이 문서는 유클리드 공간 위의 내적에 관한 것입니다. 벡터 공간 위의 내적에 대해서는
내적 공간 문서를, 벡터와 스칼라의 곱셈에 대해서는
스칼라 곱셈 문서를 참고하십시오.
선형대수학 에서 스칼라곱 (scalar곱, 영어 : scalar product ) 또는 점곱 (영어 : dot product )은 유클리드 공간 의 두 벡터로부터 실수 스칼라 를 얻는 연산이다. 스칼라곱이 유클리드 공간의 내적 을 이루므로, 이를 단순히 '내적'이라고 부르기도 한다. 스칼라곱의 개념의 물리학 배경은 주어진 힘 이 주어진 변위 의 물체에 가한 일 을 구하는 문제이다.
정의
차원 이 n {\displaystyle n} 인 유클리드 공간 R n {\displaystyle \mathbb {R} ^{n}} 의 두 벡터 a , b ∈ R n {\displaystyle \mathbf {a} ,\mathbf {b} \in \mathbb {R} ^{n}} 의 스칼라곱 a ⋅ b ∈ R {\displaystyle \mathbf {a} \cdot \mathbf {b} \in \mathbb {R} } 은 두 가지로 정의할 수 있으며, 이 두 정의는 서로 동치이다. 스칼라곱의 기호에는 가운뎃점 '⋅'을 사용하며, 수의 곱셈 기호와는 다르게 생략할 수 없다.
대수적 정의 두 벡터의 좌표가 각각 a = ( a 1 , a 2 , … , a n ) {\displaystyle \mathbf {a} =(a_{1},a_{2},\dots ,a_{n})} 와 b = ( b 1 , b 2 , … , b n ) {\displaystyle \mathbf {b} =(b_{1},b_{2},\dots ,b_{n})} 라면, 이 둘의 스칼라곱은 같은 위치의 성분을 곱한 뒤 모두 합하여 얻는 값이다.
a ⋅ b = a 1 b 1 + a 2 b 2 + ⋯ + a n b n {\displaystyle \mathbf {a} \cdot \mathbf {b} =a_{1}b_{1}+a_{2}b_{2}+\cdots +a_{n}b_{n}} 예를 들어, 두 3차원 벡터 ( 1 , 3 , − 2 ) , ( 4 , 2 , 1 ) ∈ R 3 {\displaystyle (1,3,-2),(4,2,1)\in \mathbb {R} ^{3}} 의 스칼라곱은 다음과 같다.
( 1 , 3 , − 2 ) ⋅ ( 4 , 2 , 1 ) = 1 × 4 + 3 × 2 + ( − 2 ) × 1 = 8 {\displaystyle (1,3,-2)\cdot (4,2,1)=1\times 4+3\times 2+(-2)\times 1=8} 이 경우 스칼라곱의 정의는 벡터의 좌표에 의존하여 정의하지만, R n {\displaystyle \mathbb {R} ^{n}} 에 기존의 좌표계가 아닌 새로운 좌표계를 주더라도, 이 좌표계가 정규 직교 좌표계 라면, 스칼라곱을 나타내는 공식은 바뀌지 않는다. 즉, 임의의 정규 직교 좌표계 아래 스칼라곱은 위치가 같은 두 좌표의 곱을 합한 것과 같다.
유클리드 공간의 벡터는 종종 열벡터 로 간주되며, 이 경우 두 벡터 a , b {\displaystyle \mathbf {a} ,\mathbf {b} } 의 스칼라곱은 다음과 같이 나타낼 수 있다.
a ⋅ b = a T b {\displaystyle \mathbf {a} \cdot \mathbf {b} =\mathbf {a} ^{\operatorname {T} }\mathbf {b} } 여기서 우변의 a T {\displaystyle \mathbf {a} ^{\operatorname {T} }} 는 a {\displaystyle \mathbf {a} } 의 전치 행렬 이며, 곱셈 기호가 생략된 곱셈은 행렬 곱셈 이다.
이 경우 앞선 예시에서의 내적은 다음과 같이 나타낼 수 있다.
( 1 3 − 2 ) ( 4 2 1 ) = 1 × 4 + 3 × 2 + ( − 2 ) × 1 = 8 {\displaystyle {\begin{pmatrix}1&3&-2\end{pmatrix}}{\begin{pmatrix}4\\2\\1\end{pmatrix}}=1\times 4+3\times 2+(-2)\times 1=8} 기하학적 정의 스칼라곱은 기하학적 성질인 '길이 '와 '각도 '를 통해 다음과 같이 정의할 수 있다.
a ⋅ b = { ‖ a ‖ ‖ b ‖ cos ∡ ( a , b ) a ≠ 0 ∧ b ≠ 0 0 a = 0 ∨ b = 0 {\displaystyle \mathbf {a} \cdot \mathbf {b} ={\begin{cases}\Vert \mathbf {a} \Vert \Vert \mathbf {b} \Vert \cos \measuredangle (\mathbf {a} ,\mathbf {b} )&\mathbf {a} \neq \mathbf {0} \land \mathbf {b} \neq \mathbf {0} \\0&\mathbf {a} =\mathbf {0} \lor \mathbf {b} =\mathbf {0} \end{cases}}} 여기서
‖ a ‖ {\displaystyle \Vert \mathbf {a} \Vert } 는 | a | {\displaystyle |\mathbf {a} |} 로 표기하기도 하며, 벡터 a {\displaystyle \mathbf {a} } 의 노름 을 뜻한다. 이는 a {\displaystyle \mathbf {a} } 의 길이 또는 크기를 나타낸다. ‖ b ‖ {\displaystyle \Vert \mathbf {b} \Vert } 역시 마찬가지이다. ∡ ( a , b ) {\displaystyle \measuredangle (\mathbf {a} ,\mathbf {b} )} 는 두 벡터 a , b {\displaystyle \mathbf {a} ,\mathbf {b} } 사이의 각도이다. 이는 두 벡터가 모두 0이 아닐 때에만 정의되며, 보통 [ 0 , π ] {\displaystyle [0,\pi ]} 에서 값을 취한다. cos ∡ ( a , b ) {\displaystyle \cos \measuredangle (\mathbf {a} ,\mathbf {b} )} 는 코사인 이며, 직각 삼각형 의 이웃변과 빗변의 길이의 비로 정의하거나, 테일러 급수 전개식을 통해 정의할 수 있다. 예를 들어, 만약 두 벡터의 길이가 모두 2이며, 둘 사이의 각도의 코사인 값이 1/2이라면, 이 두 벡터의 스칼라곱은 2 × 2 × 1/2 = 2이다.
이 정의에서 스칼라곱은 두 벡터의 길이와 위치 관계에만 의존하므로, 스칼라곱이 좌표계와 무관함이 더욱 뚜렷하다. 반대로 두 벡터를 똑같은 등거리 변환 에 의하여 변환시켰을 때, 두 벡터의 스칼라곱은 변하지 않는다는 점 역시 정의로부터 자명하다.
몇 가지 특수한 각도의 경우는 다음과 같다.
만약 ∡ ( a , b ) = 0 {\displaystyle \measuredangle (\mathbf {a} ,\mathbf {b} )=0} 이라면, (즉, 두 벡터의 방향이 같다면,) cos ∡ ( a , b ) = 1 {\displaystyle \cos \measuredangle (\mathbf {a} ,\mathbf {b} )=1} 이므로, 내적은 단순히 두 벡터의 길이의 곱이다. a ⋅ b = ‖ a ‖ ‖ b ‖ {\displaystyle \mathbf {a} \cdot \mathbf {b} =\Vert \mathbf {a} \Vert \Vert \mathbf {b} \Vert } 만약 ∡ ( a , b ) = 90 ∘ = π / 2 {\displaystyle \measuredangle (\mathbf {a} ,\mathbf {b} )=90^{\circ }=\pi /2} 라면, (즉, 두 벡터가 서로 수직 이라면,) cos ∡ ( a , b ) = 0 {\displaystyle \cos \measuredangle (\mathbf {a} ,\mathbf {b} )=0} 이므로, 내적은 0이다. a ⋅ b = 0 {\displaystyle \mathbf {a} \cdot \mathbf {b} =0} 만약 ∡ ( a , b ) = 180 ∘ = π {\displaystyle \measuredangle (\mathbf {a} ,\mathbf {b} )=180^{\circ }=\pi } 라면, (즉 두 벡터의 방향이 서로 반대라면,) cos ∡ ( a , b ) = − 1 {\displaystyle \cos \measuredangle (\mathbf {a} ,\mathbf {b} )=-1} 이므로, a {\displaystyle \mathbf {a} } 와 b {\displaystyle \mathbf {b} } 의 내적은 다음과 같다. a ⋅ b = − ‖ a ‖ ‖ b ‖ {\displaystyle \mathbf {a} \cdot \mathbf {b} =-\Vert \mathbf {a} \Vert \Vert \mathbf {b} \Vert } 또한, 이 정의로부터 두 벡터 사이의 각도를 구하는 다음과 같은 공식을 얻을 수 있다.
cos ∡ ( a , b ) = a ⋅ b ‖ a ‖ ‖ b ‖ ( a ≠ 0 , b ≠ 0 ) {\displaystyle \cos \measuredangle (\mathbf {a} ,\mathbf {b} )={\frac {\mathbf {a} \cdot \mathbf {b} }{\Vert \mathbf {a} \Vert \Vert \mathbf {b} \Vert }}\qquad (\mathbf {a} \neq \mathbf {0} ,\;\mathbf {b} \neq \mathbf {0} )} 성질
임의의 벡터 a , b , c ∈ R n {\displaystyle \mathbf {a} ,\mathbf {b} ,\mathbf {c} \in \mathbb {R} ^{n}} 및 스칼라 k ∈ R {\displaystyle k\in \mathbb {R} } 에 대하여, 다음 성질들이 성립한다.
교환 법칙 a ⋅ b = b ⋅ a {\displaystyle \mathbf {a} \cdot \mathbf {b} =\mathbf {b} \cdot \mathbf {a} } 왼쪽 분배 법칙 ( a + b ) ⋅ c = a ⋅ c + b ⋅ c {\displaystyle (\mathbf {a} +\mathbf {b} )\cdot \mathbf {c} =\mathbf {a} \cdot \mathbf {c} +\mathbf {b} \cdot \mathbf {c} } 오른쪽 분배 법칙 a ⋅ ( b + c ) = a ⋅ b + a ⋅ c {\displaystyle \mathbf {a} \cdot (\mathbf {b} +\mathbf {c} )=\mathbf {a} \cdot \mathbf {b} +\mathbf {a} \cdot \mathbf {c} } 스칼라 곱셈 의 보존 ( k a ) ⋅ b = a ⋅ ( k b ) = k a ⋅ b {\displaystyle (k\mathbf {a} )\cdot \mathbf {b} =\mathbf {a} \cdot (k\mathbf {b} )=k\mathbf {a} \cdot \mathbf {b} } 위 네 가지 성질에 따라, 스칼라곱은 대칭 쌍선형 형식 이다. 자기 자신과의 스칼라곱은 음이 아닌 실수이다. a ⋅ a = ‖ a ‖ 2 ≥ 0 {\displaystyle \mathbf {a} \cdot \mathbf {a} =\Vert \mathbf {a} \Vert ^{2}\geq 0} 영벡터와의 스칼라곱은 0이다. 0 ⋅ a = a ⋅ 0 = 0 {\displaystyle \mathbf {0} \cdot \mathbf {a} =\mathbf {a} \cdot \mathbf {0} =0} 자기 자신과의 스칼라곱이 0인 벡터는 영벡터뿐이다. a ⋅ a = 0 ⟺ a = 0 {\displaystyle \mathbf {a} \cdot \mathbf {a} =0\iff \mathbf {a} =\mathbf {0} } 위 세 가지 성질에 따라, 스칼라곱은 양의 정부호 형식 이다. a ⋅ b = 0 {\displaystyle \mathbf {a} \cdot \mathbf {b} =0} 일 필요충분조건은 a ⊥ b {\displaystyle \mathbf {a} \perp \mathbf {b} } 이다. a ⋅ b > 0 {\displaystyle \mathbf {a} \cdot \mathbf {b} >0} 일 필요충분조건은 ∡ ( a , b ) < 90 ∘ {\displaystyle \measuredangle (\mathbf {a} ,\mathbf {b} )<90^{\circ }} 이다. a ⋅ b < 0 {\displaystyle \mathbf {a} \cdot \mathbf {b} <0} 일 필요충분조건은 ∡ ( a , b ) > 90 ∘ {\displaystyle \measuredangle (\mathbf {a} ,\mathbf {b} )>90^{\circ }} 이다. 반면 스칼라곱이 만족시키지 않는 성질에는 다음이 있다.
결합 법칙 은 (1차원 유클리드 공간 R 1 = R {\displaystyle \mathbb {R} ^{1}=\mathbb {R} } 을 제외하면) 성립하지 않는다. 이는 두 벡터의 스칼라곱이 벡터가 아닌 스칼라이므로, ( a ⋅ b ) ⋅ c {\displaystyle (\mathbf {a} \cdot \mathbf {b} )\cdot \mathbf {c} } 나 a ⋅ ( b ⋅ c ) {\displaystyle \mathbf {a} \cdot (\mathbf {b} \cdot \mathbf {c} )} 가 무의미한 수식이기 때문이다. 소거 법칙은 (1차원 유클리드 공간 R 1 = R {\displaystyle \mathbb {R} ^{1}=\mathbb {R} } 을 제외하면) 성립하지 않는다. 예를 들어, R 2 {\displaystyle \mathbb {R} ^{2}} 에서, a = ( 1 , 0 ) {\displaystyle \mathbf {a} =(1,0)} , b = ( 1 , 1 ) {\displaystyle \mathbf {b} =(1,1)} , c = ( 1 , − 1 ) {\displaystyle \mathbf {c} =(1,-1)} 이라면, a ⋅ b = a ⋅ c = 1 {\displaystyle \mathbf {a} \cdot \mathbf {b} =\mathbf {a} \cdot \mathbf {c} =1} , a ≠ 0 {\displaystyle \mathbf {a} \neq \mathbf {0} } 이지만, b ≠ c {\displaystyle \mathbf {b} \neq \mathbf {c} } 이다. 사실, a ⋅ b = a ⋅ c {\displaystyle \mathbf {a} \cdot \mathbf {b} =\mathbf {a} \cdot \mathbf {c} } 일 필요충분조건은 a ⊥ ( b − c ) {\displaystyle \mathbf {a} \perp (\mathbf {b} -\mathbf {c} )} 이다. 응용
스칼라 사영 벡터 a {\displaystyle \mathbf {a} } 의 벡터 b {\displaystyle \mathbf {b} } 위의 스칼라 사영 (영어 : scalar projection ) a b {\displaystyle \mathbf {a} _{\mathbf {b} }} 은 a {\displaystyle \mathbf {a} } 를 b {\displaystyle \mathbf {b} } 로 수직 사영하여 얻는 벡터의 길이이다.
a b = { ‖ a ‖ cos ∡ ( a , b ) a ≠ 0 ∧ b ≠ 0 0 a = 0 ∨ b = 0 {\displaystyle \mathbf {a} _{\mathbf {b} }={\begin{cases}\Vert a\Vert \cos \measuredangle (\mathbf {a} ,\mathbf {b} )&\mathbf {a} \neq \mathbf {0} \land \mathbf {b} \neq \mathbf {0} \\0&\mathbf {a} =\mathbf {0} \lor \mathbf {b} =\mathbf {0} \end{cases}}} 스칼라 사영은 다음과 같이 단위 벡터 와의 스칼라곱으로 나타낼 수 있다.
a b = a ⋅ b ‖ b ‖ {\displaystyle \mathbf {a} _{\mathbf {b} }=\mathbf {a} \cdot {\frac {\mathbf {b} }{\Vert \mathbf {b} \Vert }}} 반대로, 스칼라곱은 다음과 같이 스칼라 사영과 벡터의 길이의 곱으로 나타낼 수 있다.
a ⋅ b = ‖ b ‖ a b = ‖ a ‖ b a {\displaystyle \mathbf {a} \cdot \mathbf {b} =\Vert \mathbf {b} \Vert \mathbf {a} _{\mathbf {b} }=\Vert \mathbf {a} \Vert \mathbf {b} _{\mathbf {a} }} 코사인 법칙 삼각형의 세 변에 대응하는 세 벡터 a , b , c 와 이들 가운데 두 벡터의 각도 θ . 삼각형의 세 변 a , b , c {\displaystyle a,b,c} 와 c {\displaystyle c} 가 마주보는 각 θ {\displaystyle \theta } 에 대한 코사인 법칙 은 스칼라곱의 성질을 통해 유도할 수 있다. 벡터 a , b , c {\displaystyle \mathbf {a} ,\mathbf {b} ,\mathbf {c} } 가 그림과 같다고 하면, 코사인 법칙은 다음과 같이 증명된다.
c 2 = c ⋅ c = ( a − b ) ⋅ ( a − b ) = a ⋅ a − b ⋅ a − a ⋅ b + b ⋅ b = a ⋅ a − 2 a ⋅ b + b ⋅ b = a 2 − 2 a b cos θ + b 2 {\displaystyle {\begin{aligned}c^{2}&=\mathbf {c} \cdot \mathbf {c} \\&=(\mathbf {a} -\mathbf {b} )\cdot (\mathbf {a} -\mathbf {b} )\\&=\mathbf {a} \cdot \mathbf {a} -\mathbf {b} \cdot \mathbf {a} -\mathbf {a} \cdot \mathbf {b} +\mathbf {b} \cdot \mathbf {b} \\&=\mathbf {a} \cdot \mathbf {a} -2\mathbf {a} \cdot \mathbf {b} +\mathbf {b} \cdot \mathbf {b} \\&=a^{2}-2ab\cos \theta +b^{2}\end{aligned}}} 삼중곱 3차원 유클리드 공간 R 3 {\displaystyle \mathbb {R} ^{3}} 속 벡터에 대한 곱셈은 그 밖에도 여럿 존재한다. 예를 들어, 두 벡터 a , b ∈ R 3 {\displaystyle \mathbf {a} ,\mathbf {b} \in \mathbb {R} ^{3}} 의 벡터곱 a × b ∈ R 3 {\displaystyle \mathbf {a} \times \mathbf {b} \in \mathbb {R} ^{3}} 은 스칼라곱과 달리 두 벡터로부터 또 다른 벡터를 얻는다. 그러나 이는 3차원이 아닌 유클리드 공간에서 의미를 잃는다.
스칼라 삼중곱 은 R 3 {\displaystyle \mathbb {R} ^{3}} 에서 스칼라곱과 벡터곱을 사용하여 정의된다. 세 벡터 a , b , c ∈ R 3 {\displaystyle \mathbf {a} ,\mathbf {b} ,\mathbf {c} \in \mathbb {R} ^{3}} 의 스칼라 삼중곱은 a ⋅ ( b × c ) ∈ R {\displaystyle \mathbf {a} \cdot (\mathbf {b} \times \mathbf {c} )\in \mathbb {R} } 로 정의된다.
벡터 삼중곱 은 R 3 {\displaystyle \mathbb {R} ^{3}} 에서 두 번의 벡터 곱으로 정의된다. 세 벡터 a , b , c ∈ R 3 {\displaystyle \mathbf {a} ,\mathbf {b} ,\mathbf {c} \in \mathbb {R} ^{3}} 의 벡터 삼중곱은 스칼라곱을 계수로 하는 선형 결합 전개식으로 나타낼 수 있으며, 이는 다음과 같다.
a × ( b × c ) = ( a ⋅ c ) b − ( a ⋅ b ) c {\displaystyle \mathbf {a} \times (\mathbf {b} \times \mathbf {c} )=(\mathbf {a} \cdot \mathbf {c} )\mathbf {b} -(\mathbf {a} \cdot \mathbf {b} )\mathbf {c} } 물리학 물리학 의 여러 가지 개념은 스칼라곱을 통해 정의된다. 예를 들어, 일 은 힘 과 변위 의 스칼라곱이며, 자기 선속 은 자기 선속 밀도 와 면적 벡터의 스칼라곱이다. 물론 변하는 힘이나 일정하지 않은 자기 선속의 경우 적분 을 사용한다.
일반화
복소수 벡터의 경우 차원이 n {\displaystyle n} 인 복소수 곱공간 C n {\displaystyle \mathbb {C} ^{n}} 속의 벡터 u , v ∈ C n {\displaystyle \mathbf {u} ,\mathbf {v} \in \mathbb {C} ^{n}} 에 대하여 스칼라곱과 비슷한 함수를 정의할 수 있으며, 이는 다음과 같다.
u ⋅ v = u ∗ v = u 1 ¯ v 1 + ⋯ + u n ¯ v n {\displaystyle \mathbf {u} \cdot \mathbf {v} =\mathbf {u} ^{*}\mathbf {v} ={\overline {u_{1}}}v_{1}+\cdots +{\overline {u_{n}}}v_{n}} 여기서 u ∗ {\displaystyle \mathbf {u} ^{*}} 는 u {\displaystyle \mathbf {u} } 의 (열벡터로서의) 켤레전치 이며, u k ¯ {\displaystyle {\overline {u_{k}}}} 는 u k {\displaystyle u_{k}} 의 켤레 복소수 이다. 이러한 함수는 양의 정부호성 을 만족시킨다. 즉, 영벡터가 아닌 복소수 벡터와 자기 자신의 스칼라곱은 항상 실수이며 0보다 크다. 그러나 실수 벡터의 스칼라곱과 달리 쌍선형성을 만족시키지 않으며, 대신 다음과 같은 반쌍선형성 을 만족시킨다. 임의의 u , v , w ∈ C n {\displaystyle \mathbf {u} ,\mathbf {v} ,\mathbf {w} \in \mathbb {C} ^{n}} 및 c ∈ C {\displaystyle c\in \mathbb {C} } 에 대하여,
( c u + v ) ⋅ w = c ¯ u ⋅ w + v ⋅ w {\displaystyle (c\mathbf {u} +\mathbf {v} )\cdot \mathbf {w} ={\bar {c}}\mathbf {u} \cdot \mathbf {w} +\mathbf {v} \cdot \mathbf {w} } u ⋅ ( c v + w ) = c u ⋅ v + u ⋅ w {\displaystyle \mathbf {u} \cdot (c\mathbf {v} +\mathbf {w} )=c\mathbf {u} \cdot \mathbf {v} +\mathbf {u} \cdot \mathbf {w} } 또한 대칭성(교환 법칙) 대신 다음과 같은 켤레 대칭성을 만족시킨다.
u ⋅ v = v ⋅ u ¯ {\displaystyle \mathbf {u} \cdot \mathbf {v} ={\overline {\mathbf {v} \cdot \mathbf {u} }}} 이 경우, 영벡터가 아닌 두 복소수 벡터의 사잇각을 나타내는 공식은 다음과 같다.
cos ∡ ( u , v ) = Re ( u ⋅ v ) ‖ u ‖ ‖ v ‖ {\displaystyle \cos \measuredangle (\mathbf {u} ,\mathbf {v} )={\frac {\operatorname {Re} (\mathbf {u} \cdot \mathbf {v} )}{\Vert \mathbf {u} \Vert \Vert \mathbf {v} \Vert }}} 여기서 Re ( u ⋅ v ) {\displaystyle \operatorname {Re} (\mathbf {u} \cdot \mathbf {v} )} 는 복소수 u ⋅ v {\displaystyle \mathbf {u} \cdot \mathbf {v} } 의 실수부 이다.
만약 이 함수의 정의에서 켤레 복소수를 생략한다면, 이는 쌍선형성과 대칭성을 유지하지만 양의 정부호성을 잃는다. 이는 대략 i 2 = − 1 < 0 {\displaystyle i^{2}=-1<0} 이기 때문이다. 사실, 양의 정부호성과 대칭성과 쌍선형성을 동시에 만족시키는 함수 B : C n × C n → C {\displaystyle B\colon \mathbb {C} ^{n}\times \mathbb {C} ^{n}\to \mathbb {C} } 는 존재하지 않는다. 이는 이러한 함수의 존재가 다음과 같은 모순을 가져오기 때문이다.
0 < B ( i v , i v ) = i 2 B ( v , v ) = − B ( v , v ) < 0 ( v ∈ C n ∖ { 0 } ) {\displaystyle 0 내적 유클리드 공간이나 복소수 곱공간의 스칼라곱을 일반화하여 내적 의 개념을 얻을 수 있다. 실수 벡터 공간 V {\displaystyle V} 에서, 두 벡터 u , v ∈ V {\displaystyle u,v\in V} 로부터 실수 스칼라 ⟨ u , v ⟩ ∈ R {\displaystyle \langle u,v\rangle \in \mathbb {R} } 를 얻는 연산이 양의 정부호성과 대칭성과 쌍선형성을 만족시킬 경우, 이를 V {\displaystyle V} 위의 내적 이라고 한다. 복소수 벡터 공간 V {\displaystyle V} 의 두 벡터 u , v ∈ V {\displaystyle u,v\in V} 로부터 복소수 스칼라 ⟨ u , v ⟩ ∈ C {\displaystyle \langle u,v\rangle \in \mathbb {C} } 를 얻는 연산이 양의 정부호성과 켤레 대칭성과 반쌍선형성을 만족시킬 경우, 이를 V {\displaystyle V} 위의 내적이라고 한다. 예를 들어, R 2 {\displaystyle \mathbb {R} ^{2}} 에 다음과 같은 함수를 정의하면, 이는 내적을 이룬다.:271
( a 1 , a 2 ) ⋅ ( b 1 , b 2 ) = a 1 b 1 − a 2 b 1 − a 1 b 2 + 4 a 2 b 2 {\displaystyle (a_{1},a_{2})\cdot (b_{1},b_{2})=a_{1}b_{1}-a_{2}b_{1}-a_{1}b_{2}+4a_{2}b_{2}} 함수의 경우 두 실숫값 함수 f , g : [ a , b ] → R {\displaystyle f,g\colon [a,b]\to \mathbb {R} } 의 내적 ⟨ f , g ⟩ ∈ R {\displaystyle \langle f,g\rangle \in \mathbb {R} } 은 급수 대신 적분 을 사용하여 다음과 같이 정의할 수 있으며, 이 역시 양의 정부호성과 대칭성과 쌍선형성을 만족시킨다.
⟨ f , g ⟩ = ∫ a b f ( x ) g ( x ) d x {\displaystyle \langle f,g\rangle =\int _{a}^{b}f(x)g(x)dx} 보다 일반적으로, 두 복소숫값 함수 f , g : [ a , b ] → C {\displaystyle f,g\colon [a,b]\to \mathbb {C} } 의 내적 ⟨ f , g ⟩ ∈ C {\displaystyle \langle f,g\rangle \in \mathbb {C} } 은 다음과 같으며, 이는 양의 정부호성과 켤레 대칭성과 반쌍선형성을 만족시킨다.
⟨ f , g ⟩ = ∫ a b f ( x ) ¯ g ( x ) d x {\displaystyle \langle f,g\rangle =\int _{a}^{b}{\overline {f(x)}}g(x)dx} 행렬의 경우 사이즈가 같은 두 실수 행렬 A , B {\displaystyle A,B} 의 프로베니우스 내적 (영어 : Frobenius inner product ) A : B {\displaystyle A:B} 은 위치가 같은 두 성분의 곱들을 합한 결과이며, 대각합 과 행렬 곱셈 을 통해 나타낼 수도 있다. 즉, 다음과 같다.
A : B = tr ( A T B ) = ∑ i , j A i j B i j {\displaystyle A:B=\operatorname {tr} (A^{\operatorname {T} }B)=\sum _{i,j}A_{ij}B_{ij}} 보다 일반적으로, 두 복소수 행렬 A , B {\displaystyle A,B} 의 프로베니우스 내적은 다음과 같다.
A : B = tr ( A ∗ B ) = ∑ i , j A i j ¯ B i j {\displaystyle A:B=\operatorname {tr} (A^{*}B)=\sum _{i,j}{\overline {A_{ij}}}B_{ij}} 여기서 A ∗ {\displaystyle A^{*}} 는 A {\displaystyle A} 의 켤레전치 이다.
각주