As Little Work As Possible

Calculus Level 5

A particle can move on an infinite two-dimensional plane, which is given an O x y Oxy -coordinate system. As the particle moves on the plane, when it is at the point with coordinates ( x , y ) (x,y) , it experiences a resistance force equal to e y e^{-y} Newtons in a direction opposite to that of motion of the particle. Thus the work done moving the particle along a particular path C C is C e y d s \int_C e^{-y}\,ds Joules, where s s represents the arc-length, measured in metres.

The particle has to be moved smoothly from the point A = ( 0 , 0 ) A=(0,0) to the point B = ( 1 2 π , 1 ) B=\big(\tfrac12\pi,1\big) , in such a way that the x x -coordinate of its position never decreases. Thus a possible path from A A to B B is of the form ( x , y ( x ) ) (x,y(x)) , where y : [ 0 , 1 2 π ] R y\,:\, \left[0,\tfrac12\pi\right] \to \mathbb{R} is a smooth function such that y ( 0 ) = 0 y(0) = 0 and y ( 1 2 π ) = 1 y\big(\tfrac12\pi\big) = 1 .

Of all the possible paths from A A to B B there is one for which the work done against the resistance force in moving the particle from A A to B B is as small as possible. The length of this path can be shown to be P tanh 1 ( e Q + R S e + T ) , P \tanh^{-1}\left( \frac{\sqrt{e^Q + R}}{Se + T}\right), where P , Q , R , S , T P,Q,R,S,T are positive integers. What is the value of P + Q + R + S + T P+Q+R+S+T ?


The answer is 7.

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.

1 solution

Mark Hennings
Oct 6, 2016

If y : [ 0 , 1 2 π ] R y\,:\, [0,\tfrac12\pi] \to \mathbb{R} is a smooth map such that y ( 0 ) = 0 y(0) = 0 and y ( 1 2 π ) = 1 y(\tfrac12\pi) = 1 , then the work done moving the particle along the path defined by the function y y is W [ y ] = 0 1 2 π e y ( x ) d s = 0 1 2 π e y ( x ) 1 + ( y ( x ) ) 2 d x W[y] \; = \; \int_0^{\frac12\pi} e^{-y(x)} \,ds \; =\; \int_0^{\frac12\pi} e^{-y(x)} \sqrt{1 + \big(y'(x)\big)^2}\,dx The theory of Calculus of Variations tells us that the function y y which extremizes W [ y ] W[y] must satisfy the Euler equation d d x ( F y ) F y = 0 \frac{d}{dx}\left(\frac{\partial F}{\partial y'}\right) - \frac{\partial F}{\partial y}\; = \; 0 where F ( x , y , y ) = e y 1 + ( y ) 2 F(x,y,y') \; = \; e^{-y} \sqrt{1 + (y')^2} Thus d d x ( y e y 1 + ( y ) 2 ) + e y 1 + ( y ) 2 = 0 ( y ( y ) 2 ) e y 1 + ( y ) 2 ( y ) 2 y e y ( 1 + ( y ) 2 ) 3 2 + e y 1 + ( y ) 2 = 0 ( y ( y ) 2 ) ( 1 + ( y ) 2 ) ( y ) 2 y + ( 1 + ( y ) 2 ) 2 = 0 y + 1 + ( y ) 2 = 0 \begin{array}{rcl} \displaystyle \frac{d}{dx}\left(\frac{y' e^{-y}}{\sqrt{1 + (y')^2}}\right) + e^{-y}\sqrt{1 + (y')^2} & = & 0 \\ \displaystyle \frac{(y'' - (y')^2)e^{-y}}{\sqrt{1 + (y')^2}} - \frac{(y')^2y'' e^{-y}}{(1 + (y')^2)^{\frac32}} + e^{-y}\sqrt{1 + (y')^2} & = & 0 \\ \displaystyle \big(y'' - (y')^2\big)\big(1 + (y')^2\big) - (y')^2 y'' + \big(1 + (y')^2\big)^2 & = & 0 \\ \displaystyle y'' + 1 + (y')^2 & = & 0 \end{array} Putting z = e y z = e^y , this equation becomes z + z = 0 z'' + z \; = \; 0 and hence z = a sin x + b cos x z = a\sin x + b\cos x for some constants a , b a,b . Since y ( 0 ) = 0 y(0) = 0 and y ( 1 2 π ) = 1 y(\tfrac12\pi) = 1 we deduce that e y = z = e sin x + cos x e^y\,=\, z \,=\, e\sin x + \cos x , and so the optimal curve is y ( x ) = ln ( e sin x + cos x ) 0 x 1 2 π . y(x) \; = \; \ln\big(e\sin x + \cos x) \hspace{1cm} 0 \le x \le \tfrac12\pi\;. For this curve we calculate ( d s d x ) 2 = 1 + ( d y d x ) 2 = e 2 + 1 ( e sin x + cos x ) 2 \left(\frac{ds}{dx}\right)^2 \; = \; 1 + \left(\frac{dy}{dx}\right)^2 \; = \; \frac{e^2+1}{(e\sin x + \cos x)^2} and hence that d s d x = e 2 + 1 e sin x + cos x = sec ( x α ) \frac{ds}{dx} \; = \; \frac{\sqrt{e^2+1}}{e\sin x + \cos x} \; = \; \sec(x-\alpha) where α = tan 1 e \alpha = \tan^{-1}e .

Thus the length of the optimal curve is s = 0 1 2 π sec ( x α ) d x = [ ln ( sec ( x α ) + tan ( x α ) ) ] 0 1 2 π = ln ( c o s e c α + cot α ) ln ( sec α tan α ) = ln [ ( sec α + tan α ) ( c o s e c α + cot α ) ] = ln ( 1 + e + e 2 + ( 1 + e ) e 2 + 1 e ) \begin{array}{rcl} \displaystyle s & = & \displaystyle \int_0^{\frac12\pi} \sec(x - \alpha)\,dx \; = \; \Big[ \ln(\sec (x-\alpha) + \tan (x-\alpha)) \Big]_0^{\frac12\pi} \\ & = & \displaystyle \ln(\mathrm{cosec}\,\alpha + \cot\alpha) - \ln(\sec\alpha - \tan\alpha) \; = \; \ln\big[(\sec\alpha+\tan\alpha)(\mathrm{cosec}\,\alpha + \cot\alpha)\big] \\ & = & \displaystyle\ln \left( \frac{1 + e + e^2 + (1+e)\sqrt{e^2+1}}{e}\right) \end{array} If we write s = 2 tanh 1 v s = 2\tanh^{-1}v then v = tanh 1 2 s = e s 1 e s + 1 = 1 + e 2 + ( 1 + e ) e 2 + 1 ( 1 + e ) 2 + ( 1 + e ) e 2 + 1 = e 2 + 1 e + 1 v \; = \; \tanh\tfrac12s \; =\; \frac{e^s-1}{e^s+1} \; = \; \frac{1+e^2 + (1+e)\sqrt{e^2+1}}{(1+e)^2 + (1+e)\sqrt{e^2+1}} \; = \; \frac{\sqrt{e^2+1}}{e+1} so we deduce that s = 2 tanh 1 ( e 2 + 1 e + 1 ) s \; = \; 2\tanh^{-1}\left(\frac{\sqrt{e^2+1}}{e+1}\right) making the answer 2 + 2 + 1 + 1 + 1 = 7 2 + 2 + 1 + 1 + 1 = \boxed{7} .

@Mark Hennings Is calculus of variations the only applicable tool here?

A Former Brilliant Member - 4 years, 8 months ago

Log in to reply

It's the only one I can think of.

Mark Hennings - 4 years, 8 months ago

phew...............it caused such a pain in the neck ! @Mark Hennings

A Former Brilliant Member - 4 years, 8 months ago

Log in to reply

Glad you (I think) enjoyed it!

Mark Hennings - 4 years, 8 months ago

Log in to reply

of courseeeeeeeeeeeeeeee , i enjoy ur every ques !

A Former Brilliant Member - 4 years, 8 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...