Quant Academy Math · Probability · Quant Finance
Phase 3 - Lesson 3.1

Vectors, Multivariable Functions, and Level Sets

From points in R^n to scalar fields: how to picture a function of several variables through its graph, level sets, and sections.

⏱ 45 min● Intermediate🔗 Prereqs: 2.x single-variable calculus
↖ Phase 3 hub
Builds on: Single-variable functions and limits from Phase 2.
Leads to: Partial derivatives (3.2) differentiate the scalar fields introduced here.

Learning Objectives

Click a status chip to cycle: Not started → In progress → Studied → Practiced → Needs review → Mastered.

Key Vocabulary

Euclidean space
The set \(\R^n\) of ordered n-tuples with vector addition, scalar multiplication, and the dot product.
Scalar field
A function \(f:\R^n\to\R\) assigning a number to each point, e.g. temperature or a portfolio’s loss.
Vector field
A function \(F:\R^n\to\R^m\) assigning a vector to each point, e.g. a gradient or a force.
Level set
The set \(\{x: f(x)=c\}\); for \(n=2\) a level curve, for \(n=3\) a level surface.
Norm
The length \(\lVert x\rVert=\sqrt{x\cdot x}\); measures distance and induces the metric on \(\R^n\).
Path limit
The value approached by \(f\) as \(x\) tends to a point along a chosen curve; all paths must agree for a limit to exist.

Intuition & Motivation

Intuition
A function of one variable lives on a line and draws a curve. A function of two variables lives on a plane and draws a surface floating above it. You cannot easily see four dimensions, so instead of the surface we read its contour map: slice the surface at constant heights and project the slices down. Those level curves are exactly what a topographic map or an option’s profit diagram shows.

The dot product is the one tool that turns geometry into algebra: it measures length, angle, and projection all at once. Almost every idea in the rest of this phase - gradients, orthogonality of constraints, least squares - is the dot product wearing a different hat.

Vectors and the dot product

A point \(x=(x_1,\dots,x_n)\) in \(\R^n\) is simultaneously a location and a displacement. The two operations that matter are the norm and the dot product:

\[x\cdot y=\sum_{i=1}^n x_i y_i,\qquad \lVert x\rVert=\sqrt{x\cdot x},\qquad x\cdot y=\lVert x\rVert\,\lVert y\rVert\cos\theta.\] (3.1)

The last identity defines the angle \(\theta\) between vectors and gives the Cauchy–Schwarz bound \(|x\cdot y|\le \lVert x\rVert\,\lVert y\rVert\). Two vectors are orthogonal exactly when \(x\cdot y=0\).

Definition - Scalar and vector fields
A scalar field is \(f:D\subseteq\R^n\to\R\). A vector field is \(F:D\to\R^m\). A path or curve is \(r:\R\to\R^n\), a function of one variable with vector output.

Graphs, level sets, and sections

For \(f:\R^2\to\R\) the graph is the surface \(z=f(x,y)\) in \(\R^3\). Two lower-dimensional shadows make it readable:

Worked Example - Level curves of a quadratic loss
1
Let \(f(x,y)=x^2+4y^2\), a stylized portfolio loss in two risk factors. Set \(f=c\) with \(c\gt 0\).
2
The level set \(x^2+4y^2=c\) is an ellipse with semi-axes \(\sqrt{c}\) and \(\tfrac12\sqrt{c}\).
3
As \(c\) grows the ellipses nest outward; their elongation along \(x\) says the loss is less sensitive to \(x\) than to \(y\) (a unit move in \(y\) costs four times as much).
4
The vertical section at \(y=0\) is the parabola \(z=x^2\) - the surface really is a bowl.

Limits and continuity in several variables

We say \(\lim_{x\to a}f(x)=L\) if \(f(x)\) can be forced within any tolerance of \(L\) by taking \(x\) close enough to \(a\) - along every path. Unlike one dimension there are infinitely many approach directions, so a limit fails if two paths disagree.

Key Idea
To disprove a limit, exhibit two paths giving different values. To prove one, bound \(|f(x)-L|\) by a quantity depending only on \(\lVert x-a\rVert\) (e.g. via polar coordinates).
Common Mistakes to Avoid
  • Concluding a two-variable limit exists after checking only straight-line paths - a parabolic path can still break it (classic: \(f=xy^2/(x^2+y^4)\)).
  • Confusing the graph (a surface in \(\R^3\)) with a level set (a curve in \(\R^2\)).
  • Writing \(\lVert x\rVert^2=x^2+y^2+\dots\) but then forgetting the square root when a distance is needed.
  • Assuming orthogonal means ‘perpendicular on the page’ only; algebraically it is exactly \(x\cdot y=0\) in any dimension.
Quant Practitioner Tips
  • Read any risk report as a contour map: nested ellipses are a quadratic loss, and their axes are the principal risk directions (Phase 4 makes this precise).
  • Polar substitution \(x=r\cos\theta,\ y=r\sin\theta\) turns many 2-D limits into a single-variable limit in \(r\).
  • The dot product is your Swiss-army knife: length, angle, projection, and later the gradient’s directional derivative all reduce to it.

Practice this in the Euler Lab

Computational problems that exercise exactly this technique. Each opens in the Euler Lab with a Python workbench, a progressive hint ladder, and answer checking. Tier A/B run at full scale in the browser.

Warm-up:
#6 Sum Square Difference (1%, tier A) #9 Special Pythagorean Triplet (1%, tier A) #28 Number Spiral Diagonals (2%, tier A) #39 Integer Right Triangles (2%, tier A)

Applied:
#504 Square on the Inside (16%, tier B) #265 Binary Circles (17%, tier B) #679 Freefarea (17%, tier B)

Challenge:
#147 Rectangles in Cross-hatched Grids (41%, tier C) #184 Triangles Containing the Origin (41%, tier C)

273 Project Euler problems in total are mapped to this lesson. Open the Euler Lab to filter them all.

Knowledge Check

Q1 Easy
The level set \(\{(x,y): x^2+4y^2=9\}\) is:
A circle of radius 3
An ellipse with semi-axes 3 and 3/2
A parabola
Two straight lines
Q2 Easy
Two nonzero vectors satisfy \(x\cdot y=0\). This means:
They point the same way
They are orthogonal (angle 90°)
One is the zero vector
Their norms are equal
Q3 Medium
To show \(\lim_{(x,y)\to(0,0)} f(x,y)\) does NOT exist, the cleanest strategy is:
Check the limit along the x-axis only
Find two approach paths giving different limits
Show f is unbounded
Evaluate f at the origin

Practical Exercise

Let \(f(x,y)=\dfrac{xy}{x^2+y^2}\) for \((x,y)\neq(0,0)\). (a) Compute the limit as \((x,y)\to(0,0)\) along the line \(y=mx\). (b) Does \(\lim_{(x,y)\to 0}f\) exist? Justify.

▶ Show full solution

(a) Substitute \(y=mx\):

\[f(x,mx)=\frac{x\cdot mx}{x^2+m^2x^2}=\frac{m x^2}{(1+m^2)x^2}=\frac{m}{1+m^2},\]

which is independent of \(x\), so the limit along \(y=mx\) is \(m/(1+m^2)\).

(b) The path limit depends on the slope \(m\): it is \(0\) along \(y=0\) (m=0) but \(\tfrac12\) along \(y=x\) (m=1). Two paths disagree, so the two-variable limit does not exist, even though every straight-line restriction is finite.

After the reveal, answer for yourself: Notice each single-line limit exists yet the joint limit fails - multivariable limits are strictly stronger.

Lesson Summary

A function of several variables is a scalar field on \(\R^n\); we read it through its graph, level sets, and sections. The dot product encodes length, angle, and orthogonality, and multivariable limits must hold along every path - a single disagreeing path disproves existence.

Formula Sheet Additions

Dot product & angle
\[x\cdot y=\lVert x\rVert\,\lVert y\rVert\cos\theta\]
Length, angle, and orthogonality in one identity.

Retrieval Practice

Close the lesson and answer from memory before checking. This is deliberate, effortful recall - the single highest-yield study action.

▶ Show retrieval prompts & answers
Q: What is a level set and what shape does \(x^2+4y^2=c\) give?
A: The set of inputs mapping to a fixed value c; here a family of ellipses with semi-axes sqrt(c) and (1/2)sqrt(c).
Q: How do you disprove a multivariable limit?
A: Exhibit two approach paths (e.g. y=0 and y=x) along which f tends to different values.

Completion Checklist

Confidence / mastery rating
Personal notes

Source References

This lesson synthesizes and paraphrases concepts from the sources below. No copyrighted text, problem sets, or solutions are reproduced. Return to the originals for full depth.

← 2.7Next: 3.2 →