梯度定理(gradient theorem),係向量微積分裏面嘅一個定理,亦都係微積分基本定理拓展咗之後對於路徑積分嘅一個廣義版本。
根據梯度定理,如果有一個喺 R n {\displaystyle \mathbb {R} ^{n}} 向量空間嘅連續可微分標量函數 ϕ : U ⊆ R n → R {\displaystyle \phi :\mathbf {U} \subseteq \mathbb {R} ^{n}\rightarrow \mathbb {R} } ,對於任何喺 U {\displaystyle \mathbf {U} } 裏面起點係 p → {\displaystyle {\vec {p}}} 同埋終點係 q → {\displaystyle {\vec {q}}} 嘅曲線 γ {\displaystyle \gamma } ,有:
∫ γ ∇ ϕ ( r → ) ⋅ d r → = ϕ ( q → ) − ϕ ( p → ) {\displaystyle \int _{\gamma }\nabla \phi ({\vec {r}})\cdot d{\vec {r}}=\phi ({\vec {q}})-\phi ({\vec {p}})}
其中 ∇ ϕ ( r → ) {\displaystyle \nabla \phi ({\vec {r}})} 就係 ϕ {\displaystyle \phi } 嘅梯度,係一個向量場。
將 R n {\displaystyle \mathbb {R} ^{n}} 裏面嘅所有維度分別叫做 x i {\displaystyle x_{i}} ,所有單位向量分別叫做 e i → {\displaystyle {\vec {e_{i}}}} ,其中 i ∈ N : 1 ≤ i ≤ n {\displaystyle i\in \mathbb {N} :1\leq i\leq n} 。噉樣 ϕ {\displaystyle \phi } 嘅全微分可以寫做:
d ϕ = ∑ i = 1 n ∂ ϕ ∂ x i d x i {\displaystyle d\phi =\sum _{i=1}^{n}{\frac {\partial {\phi }}{\partial x_{i}}}dx_{i}}
喺 R n {\displaystyle \mathbb {R} ^{n}} 裏面梯度嘅定義係:
∇ ϕ = ∑ i = 1 n ∂ ϕ ∂ x i e i → {\displaystyle \nabla \phi =\sum _{i=1}^{n}{\frac {\partial {\phi }}{\partial x_{i}}}{\vec {e_{i}}}}
而位置向量同佢嘅微分嘅定義分別係:
r → = ∑ i = 1 n x i e i → d r → = ∑ i = 1 n d x i e i → {\displaystyle {\begin{aligned}&{\vec {r}}=\sum _{i=1}^{n}x_{i}{\vec {e_{i}}}\\&d{\vec {r}}=\sum _{i=1}^{n}dx_{i}{\vec {e_{i}}}\end{aligned}}}
所以有:
d ϕ = ∇ ϕ ⋅ d r → {\displaystyle d\phi =\nabla \phi \cdot d{\vec {r}}}
於是有:
∫ γ ∇ ϕ ( r → ) ⋅ d r → = ∫ ϕ ( p → ) ϕ ( q → ) d ϕ ( r → ) {\displaystyle {\begin{aligned}&\int _{\gamma }\nabla \phi ({\vec {r}})\cdot d{\vec {r}}\\&=\int _{\phi ({\vec {p}})}^{\phi ({\vec {q}})}d\phi ({\vec {r}})\end{aligned}}} (積分上下界轉換)
= [ ϕ ] ϕ ( p → ) ϕ ( q → ) = ϕ ( q → ) − ϕ ( p → ) {\displaystyle {\begin{aligned}&=[\phi ]_{\phi ({\vec {p}})}^{\phi ({\vec {q}})}\\&=\phi ({\vec {q}})-\phi ({\vec {p}})\end{aligned}}}
證完。