How to Find the Limit of Multivariable Functions
Understanding how to find the limit of multivariable functions is a cornerstone of advanced calculus and essential for fields such as physics, engineering, and economics. Unlike single‑variable limits, where you only need to consider left‑ and right‑hand approaches, a multivariable limit requires the function to approach the same value regardless of the path taken toward the point of interest. This article walks you through the concepts, strategies, and step‑by‑step methods you can use to evaluate these limits confidently.
1. What Is a Multivariable Limit?
For a function (f(x,y)) (or with more variables), the limit as ((x,y)\to(a,b)) is defined as:
[ \lim_{(x,y)\to(a,b)} f(x,y)=L ]
if for every (\varepsilon>0) there exists a (\delta>0) such that whenever
[ 0<\sqrt{(x-a)^2+(y-b)^2}<\delta\quad\text{we have}\quad|f(x,y)-L|<\varepsilon . ]
In plain language: no matter how you get arbitrarily close to ((a,b)), the function values must get arbitrarily close to (L). If two different paths give different results, the limit does not exist (DNE) But it adds up..
2. Preliminary Checks Before Diving In
Before applying heavy machinery, run these quick tests:
| Check | What to Do | Outcome |
|---|---|---|
| Direct substitution | Plug ((a,b)) into (f). | If the function is defined and yields a finite number, that is often the limit (provided the function is continuous at that point). Also, |
| Obvious indeterminate form | You get (\frac{0}{0}), (\frac{\infty}{\infty}), (0\cdot\infty), etc. ) and the denominator is non‑zero at ((a,b)). And | Then the limit equals the function value. |
| Continuity | Recognize if (f) is a sum, product, quotient, or composition of continuous functions (polynomials, exponentials, sine/cosine, etc. | Proceed to path analysis or algebraic manipulation. |
Honestly, this part trips people up more than it should Small thing, real impact..
3. Core Strategies for Evaluating the Limit
3.1 Path (or Two‑Path) Test
The simplest way to show a limit does NOT exist is to find two different paths that give different limiting values.
Steps
- Choose simple paths, e.g., (y=mx) (straight lines), (y=x^2) (parabola), or (x=0), (y=0).
- Substitute the path into (f(x,y)) to obtain a single‑variable function.
- Compute the limit as the variable approaches the relevant value.
- If two paths yield different results, the limit DNE.
Example:
[
f(x,y)=\frac{xy}{x^2+y^2},\quad (x,y)\to(0,0).
]
- Along (y=0): (f(x,0)=0\Rightarrow) limit (0).
- Along (y=x): (f(x,x)=\frac{x^2}{2x^2}= \frac12\Rightarrow) limit (\frac12).
Since the results differ, the limit does not exist.
3.2 Polar Coordinates
When the point of interest is the origin (or can be shifted to the origin), converting to polar coordinates often simplifies the expression because the distance to the point is captured by a single variable (r).
Transformation:
[
x=r\cos\theta,\qquad y=r\sin\theta,\qquad r=\sqrt{x^2+y^2}.
]
Then examine (\displaystyle\lim_{r\to0^+} f(r\cos\theta,r\sin\theta)). If the resulting expression does not depend on (\theta) and tends to a finite value (L), the limit exists and equals (L). If the limit still varies with (\theta), the limit DNE Still holds up..
Example:
[
f(x,y)=\frac{x^2y}{x^4+y^2}.
]
In polar:
[
f(r\cos\theta,r\sin\theta)=\frac{r^3\cos^2\theta\sin\theta}{r^4\cos^4\theta+r^2\sin^2\theta}
= \frac{r\cos^2\theta\sin\theta}{r^2\cos^4\theta+\sin^2\theta}.
]
As (r\to0), numerator (\to0) while denominator (\to\sin^2\theta) (unless (\sin\theta=0)). The limit becomes (0) for all (\theta) where (\sin\theta\neq0). For (\theta=0,\pi) (i.e., (y=0)), the expression is identically (0). Hence the limit is (0) regardless of (\theta); the limit exists and equals (0) Easy to understand, harder to ignore..
3.3 Squeeze (Sandwich) Theorem
If you can bound (|f(x,y)-L|) between two functions that both go to zero as ((x,y)\to(a,b)), then the limit is (L).
Typical bounds: use inequalities like (|xy|\le \frac12(x^2+y^2)) or (|\sin t|\le|t|).
Example:
[
f(x,y)=\frac{x^2y}{x^2+y^2}.
]
Notice (|x^2y|\le x^2\sqrt{x^2+y^2}) because (|y|\le\sqrt{x^2+y^2}). Then
[
|f(x,y)|=\frac{|x^2y|}{x^2+y^2}\le\frac{x^2\sqrt{x^2+y^2}}{x^2+y^2}
= \frac{x^2}{\sqrt{x^2+y^2}}.
]
As ((x,y)\to(0,0)), the right‑hand side goes to (0). By the squeeze theorem, (\displaystyle\lim_{(x,y)\to(0,0)} f(x,y)=0) It's one of those things that adds up..
3.4 Algebraic Manipulation & Factorization
Sometimes factoring cancels the problematic part, turning an indeterminate form into a determinate one It's one of those things that adds up..
Example:
[
f(x,y)=\frac{x^2-y^2}{x-y},\quad (x,y)\to(1,1).
]
Factor numerator: ((x-y)(x+y)). Cancel (x-y) (valid except at the point itself):
[
f(x,y)=x+y\quad\text{for }x\neq y.
Day to day, ]
Now direct substitution gives (1+1=2). Hence the limit is (2) Easy to understand, harder to ignore. But it adds up..
3.5 Epsilon‑Delta Proofs (Theoretical)
For rigorous justification, especially in proofs, you may need to construct an (\varepsilon)-(\delta) argument. While rarely required in routine homework, understanding the logic helps you appreciate why the path and squeeze methods work Not complicated — just consistent..
Outline:
- Assume a candidate limit (L).
- Given (\varepsilon>0), find (\delta>0)
To carry out an ε‑δ verification for the polar‑coordinate example we first rewrite the difference between the function and its proposed limit. In the case (L=0) we have
[ \bigl|f(r\cos\theta,r\sin\theta)-0\bigr| =\left|\frac{r^{3}\cos ^2!\theta,\sin\theta} {r^{4}\cos ^4!Practically speaking, \theta+r^{2}\sin ^2! \theta}\right| =\frac{r,|\cos ^2!In real terms, \theta,\sin\theta|} {\bigl|r^{2}\cos ^4! \theta+r^{2}\sin ^2!\theta\bigr|} =\frac{|\cos ^2!\theta,\sin\theta|} {r\bigl|\cos ^4!Practically speaking, \theta+r^{-2}\sin ^2! \theta\bigr|}.
Since (\cos ^2!In real terms, \theta+r^{-2}\sin ^2! \theta|\ge r^{-2}|\sin ^2!\theta\le 1) and (|\sin\theta|\le 1), the absolute value of the numerator stays bounded by 1. A convenient way to obtain a uniform bound for every (\theta) is to note that (|\cos ^4!\theta|).
[ |f(r\cos\theta,r\sin\theta)| \le \frac{1}{r\cdot r^{-2}} |\sin\theta| = r,|\sin\theta|. ]
Because (|\sin\theta|\le 1), we obtain the simple estimate
[ |f(r\cos\theta,r\sin\theta)|\le r . ]
Hence, whenever the radial distance satisfies (r<\varepsilon), the quantity above is automatically smaller than (\varepsilon). Choosing (\delta=\varepsilon) therefore guarantees
[ 0<\sqrt{x^{2}+y^{2}}<\delta\quad\Longrightarrow\quad \bigl|f(x,y)-0\bigr|<\varepsilon, ]
which proves that (\displaystyle\lim_{(x,y)\to(0,0)}f(x,y)=0) rigorously The details matter here..
Beyond the three core techniques already presented—polar conversion, the
Beyond the three core techniques already presented—polar conversion, the squeeze theorem, and algebraic manipulation/factorization—there are several additional strategies that often prove useful when dealing with multivariable limits And it works..
Path‑wise examination for non‑existence.
If two different approaches to the point yield different limiting values, the limit does not exist. Commonly tested paths include straight lines (y=mx), parabolas (y=x^{2}), or more general curves (y=x^{n}). Because the limit must be independent of the path, finding a single counter‑example suffices to disprove existence. To give you an idea, for
[
g(x,y)=\frac{xy}{x^{2}+y^{2}},
]
approaching along (y=x) gives (\tfrac12), while along (y=0) gives (0); hence the limit at ((0,0)) fails to exist.
Iterated limits and continuity of elementary functions.
When a function can be expressed as a composition of continuous single‑variable functions (polynomials, exponentials, trigonometric functions, rational functions with non‑zero denominator), the limit can often be obtained by successive substitution. If the inner limit exists and the outer function is continuous at that value, the overall limit follows directly. This technique is especially handy after a simplification step that removes the indeterminate form Most people skip this — try not to. Surprisingly effective..
Norm‑based estimates and homogeneity.
Many multivariable expressions are homogeneous of some degree; exploiting this property allows one to factor out a power of the radius (r=\sqrt{x^{2}+y^{2}}). For a homogeneous function (H(x,y)) of degree (k), we have (H(r\cos\theta,r\sin\theta)=r^{k}H(\cos\theta,\sin\theta)). If (k>0) and the angular factor remains bounded, the whole expression tends to zero as (r\to0). Conversely, if (k<0) the limit may blow up unless the angular factor vanishes sufficiently fast, prompting a more refined angular analysis Took long enough..
Use of inequalities (AM‑GM, Cauchy‑Schwarz, etc.).
Bounding the numerator or denominator by simpler expressions often leads to a squeeze‑type argument. As an example, the inequality (|xy|\le \tfrac12(x^{2}+y^{2})) quickly shows that
[
\left|\frac{xy}{x^{2}+y^{2}}\right|\le \tfrac12,
]
and combined with a factor that goes to zero (such as an extra (r) term) yields a limit of zero That alone is useful..
Sequence method.
A limit exists iff every sequence ((x_n,y_n)\to(0,0)) gives the same limit of (f(x_n,y_n)). Constructing specific sequences (e.g., ((t, t^{2})) or ((t, \sin t))) can reveal path‑dependent behavior that might be missed by straight‑line tests.
In practice, one begins with the simplest tools: direct substitution, factoring, and polar conversion. If these fail to give a clear answer, move to path tests to check for non‑existence, or apply norm‑based estimates and inequalities to invoke the squeeze theorem. When a rigorous justification is required, construct an (\varepsilon)-(\delta) argument based on the bounds obtained from the previous steps Practical, not theoretical..
A natural next step after establishing the basic toolbox is to examine iterated limits in more depth. Suppose we have already shown that the two one‑dimensional limits along the coordinate axes differ, as in the opening example. Even if the full two‑variable limit does not exist, the iterated limits
[ \lim_{x\to0}\Bigl(;\lim_{y\to0}\frac{xy}{x^{2}+y^{2}}\Bigr),\qquad \lim_{y\to0}\Bigl(;\lim_{x\to0}\frac{xy}{x^{2}+y^{2}}\Bigr) ]
can be investigated. In many cases the inner limit exists and the outer limit is then evaluated by ordinary continuity. For the function above, fixing (x) and letting (y\to0) gives
[ \lim_{y\to0}\frac{xy}{x^{2}+y^{2}}=\frac{0}{x^{2}}=0, ]
so the first iterated limit equals (\displaystyle\lim_{x\to0}0=0). Reversing the order, however, yields
[ \lim_{x\to0}\frac{xy}{x^{2}+y^{2}}=\frac{0}{y^{2}}=0, ]
and consequently (\displaystyle\lim_{y\to0}0=0) as well. Both iterated limits coincide, yet the joint limit fails to exist—a reminder that iterated limits are not a substitute for the full limit.
Continuity of elementary building blocks
When a function is expressed as a composition of continuous one‑variable maps, the limit can be read off by successive substitution. Here's a good example: consider
[ f(x,y)=\exp!\bigl(, \ln(x^{2}+y^{2}),\bigr)=\frac{1}{x^{2}+y^{2}}. ]
The inner expression (x^{2}+y^{2}) tends to (0) as ((x,y)\to(0,0)). Because the exponential function is continuous on ((0,\infty)), we may write
[ \lim_{(x,y)\to(0,0)}f(x,y)=\exp!\bigl(,\lim_{(x,y)\to(0,0)}\ln(x^{2}+y^{2}),\bigr). ]
The logarithm diverges to (-\infty), and (\exp(-\infty)=0); thus the limit is (0). In more delicate cases, such as
[ g(x,y)=\frac{\sin(x^{2}y)}{x^{2}+y^{2}}, ]
the continuity of (\sin t) at (t=0) lets us replace the numerator by its argument once we know that (x^{2}y\to0) along any path that makes the denominator vanish at the same rate. The key idea is to identify a continuous outer function and verify that its argument approaches a finite value.
Homogeneity revisited
Homogeneous functions often admit a clean polar description. If (H(x,y)) is homogeneous of degree (k), then in polar coordinates
[ H(r\cos\theta,r\sin\theta)=r^{k},H(\cos\theta,\sin\theta). ]
When (k>0) the factor (r^{k}) forces the whole expression to zero as (r\to0), regardless of the angular term, provided the angular factor stays bounded (which it always does for continuous (H)). This means any homogeneous function of positive degree automatically has limit (0) at the origin. Conversely, a homogeneous function of negative degree behaves like (1/r^{|k|}) and can blow up unless the angular part cancels the singularity.
[ \frac{H(x,y)}{r^{m}} = H(\cos\theta,\sin\theta), ]
and inspects whether the angular factor tends to zero fast enough to offset the (r^{-m}) growth.
Inequality‑driven squeeze arguments
Beyond the simple bound (|xy|\le\frac12(x^{2}+y^{2})), a whole family of classical inequalities can be marshaled to trap a problematic expression between two simpler ones that share the same limit. Take this: the Cauchy–Schwarz inequality yields
[ |x^{2}+y^{2}|,|u^{2}+v^{2}| \ge (xu+yv)^{2}, ]
which can be rearranged to give
[ \left|\frac{xu+yv}{x^{2}+y^{2}}\right| \le \frac{\sqrt{u^{2}+v^{2}}}{\sqrt{x^{2}+y^{2}}}. ]
If we can show that the right‑hand side tends to zero—say, by proving (\sqrt{u^{2}+v^{2}} = O(r)) while (\sqrt{x^{2}+y^{2}} = r)—the squeeze theorem forces the original quotient to zero. Such estimates are especially handy when the numerator is a product or a sum of powers, because they make it possible to replace a complicated expression by a monomial in (r) Simple, but easy to overlook..
Quick note before moving on.
Sequence‑based verification
Even after exhaustive path testing, a rigorous proof often requires a sequence argument. Suppose we have already bounded the function by a multiple of (r); then for any sequence ((x_n,y_n)\to(0,0)) we have
[ |f(x_n,y_n)| \le C,r_n \quad\text{with } r_n=\sqrt{x_n^{2}+y_n^{2}}\to0. ]
Hence (\displaystyle\lim_{n\to\infty}f(x_n,y_n)=0). If, on the other hand, we can produce two sequences approaching the origin that yield different limits—e.g.Think about it: , ((t, t^{2})) giving (0) while ((t, t)) gives (\frac12)—the limit does not exist. The sequence method therefore complements the geometric intuition supplied by path analysis.
Putting the pieces together
A systematic workflow emerges:
- Simplify the expression algebraically (factor, cancel, rewrite).
- Attempt direct substitution; if the result is finite, the limit exists.
- Convert to polar coordinates to expose radial dependence; use homogeneity or factor out (r^{k}) when possible.
- Apply inequalities (AM‑GM, Cauchy‑Schwarz, basic bounds) to obtain a squeeze‑type estimate.
- Test special paths (lines, parabolas, sinusoidal curves, sequences) to detect path‑dependence.
- If a candidate limit (L) is identified, construct an (\varepsilon)–(\delta) proof using the bounds obtained in steps 4–5, thereby confirming that every approach yields (L).
When these steps are followed, the evaluation of multivariable limits becomes a matter of matching the structure of the function with the appropriate technique. The combination of algebraic manipulation, continuity arguments, homogeneity, inequality bounds, and sequential verification equips the analyst with a versatile toolkit capable of handling virtually any elementary expression But it adds up..
Conclusion
The short version: the existence of a multivariable limit at a point can be decided by first looking for continuity after simplification, then exploiting homogeneity or polar form to reduce the problem to a radial factor, and finally buttressing the analysis with suitable inequalities or sequence constructions. By moving fluidly among these methods—never relying on a single test in isolation—one can rigorously determine whether a limit exists and, when it does, identify its precise value.
