Can an endpoint be a local maximum? Understanding the role of boundaries in calculus
When studying functions and their extrema, one of the most common questions students encounter is whether a point that lies at the very edge of a domain—an endpoint—can qualify as a local maximum. The answer hinges on precise definitions, the nature of the interval under consideration, and the behavior of the function near that boundary. In this article we unpack the concept step by step, clarify the distinction between local and global extrema, and illustrate with concrete examples when an endpoint can, and cannot, serve as a local maximum Worth knowing..
Introduction: Setting the Stage
In calculus, a local maximum (also called a relative maximum) of a function (f) at a point (c) means that there exists some open interval ((c-\delta, c+\delta)) such that (f(x) \le f(c)) for all (x) in that interval where the function is defined. The key idea is that the function’s value at (c) is at least as large as the values of nearby points.
When the domain of (f) is restricted—say, to a closed interval ([a, b])—the points (a) and (b) are endpoints. In real terms, because they lack a full two‑sided neighbourhood, the standard definition of a local maximum must be examined carefully. The question “can an endpoint be a local maximum?” is therefore not merely a curiosity; it touches on the foundations of optimization problems, especially those arising in applied mathematics, economics, and engineering where constraints naturally create boundaries Nothing fancy..
Definitions: Local Maximum vs. Endpoint
What is a Local Maximum?
Formally, let (f: D \subseteq \mathbb{R} \to \mathbb{R}). Still, a point (c \in D) is a local maximum if there exists (\epsilon > 0) such that for all (x \in D) with (|x-c| < \epsilon), we have (f(x) \le f(c)). Notice that the condition only requires comparison with points inside the domain that lie within the epsilon‑neighbourhood of (c).
What is an Endpoint?
If the domain (D) is an interval, the endpoints are the smallest and largest elements of that interval (when they exist). Day to day, for a closed interval ([a, b]), the endpoints are (a) and (b). For an open interval ((a, b)) or a half‑open interval ([a, b)), only the included side counts as an endpoint; the other side is not part of the domain.
Because an endpoint lacks points on one side, the usual two‑sided neighbourhood cannot be centered at that point. Instead, we consider a one‑sided neighbourhood: for the left endpoint (a), we look at points (x) with (a \le x < a+\epsilon); for the right endpoint (b), we look at (b-\epsilon < x \le b).
Theoretical Background: Why Endpoints Can Be Local Maxima
Fermat’s Theorem and Its Limitations
Fermat’s theorem states that if (f) has a local extremum at an interior point (c) and (f) is differentiable at (c), then (f'(c)=0). This theorem does not apply to endpoints because the derivative may not exist in the usual two‑sided sense, and the condition (f'(c)=0) is not required for an extremum at a boundary Worth keeping that in mind..
Extreme Value Theorem (EVT)
The EVT guarantees that a continuous function on a closed and bounded interval ([a, b]) attains both a global maximum and a global minimum somewhere on that interval. These global extrema may occur at interior points (where Fermat’s theorem applies) or at the endpoints. Since a global maximum is, by definition, at least as large as any other value in the domain, it automatically satisfies the local maximum condition when we restrict attention to a sufficiently small one‑sided neighbourhood.
Thus, continuity on a closed interval provides a straightforward scenario where an endpoint can be a local (indeed, global) maximum.
When an Endpoint Is a Local Maximum: Conditions and Examples
Condition 1: The Function Does Not Increase Beyond the Endpoint
For the left endpoint (a) to be a local maximum, we need: [ \exists \epsilon > 0 \text{ such that } f(x) \le f(a) \quad \forall x \in [a, a+\epsilon). ] Analogously, for the right endpoint (b): [ \exists \epsilon > 0 \text{ such that } f(x) \le f(b) \quad \forall x \in (b-\epsilon, b]. ]
In plain language: the function must not rise immediately after leaving the endpoint And it works..
Example 1: A Simple Quadratic on a Closed Interval
Consider (f(x) = -(x-2)^2 + 5) on the interval ([0, 3]).
Practically speaking, - The vertex of the parabola is at (x=2), where (f(2)=5) (the global maximum). Practically speaking, - At the left endpoint (x=0), (f(0)=-(0-2)^2+5 = 1). And - For any small (\epsilon>0), points just to the right of 0 give values (f(\epsilon) = -( \epsilon-2)^2+5). Since (( \epsilon-2)^2 < 4) for (\epsilon<2), we have (f(\epsilon) > 1). Worth adding: hence, (f) increases as we move right from 0, so (x=0) is not a local maximum. - At the right endpoint (x=3), (f(3)=-(3-2)^2+5 = 4). Plus, for points just left of 3, say (x=3-\epsilon), we have (f(3-\epsilon) = -(1-\epsilon)^2+5 = 4 - (2\epsilon - \epsilon^2)). For sufficiently small (\epsilon), this is less than 4, so the function decreases as we approach 3 from the left. So, (x=3) is a local maximum (in fact, it is also the global maximum on ([0,2]) but not on the whole interval because the interior point 2 gives a higher value).
Example 2: A Constant Function
Let (f(x)=3) on ([1,4]). And every point yields the same value. - At the left endpoint (x=1), for any (\epsilon>0), all points (x\in[1,1+\epsilon)) satisfy (f(x)=3 \le f(1)=3). Consider this: hence, (x=1) meets the definition of a local maximum (and also a local minimum). Even so, - The same holds for the right endpoint (x=4). Thus, endpoints can be local maxima when the function is flat or non‑increasing near the boundary.
Example 3: A Function with a Jump Discontinuity
Consider [ f(x)=\begin{cases} 2, & x=0\ x+1, & 0<x\le 2 \end{cases} ] on the domain ([0,2]).
- At
Continuingwith the piecewise‑defined function above, we examine the behavior at the left endpoint (x=0).
For any (\varepsilon>0) and any (x\in(0,\varepsilon)) we have (f(x)=x+1>1). Since (f(0)=2), it follows that (f(x)>f(0)) for all sufficiently small positive (x). This means (x=0) fails the local‑maximum test; it is not a maximum at all.
Now shift our focus to the right endpoint (x=2). Because the definition of the function on ((0,2]) is simply (f(x)=x+1), the values immediately to the left of 2 are (f(2-\varepsilon)=2-\varepsilon+1=3-\varepsilon), which are strictly less than (f(2)=3). Practically speaking, hence there exists an (\varepsilon>0) such that every point in ((2-\varepsilon,2]) satisfies (f(x)\le f(2)). Simply put, the right endpoint is a local maximum (indeed, it is also the absolute maximum on the whole domain, since no interior point can exceed the value 3) Still holds up..
Easier said than done, but still worth knowing.
These three illustrations — constant functions, functions that are non‑increasing at a boundary, and functions that rise away from one side but fall toward the opposite side — capture the essential ways in which an endpoint can qualify as a local maximum. The key ingredients are:
- One‑sided monotonicity – the function must not increase when moving inward from the endpoint. 2. Existence of a neighbourhood – there must be some (\varepsilon>0) for which the inequality (f(x)\le f(\text{endpoint})) holds throughout that neighbourhood.
- Domain‑specific interpretation – on a closed interval the endpoint’s one‑sided neighbourhood is all that is available, so the condition is automatically satisfied when the function stays flat or declines right away.
In practice, identifying whether an endpoint can serve as a local maximum reduces to checking these simple criteria. Here's the thing — when they hold, the endpoint contributes a candidate for optimization, especially in constrained problems where the feasible set is a closed interval or a more general closed set. Recognizing the possibility of endpoint maxima thus completes the picture of how local extrema arise in elementary calculus and prepares the ground for more sophisticated optimization techniques.
