete solution for finding absolute extrema of continuous functions on closed intervals, but many real-world optimization problems do not fit neatly into this framework. In multivariable calculus, the analogous procedure uses Lagrange multipliers and examines boundary curves of closed, bounded regions in ℝ². In applied optimization within single-variable calculus, you often encounter problems on open or unbounded domains where the Extreme Value Theorem does not directly apply, requiring additional arguments—such as behavior as x → ±∞ or the sole critical point test—to verify that a local extremum is actually global.
| Scenario | Candidates Test Applicable? | Alternative Approach |
|---|---|---|
| f continuous on [a, b] | Yes — ideal case | N/A |
| f continuous on (a, b) — open interval | No | Evaluate limits as x → a⁺ and x → b⁻; compare with critical values |
| f on (−∞, ∞) — unbounded domain | No | Analyze end behavior via limits; use sole critical point theorem if applicable |
| f has a discontinuity in [a, b] | No | Analyze one-sided limits at discontinuities; split into subintervals if possible |
| f(x, y) continuous on closed, bounded region in ℝ² | Generalized version applies | Lagrange multipliers for boundary; set ∇f = 0 for interior |
For the AP Calculus BC exam, applied optimization problems frequently require you to first set up the function and interval from context, then apply the Candidates Test (or, for open domains, justify global extrema through other means). Recognizing when the Candidates Test applies—and when it does not—is itself a testable skill. In later courses such as real analysis, you will encounter the full proof of the Extreme Value Theorem using the completeness property of the real numbers, providing a deeper appreciation for why the closed-interval hypothesis is non-negotiable.
The Candidates Test is a systematic method for finding the absolute (global) maximum and absolute (global) minimum of a continuous function on a closed interval [a, b]. Its validity rests on the Extreme Value Theorem, which guarantees that such extrema exist. The procedure has three steps: (1) find all critical numbers in the open interval (a, b) by solving f′(x) = 0 and identifying where f′(x) does not exist; (2) evaluate f at each critical number and at both endpoints; and (3) compare all values to identify the largest and smallest.
This method is distinct from the First Derivative Test and Second Derivative Test, which classify local (relative) extrema at individual critical numbers. The Candidates Test is the go-to strategy whenever an AP problem asks for absolute extrema on a closed interval, and it requires no sign charts or second derivatives—just careful evaluation and comparison. Remember: if the interval is not closed, or if f is not continuous, the Extreme Value Theorem does not apply, and you must use alternative reasoning to determine whether absolute extrema exist and where they occur.