📚 Core Content and Study Guide for Multivariable Calculus | 多元微积分核心内容与学习指南
Multivariable calculus extends the ideas of single-variable calculus to functions of two, three, or more variables. It provides the essential language and tools for describing change, rates, motion, geometry and physical laws in higher-dimensional spaces, and it underpins subjects such as electromagnetism, fluid dynamics, economics and machine learning. This guide walks you through the core topics - from partial derivatives and multiple integrals to the great vector calculus theorems - and offers concrete study strategies, to help you succeed, whether you are preparing for A-Level Further Mathematics, the IB Diploma, AP Calculus BC, or the first year of a university mathematics degree.
多元微积分把单变量微积分的思想推广到两个、三个或更多变量的函数。它为描述高维空间中的变化、变化率、运动、几何与物理定律提供了基本的语言和工具，同时也是电磁学、流体力学、经济学和机器学习等学科的数学基础。本指南将带你系统梳理核心主题——从偏导数、多重积分到向量微积分中的几大定理——并提供具体的学习策略，无论你是在准备 A-Level 进阶数学、IB 课程、AP 微积分 BC，还是大学一年级数学课程，都能从中受益。
1. Functions of Several Variables | 多元函数
ten z = f(x, y), assigns a unique real number to every ordered pair (x, y) in its domain D ⊂ ℝ². The graph of such a function is a surface in ℝ (x, y, z), are harder to draw but follow exactly the same logic: the domain is a region in ℝ³ and the output is a single number.
二元函数记作 z = f(x, y)，它把定义域 D ⊂ ℝ² 中每一个有序数对 (x, y) 映射为唯一的实数，其图像是 ℝ³ 中的一张曲面，值域则是所有实际取到的输出值。三元函数 w = f(x, y, z) 虽然难以作图，但逻辑完全相同：定义域是 ℝ³ 中的一块区域，输出仍然是一个数。
ithms require a strictly positive argument, and denominators must not be zero. For example, f(x, y) = √9 − x² − y² has domain the closed disc x² + y² ≤ 9 and range [0, 3].
在进行任何微分或积分之前，先检查表达式中的约束条件：平方根要求被开方数非负，对数要求真数严格大于零，分母不能为零。例如 f(x, y) = √9 − x² − y² 的定义域是闭圆盘 x² + y² ≤ 9，值域是 [0, 3]。
hout 3D technology, and they are exactly the mathematics behind topographic maps and weather charts. Common surfaces worth memorising include the paraboloid f(x, y) = x² + y², the cone f(x, y) = √x² + y², and the hyperbolic paraboloid (saddle) f(x, y) = x² − y².
等高线（等值线）是满足 f(x, y) = k（k 为固定常数）的点集。等高线图让你在不依赖三维技术的情况下可视化曲面的形状，它正是等高地形图和天气图背后的数学。值得记住的常见曲面包括抛物面 f(x, y) = x² + y²、锥面 f(x, y) = √x² + y² 以及双曲抛物面（马鞍面）f(x, y) = x² − y²。
Limits and continuity in two variables are subtler than in one variable, because a point can be approached along infinitely many paths. If two different paths give two different limits, the limit does not exist; if a function is continuous, then small changes in the input produce small changes in the output.
二元函数的极限与连续性比一元情形更微妙，因为趋近一个点可以有无数条路径。如果两条不同路径给出两个不同的极限值，则该极限不存在；若函数连续，则输入的微小变化只会引起输出的微小变化。
2. Partial Derivatives | 偏导数
by differentiating f(x, y) while treating y as a constant. Geometrically, at a point (a, b, f(a, b)) it gives the slope of the tangent line running parallel to the xz-plane. The formal definition mirrors the one-variable derivative:
f 对 x 的偏导数记作 ∂f/∂x 或 fₓ，求法是把 y 视为常数后对 x 求导。从几何上看，在点 (a, b, f(a, b)) 处，它给出与 xz 平面平行的切线斜率。其形式定义与一元导数完全对应：
∂f/∂x = lim (h → 0) [f(x + h, y) − f(x, y)] / h
e constant multiple rule, product rule, quotient rule and chain rule all still apply - only the variable of interest is differentiated.
对 y 作同样构造便得到 ∂f/∂y，即与 yz 平面平行的切线斜率。对 f(x, y) = x²y³，有 ∂f/∂x = 2xy³，∂f/∂y = 3x²y²。注意常系数法则、乘积法则、商法则和链式法则依然适用，只是仅对感兴趣的变量求导。
, while ∂²f/∂x∂y means differentiate first with respect to y and then with respect to x. Clairaut' ates that if the mixed partial derivatives are continuous, then the order does not matter:
高阶偏导数通过反复求导得到。记号 ∂²f/∂x² 表示对 x 求导两次，而 ∂²f/∂x∂y 表示先对 y 求导、再对 x 求导。克莱罗定理指出：若混合偏导数连续，则求导次序可以交换：
∂²f/∂x∂y = ∂²f/∂y∂x
s equal 6xy². In exams, the most common loss of marks is a careless sign slip when differentiating a product or quotient, so write each step down rather than doing it mentally.
对 f(x, y) = x²y³，两个混合偏导数都等于 6xy²。在考试中，最常见的失分点是在乘积或商求导时出现符号错误，因此务必把每一步写清楚，不要心算。
3. The Gradient and Directional Derivatives | 梯度与方向导数
The gradient of f collects all first partial derivatives into a vector. In two dimensions ∇f = (∂f/∂x, ∂f/∂y); in three dimensions ∇ ∂y, ∂f/∂z). The gradient points in the direction of the steepest ascent, and its magnitude |∇f| is the maximum rate of increase of the function at that point.
梯度把全部一阶偏导数汇成一个向量。二维时 ∇f = (∂f/∂x, ∂f/∂y)，三维时 ∇f = (∂f/∂x, ∂f/∂y, ∂f/∂z)。梯度指向函数上升最快的方向，其模 |∇f| 就是该点处函数的最大增长率。
The directional derivative of f at a point in the direction of a unit vector u measures the rate of change in that direction and is computed with a dot product:
函数在某点沿单位向量 u 方向的方向导数，度量了沿该方向的变化率，用点积计算：
Du f = ∇f · u （要求 |u| = 1）
Because the dot product is largest when the vectors are parallel, the maximum value of Du f is |∇f|, the minimum is −|∇f|, and Du l curve or level surface. This last fact explains why the gradient is always perpendicular to contours - a standard exam question.
由于两向量同向时点积最大，所以方向导数的最大值为 |∇f|，最小值为 −|∇f|，而在与等值线或等值面相切的方向上 Du f = 0。这一点解释了为什么梯度总与等高线垂直，也是标准考点之一。
Example: for f(x, y) = x² + y² at the point (1, 2), ∇f = (2x, 2y) = (2, 4), so the maximum rate of increase is |∇f| = √2² + 4² = √20 = 2√5, in the direction of the unit vector (1/√5, 2/√5).
例：对 f(x, y) = x² + y² 在点 (1, 2) 处，∇f = (2x, 2y) = (2, 4)，因此最大增长率为 |∇f| = √2² + 4² = √20 = 2√5，方向为单位向量 (1/√5, 2/√5)。
4. The Chain Rule and Implicit Differentiation | 链式法则与隐函数求导
ntermediate variable. If z = f(x, y) and both x and y depend on a single parameter t, then
多元链式法则追踪一个输入的变化如何经由每一个中间变量传递。若 z = f(x, y) 且 x、y 都依赖同一个参数 t，则
dz/dt = (∂z/∂x)(dx/dt) + (∂z/∂y)(dy/dt)
diagram - one branch for each intermediate variable - is the safest way to avoid missing a term.
若 x、y 依赖两个参数 s 和 t，则对每个参数各得到一个同形状的方程，从而求出 ∂z/∂s 与 ∂z/∂t。画出树状图——每个中间变量画一条分支——是避免漏项最稳妥的方法。
0. Differentiating with respect to x, and remembering that y is itself a function of x, gives the compact formula
隐函数求导可由对 F(x, y) = 0 使用链式法则得到。对 x 求导，并记住 y 本身是 x 的函数，便得到简洁公式
dy/dx = −(∂F/∂x) / (∂F/∂y)
In three variables, the same reasoning gives ∂z/∂x = −(∂F/∂x)/(∂F/∂z), provided ∂F/∂z ≠ 0.
对圆 x² + y² = 25，取 F = x² + y² − 25，得 dy/dx = −2x / 2y = −x/y。在三个变量的情形下，同样的推理给出 ∂z/∂x = −(∂F/∂x)/(∂F/∂z)，前提是 ∂F/∂z ≠ 0。
he notation is a reminder of the chain rule, not a licence for algebraic cancellation - a classic misconception that costs marks.
考试提示：切勿把 ∂ 符号当作分数直接约分。该记号只是链式法则的提醒，并不是代数约分的许可，这是一个典型的失分误区。
5. Tangent Planes and Linear Approximation | 切平面与线性近似
the tangent plane in two. At the point (a, b, f(a, b)) the tangent plane to z = f(x, y) is
正如一元导数给出切线，梯度在二元情形给出切平面。在点 (a, b, f(a, b)) 处，曲面 z = f(x, y) 的切平面为
z − f(a, b) = fₓ(a, b)(x − a) + fy(a, b)(y − b)
replaces the surface locally by its tangent plane: L(x, y) = f(a, b) + fₓ(a, b)(x − a) + fy ror analysis, since it converts a nonlinear problem into a linear one near a chosen point.
线性化（线性近似）用切平面在局部替代曲面：L(x, y) = f(a, b) + fₓ(a, b)(x − a) + fy(a, b)(y − b)。这是误差分析的核心工具，因为它把非线性问题在选定点附近转化为线性问题。
The total differential expresses the same idea in infinitesimal form: dz = fₓ dx + fy nge slightly, for example in the propagation of measurement errors in physics experiments.
全微分以无穷小形式表达同一思想：dz = fₓ dx + fy dy。它用于估计当输入量发生微小变化时结果的变化，例如物理实验中测量误差的传播。
Example: for f(x, y) = x² + y² at (1, 1, 2), fₓ = 2 and fy = 2
更多咨询请联系16621398022（同微信）
或截图后用「扫一扫 → 相册」选择本图