Functional equation

Calculus Level 5

Let f ( x ) f(x) be the functional equation that satisfies x = f ( x ) e f ( x ) x = f(x) e ^ { f(x) } for all real vaues x x . Find

0 e f ( x ) d x . \int_0^e f(x) \, dx.


The answer is 1.7182.

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

9 solutions

Kenny Lau
Oct 12, 2015

I don't care if it is the Lambert W function .

x = f ( x ) e f ( x ) f 1 x = x e x x=f(x)e^{f(x)} \implies f^{-1}x=xe^x .

I = 0 e f ( x ) d x \displaystyle I=\int_0^e f(x)\mathrm dx .

Let u = f ( x ) u=f(x) , x = u e u x=ue^u .

When x = 0 x=0 , u = 0 u=0 . When x = e x=e , u = 1 u=1 .

I = 0 1 u d x = 0 1 u d ( u e u ) \displaystyle I=\int_0^1 u\mathrm dx=\int_0^1 u\mathrm d(ue^u) .

Integration by parts:

I = [ u 2 e u ] 0 1 0 1 u e u d u I=[u^2e^u]_0^1-\int_0^1ue^u\mathrm du

I = [ u 2 e u ] 0 1 0 1 u d ( e u ) I=[u^2e^u]_0^1-\int_0^1u\mathrm d(e^u)

I = [ u 2 e u ] 0 1 [ u e u ] 0 1 + 0 1 e u d u I=[u^2e^u]_0^1-[ue^u]_0^1+\int_0^1e^u\mathrm du

I = [ u 2 e u ] 0 1 [ u e u ] 0 1 + [ e u ] 0 1 I=[u^2e^u]_0^1-[ue^u]_0^1+[e^u]_0^1

I = ( e 0 ) ( e 0 ) + ( e 1 ) I=(e-0)-(e-0)+(e-1)

I = e 1 I=e-1 .

Ben Williams
Feb 4, 2015

Let I = 0 e f ( x ) d x I = \int^{e}_{0} f(x) dx See that d d x [ f ( x ) e f ( x ) ] = 1 I = 0 e d d x [ f ( x ) e f ( x ) ] f ( x ) \frac{d}{dx} [f(x)e^{f(x)}]=1 \implies I=\int^{e}_{0} \frac{d}{dx}[f(x)e^{f(x)}]f(x)
Then: I = 0 e f ( x ) 2 e f ( x ) 0 e f ( x ) d d x [ e f ( x ) ] d x I = \bigg|^{e}_{0} f(x)^{2}e^{f(x)} - \int^{e}_{0} f(x) \frac{d}{dx} [e^{f(x)}] dx I = 0 e e f ( x ) ( f ( x ) 2 f ( x ) + 1 ) \implies I = \bigg|^{e}_{0} e^{f(x)} (f(x)^{2} - f(x) +1 ) Now we know that when x = 0 x=0 , f ( x ) e f ( x ) = 0 f(x)e^{f(x)} =0 Assuming that: lim x 0 f ( x ) \lim_{x \rightarrow 0} f(x) \neq - \infty (or more broadly, that the function is well defined at x = 0 x= 0 ) We also know that this implies f ( 0 ) = 0 f(0) = 0 . Similarly, when x = e x=e , we know: f ( e ) e f ( e ) = e f ( e ) = e 1 f ( e ) f(e)e^{f(e)} = e \implies f(e) = e^{1-f(e)} This is easily solved to yield f ( e ) = 1 f(e) = 1 Plugging the numbers back into the formula for I I , we get : I = e ( 1 1 + 1 ) 1 ( 0 + 0 + 1 ) = e 1 I = e(1-1+1)-1(0+0+1) = e-1

Jake Lai
Feb 4, 2015

f ( x ) f(x) , the inverse of x e x xe^{x} , is known as the the Lambert W function .

It can be seen by integration by parts that

0 e f ( x ) d x = e × 1 0 1 x e x d x \int_{0}^{e} f(x) dx = e \times 1 - \int_{0}^{1} xe^{x} dx

= e x e x e x d x 0 1 = e e x ( x 1 ) 0 1 = e-|xe^x - \int e^{x} dx|_{0}^{1} = e-|e^{x}(x-1)|_{0}^{1}

= e 1 = \boxed{e-1}

I meant to post this solution 6 hours ago. Somehow I got so distracted that even a simple click took 6 hours. Hooray.

Jake Lai - 6 years, 4 months ago

Log in to reply

Yay to having your solutions be saved automatically :)

Calvin Lin Staff - 6 years, 4 months ago
Amrit Anand
Oct 21, 2015

Nice solution.

Abhinav Jha - 5 years, 4 months ago
Raushan Sharma
Jun 6, 2016

I don't know the Lambert function, but here is my solution:

Take f ( x ) = y f(x) =y :

We have the equation: x = y e y x=y\cdot e^y

x y = e y l n ( x ) l n ( y ) = y \Rightarrow \frac{x}{y} = e^y \Rightarrow ln(x) - ln(y) = y [Taking log (base e) on both sides]

Now differentiating both sides w.r.t x x , we get:

1 x 1 y d y d x = d y d x \frac{1}{x} - \frac{1}{y} \cdot \frac{dy}{dx} = \frac{dy}{dx}

d y d x = y x ( y + 1 ) y d x = x ( y + 1 ) d y = y e y ( y + 1 ) d y = e y ( y 2 + y ) d y \Rightarrow \frac{dy}{dx} = \frac{y}{x(y+1)} \Rightarrow y\cdot dx = x(y+1) dy = ye^y(y+1) dy = e^y(y^2+y)dy

Now, we need to find 0 e y d x = 0 1 e y ( y 2 + y ) d y \int_0^e y \, dx = \int_0^1 e^y(y^2+y) \, dy

Applying Integration by Parts, we have:

0 1 e y ( y 2 + y ) d y = [ e y ( y 2 y + 1 ) ] 0 1 = e 1 = 1.7182 \int_0^1 e^y(y^2+y) \, dy = [e^y(y^2 -y+1)]_0^1 = e-1 = 1.7182

Santiago Hincapie
Oct 15, 2015

\int { f(x) } =xf(x)+g(x)\\ where\quad g(x)\quad was\quad a\quad function\\ f(x)=f(x)+x\frac { d }{ dx } f(x)+\frac { dg(x) }{ dx } \\ g(x)=-\int { xdf(x) } \\ g(x)=-\int { f(x)e^{ f(x) }df(x) } \\ g(x)=-[f(x)e^{ f(x) }-e^{ f(x) }]\\ then\\ \int { f(x)dx } =xf(x)+e^{ f(x) }(1-f(x))\\ and\\ \int _{ 0 }^{ e }{ f(x)dx } =[ef(e)+e^{ f(e) }(1-f(e))]-[0f(0)+e^{ f(0) }(1-f(0))]\\ note\quad that\\ f(0)=0\quad and\quad f(e)=1\\ then\\ \int _{ 0 }^{ e }{ f(x)dx } =e-1\approx 1.71828183

Augusto Bernardi
Feb 7, 2015

d x d x = d ( f ( x ) e f ( x ) ) d x = e f ( x ) f ( x ) + e f ( x ) f ( x ) f ( x ) = 1 \frac { dx }{ dx } =\frac { d(f\left( x \right) { e }^{ f\left( x \right) }) }{ dx } ={ e }^{ f\left( x \right) }f^{ ' }\left( x \right) +{ e }^{ f\left( x \right) }f^{ ' }\left( x \right) f\left( x \right) =1

Which implies that

1 f 1 ( x ) = e f ( x ) + e f ( x ) f ( x ) { \frac { 1 }{ f^{ 1 }\left( x \right) } =e }^{ f\left( x \right) }+{ e }^{ f\left( x \right) }f\left( x \right) (eq.1)

Now, looking at the integral:

f ( x ) d x = f ( x ) f ( x ) d f ( x ) \int { f\left( x \right) dx= } \int { \frac { f\left( x \right) }{ f^{ ' }\left( x \right) } } df\left( x \right)

Using eq.1

f ( x ) f ( x ) d f ( x ) = f ( x ) 2 e f ( x ) d f ( x ) + f ( x ) e f ( x ) d f ( x ) \int { \frac { f\left( x \right) }{ f^{ ' }\left( x \right) } } df\left( x \right) =\int { { f\left( x \right) }^{ 2 } } { e }^{ f\left( x \right) }df\left( x \right) +\int { { f\left( x \right) } } { e }^{ f\left( x \right) }df\left( x \right)

Integrating by parts

f ( x ) f ( x ) d f ( x ) = f ( x ) 2 e f ( x ) d f ( x ) + f ( x ) e f ( x ) d f ( x ) = \int { \frac { f\left( x \right) }{ f^{ ' }\left( x \right) } } df\left( x \right) =\int { { f\left( x \right) }^{ 2 } } { e }^{ f\left( x \right) }df\left( x \right) +\int { { f\left( x \right) } } { e }^{ f\left( x \right) }df\left( x \right)= = f ( x ) 2 e f ( x ) 2 f ( x ) e f ( x ) d f ( x ) + f ( x ) e f ( x ) d f ( x ) ={ f\left( x \right) }^{ 2 }{ e }^{ f\left( x \right) }-2\int { { f\left( x \right) } } { e }^{ f\left( x \right) }df\left( x \right) +\int { { f\left( x \right) } } { e }^{ f\left( x \right) }df\left( x \right)

Integrating by parts again

f ( x ) 2 e f ( x ) f ( x ) e f ( x ) d f ( x ) = f ( x ) 2 e f ( x ) + e f ( x ) f ( x ) e f ( x ) { f\left( x \right) }^{ 2 }{ e }^{ f\left( x \right) }-\int { { f\left( x \right) } } { e }^{ f\left( x \right) }df\left( x \right) ={ f\left( x \right) }^{ 2 }{ e }^{ f\left( x \right) }+{ e }^{ f\left( x \right) }-f\left( x \right) { e }^{ f\left( x \right) }

Therefore:

0 e f ( x ) d x = e [ f ( e ) 1 ] + e f ( e ) e f ( 0 ) \int _{ 0 }^{ e }{ f\left( x \right) dx=e[f(e)-1] } +{ e }^{ f\left( e \right) }-{ e }^{ f(0) } (eq.2)

Back to the integral:

f ( x ) d x = l n ( x ) d x l n ( f ( x ) ) d x \int { f(x)dx=\int { ln(x)dx-\int { ln(f(x))dx } } }

Using (eq.1) in the second term and using integral by parts in both terms, we find that:

f ( x ) d x = x [ l n ( x ) 1 ] + e f ( x ) [ 1 f ( x ) l n ( f ( x ) ) ] \int { f(x)dx=\quad x[ln(x)-1]+{ e }^{ f\left( x \right) } } [1-f(x)ln(f(x))]

And therefore:

0 e f ( x ) d x = e f ( e ) e l n ( f ( e ) ) e f ( 0 ) \int _{ 0 }^{ e }{ f(x)dx } ={ e }^{ f(e) }-eln(f(e))-{ e }^{ f(0) } (eq.3)

Using eq.2 and eq.3, we find that

f ( e ) 1 = l n ( f ( e ) ) f ( e ) = 1 f(e)-1=ln(f(e))\Rightarrow f(e)=1

Since e f ( 0 ) = 1 { e }^{ f\left( 0 \right) }=1 , we conclude, by eq.2 or eq.3, that

0 e f ( x ) d x = e 1 \int _{ 0 }^{ e }{ f(x)dx=e-1 }

Prakhar Agarwal
Feb 4, 2015

let y=f(x) therefore, y(e)^y=x, it is an invertible function. => ∫ydx is area of y=f(x) with x-axis from x=0 to x=e which is equal to [e*1-∫xdy] where y is from 0 to 1. =e-1

If x = y e y x=ye^y then y = W ( x ) y=W(x) , where W ( x ) W(x) is the Lambert-W function.

0 e W ( x ) d x = e 1 = 1.718 \int_0^e W(x) dx = e-1 = \boxed{1.718}

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...