Calculus of variations

Japanese: 変分法 - へんぶんほう(英語表記)calculus of variations
Calculus of variations

The extreme value problem dealt with in differential calculus involves finding the maximum or minimum value of a function f ( x1 , x2 , …, xn ) of n variables ( x1 , x2 , …, xn ) moving within a certain domain, whereas the calculus of variations involves considering problems like the following famous example (by J. Bernoulli in 1696) .

Two points A =( x 0 , y 0 ), B =( x 1 , y 1 ), x 1x 0 , y 1y 0 in the perpendicular x - y plane are connected by a smooth curve y =( x ).When a ball is slid along this curve without friction, what is the curve that will reach point B in the shortest time?

To express this mathematically, the velocity of the particle along the curve y = ( x ) is,

(where g is the acceleration of gravity), the time it takes is

If we omit the non-essential constants (2g) and y 0 , then among the continuously differentiable curves y =( x ) such that ( x 0 ) = 0, ( x 1 ) = y 1,

The problem becomes: find ( x ) that minimizes (the solution is a cycloid).

In this way, the calculus of variations is used to solve problems of the form: Find ( x ) that minimizes a real-valued function I () of function ( x ).


Then, the necessary and sufficient condition for ( x ) = u ( x ) to have a maximum or minimum is Euler's equation

The aim is to satisfy the following.

Similarly, let G be a bounded region of two-dimensional space. Among the functions on the boundary of G such that ( x , y ) = g ( x , y ) (the given function),

The problem of finding ( x , y ) = u ( x , y ) that minimizes is expressed as Euler's equation

Satisfy.

One method is to use Euler's equation to solve the calculus of variations problem, but there are also direct methods to solve the calculus of variations problem directly. In particular, it is often used to convert a differential equation into a variation problem with the Euler equation and solve it with the direct calculus of variations method.

There is the Ritz method for finding an approximate solution using a direct method. To explain it in terms of the Poisson equation, take a continuously differentiable function ( x , y ) on G with boundary value (x, y ) = g ( x , y ), and set u ( x , y ) = ( x , y ) - w ( x , y ), then

To minimize this, we need to find w , which is 0 on the boundary. As an approximation, we can take appropriate functions 1 ( x , y ), …, n ( x , y ), which are 0 on the boundary, and
w n ( x , y )= c 1 1 ( x , y )+……+ c n n ( x , y )
The coefficients c 1 , c 2 ,……, c n of

To do this, we use the following method instead of the usual extreme value problem of differential calculus:

Then, we can determine c 1 , c 2 , …, c n . In other words, the simultaneous linear equations

By determining c 1 , c 2 , …, c n , we get the approximate solution

is obtained.

[Haruo Sunouchi]

Calculus of variations (cycloid curve)
©Shogakukan ">

Calculus of variations (cycloid curve)


Source: Shogakukan Encyclopedia Nipponica About Encyclopedia Nipponica Information | Legend

Japanese:

微分法で取り扱われる極値問題は、ある領域内を動くn個の変数(x1,x2,……,xn)の関数f(x1,x2,……,xn)の最大値、最小値を求めるのに対し、変分法では、次の有名な例(1696年、J・ベルヌーイによる)のような問題を考える。

 垂直なx-y平面の二点A=(x0,y0),B=(x1,y1),x1x0,y1y0を滑らかな曲線y=(x)で結び、この曲線に沿って一つの球を摩擦を受けることなく滑り落とすとき、点Bへ最短時間で到達する曲線を求めよ。

 これを数式で表すには、曲線y=(x)に沿っての粒子の速度は、

(落下した垂直距離の平方根、gは重力の加速度)に比例することを思い出すと、所要時間は

となる。本質的でない定数(2g)やy0を略すと、(x0)=0,(x1)=y1となる連続微分可能な曲線y=(x)のうち

を最小にする(x)を求めよ、という問題になる(この解はサイクロイドである)。

 このように、関数(x)の実数値関数I()を最小にする(x)を求めよ、という形の問題を解くのが変分法である。


のとき、(x)=u(x)で最大、最小をとる必要十分条件は、オイラーの方程式

を満足することである。

 同様に、二次元空間の有界領域をGとし、Gの境界上で(x,y)=g(x,y)(与えられた関数)となる関数のうちで

を最小にする(x,y)=u(x,y)を求めよ、という問題は、オイラーの方程式として

を満足する。

 変分法の問題をオイラーの方程式を用いて解くのも一つの方法であるが、変分法の問題を直接解く直接法もある。とくに、微分方程式をオイラー方程式としてもつ変分問題に直し、変分法の直接法で解くこともよく用いられる。

 直接法で近似解を求めるのにリッツの方法がある。それをポアソンの方程式について説明をすると、境界値として(x,y)=g(x,y)となるG上の連続微分可能な関数(x,y)をとり、u(x,y)=(x,y)-w(x,y)と置くと、

を最小とし、境界上で0となるwを求めればよい。その近似として、境界上で0となる関数1(x,y),……,n(x,y)を適当にとり、
  wn(x,y)=c11(x,y)+……+cnn(x,y)
の係数c1,c2,……,cn

を最小にするように決める。それには、普通の微分法の極値問題より、

よりc1c2、……、cnを決めればよい。すなわち、連立一次方程式

よりc1c2、……、cnを決めると、近似解

が得られる。

[洲之内治男]

変分法(サイクロイド曲線)
©Shogakukan">

変分法(サイクロイド曲線)


出典 小学館 日本大百科全書(ニッポニカ)日本大百科全書(ニッポニカ)について 情報 | 凡例

<<:  Condyloma flatus - Hempeikonjiro-ma

>>:  Bian Wen-jin (English spelling)

Recommend

Hezhou

Another name for Kawachi Province (eastern Osaka P...

Church - Kyoukaido

〘noun〙 A permanent building established by Christi...

Henrik Pontoppidan

A representative Danish naturalist writer. Born t...

Feathery Bainite - Umoujoubenaito

…In steels containing alloying elements such as m...

Stochastic radiation damage

...The occurrence of leukemia and various cancers...

Cantilever beam

A cantilever is a beam with one end fixed and the ...

Curry,HB - Curry

…The amount of tape used is not limited in advanc...

Campion, T.

...From the Elizabethan period through to the rei...

NLA - Number one eleven

→Neurorep anesthesia Source: Shogakukan Encyclope...

Deccan Plateau - English spelling: Deccan Plateau

A plateau occupying the southern half of India. T...

Reconciliation - Wakai

[1] 〘 noun 〙① To make amends by ceasing a dispute....

Takamori [town] - Takamori

A town in Aso County in northeastern Kumamoto Pref...

Driving Impairment - Unten Sogai

...A railway accident is an event that occurs tha...

Conceptual Schema

See the Schema page. Source: ASCII.jp Digital Dic...

ICSU - ICSU

International Council for Science. An internationa...