Analyzing the Brachistochrone Problem for the Conic Sections and its real lfie applications
Aarush Patel
August 2024
Abstract
This paper explores the brachistochrone problem for several curves. Throughout the paper, we find the time taken by an object of any mass to travel from the initial position (1,0)(1,0) to the final position (0,1)(0,1) through any curve. Finding a generalized integral time formula applicable to all curves of the form y=f(x)y=f(x), we further apply the formula to a line and the following conic sections - parabola of form y=ax^(2)+bx+cy=a x^{2}+b x+c and x=ay^(2)+by+cx=a y^{2}+b y+c, , circle of the form (x-a)^(2)+(y-b)=r^(2)(x-a)^{2}+(y-b)=r^{2}, ellipse with xx and yy as major axes, and hyperbola with xx and yy as major axis. The parabola with equation x=ay^(2)+by+cx=a y^{2}+b y+c was found to be the curve of fastest descent among all the conic sections. These findings are also observable in real life scenarios. For example, during descent, roller coasters are generally parabolic as they would provide more thrill to the visitors than a straight one.
Contents
1 Introduction ..... 1
2 Generalized integral time formula for any curve and descent for a straight line ..... 2
3 Time taken for the descent by various conic sections ..... 3
3.1 Parabola ..... 4
3.2 Circle ..... 4
3.3 Ellipse ..... 5
3.4 Hyperbola ..... 7
4 Conclusion ..... 8
1 Introduction
The brachistochrone problem, proposed by Johann Bernoulli in 1696, aims to find, for points AA and BB with AA located above BB, a curve such that descent of an object due to gravity along that curve takes the shortest possible time. The problem considers only gravitational force on the object and ignores other forces such as friction. It was solved by a number of famous mathematicians such as Johann Bernoulli himself, Jakob Bernoulli, and Isaac Newton, all of them proposing a different approach [4]. Till date, numerous researchers have tried to expand upon the problem by including external factors such as non-uniform gravitational field [1], friction [2] and some other challenges, examining the problem further. These methods have been quite effective but require heavy machinery such as the calculus of variations and the Euler-Lagrange differential equation. Additionally, all of the papers are predominantly focused on deriving the equation of the curve with quickest descent and are not focused on other curves. In this paper, we focus on a more elementary approach and study descent along the conic sections from a point AA with coordinates (0,1)(0,1) to a point BB with coordinates (1,0)(1,0). In Section 2, we utilise a general approach for calculating the time it takes to descend along a curve defined by a function y=f(x)y=f(x). In the following section, 3 we give a detailed account of the time period in cases of parabola, circle, ellipse, and hyperbola respectively. Using an online tool [5] to find the value of integrals, we compare the findings of time periods of different curves. The curves when arranged in decreasing order of their time periods are as follows : line > parabola of form y=f(x)=y=f(x)= hyperbola with yy axis as major axis >> ellipse with yy axis as major axis >> circle >> ellipse with xx axis as major axis == hyperbola
with xx axis as major axis >> parabola of form x=f(y)x=f(y). Applying these findings in real life, we find that water parks generally have curved slides to provide a faster and more enjoyable experience to the person. Additionally, we understand that the shortest curve is not always the fastest one.
The methodology mainly comprises of independent calculations for calculating the time. Additionally, a small literature review was done to gather historical background on the problem.
2 Generalized integral time formula for any curve and descent for a straight line
Let us take an object of mass mm sliding from point AA to point BB. Throughout the paper, we will consider the initial position position (A)(A) to be (0,1)(0,1) and the final position (B)(B) to be (1,0)(1,0). In order to find the time taken, we will first use the law of conservation of energy. The energy at the initial point is just the gravitational potential energy mghm g h where hh is the height. We have h=1h=1 since our initial point is (1,0)(1,0). The energy at any other point PP on the curve will be (1)/(2)mv^(2)+mgy\frac{1}{2} m v^{2}+m g y where vv is the velocity of the object at PP and hh is the yy-coordinate. By conservation of energy,
{:[mg**1=(1)/(2)mv^(2)+mgy],[(1)/(2)mv^(2)=mg(1-y)],[v=sqrt(2g(1-y))]:}\begin{gathered}
m g * 1=\frac{1}{2} m v^{2}+m g y \\
\frac{1}{2} m v^{2}=m g(1-y) \\
v=\sqrt{2 g(1-y)}
\end{gathered}
We have found the velocity at any point. By definition, velocity is the derivative of displacement with respect to time i.e v=(ds)/(dt)v=\frac{d s}{d t}. Rearranging gives dt=(ds)/(v)d t=\frac{d s}{v} and integrating we obtain
t=int_(0)^(1)dt=int_(0)^(1)(ds)/(v)=int_(0)^(1)(ds)/(sqrt(2g(1-y)))t=\int_{0}^{1} d t=\int_{0}^{1} \frac{d s}{v}=\int_{0}^{1} \frac{d s}{\sqrt{2 g(1-y)}}
Note that dsd s is the small change in displacement in a small period of time dtd t. Our aim is to express dsd s in terms of dxd x. In the below graph, the blue line represents the curve y=x^(2)y=x^{2} and the purple line refers to the tangent to the curve at x=1x=1. The angle made by the tangent line with the positive xx axis is denoted by theta\theta. Hence, the slope of the tangent line is equal to tan theta\tan \theta. Through the right angled triangle formed with dsd s as the hypotenuse and dxd x as the base, ds=(dx)/(cos theta)d s=\frac{d x}{\cos \theta}.
Now, the derivative of the curve at x=1x=1 is the slope of the line i.e. tan theta\tan \theta. Now from equation sec^(2)theta=\sec ^{2} \theta=tan^(2)theta+1\tan ^{2} \theta+1 we have
As theta <= (pi) (2),cos theta< asciimath>\theta \leq \frac{\pi}{2}, \cos \theta=> is positive. Now
cos theta=(1)/(sqrt(tan^(2)theta+1))=(1)/(sqrt(((dy)/(dx))^(2)+1))\cos \theta=\frac{1}{\sqrt{\tan ^{2} \theta+1}}=\frac{1}{\sqrt{\left(\frac{d y}{d x}\right)^{2}+1}}
So, ds=sqrt((((dy)/(dx))^(2)+1))dxd s=\sqrt{\left(\left(\frac{d y}{d x}\right)^{2}+1\right)} d x and
{:(2.1)t=int_(0)^(1)(sqrt(1+((dy)/(dx))^(2)))/(sqrt(2g(1-y)))dx:}\begin{equation*}
t=\int_{0}^{1} \frac{\sqrt{1+\left(\frac{d y}{d x}\right)^{2}}}{\sqrt{2 g(1-y)}} d x \tag{2.1}
\end{equation*}
This is the general time formula, which is applicable to any curve y=f(x)y=f(x).
Consider the case of a line passing through the points (1,0)(1,0) and (0,1)(0,1). It is given by y=1-xy=1-x so that (dy)/(dx)=-1\frac{d y}{d x}=-1. Hence, the time taken by an object to roll down the line y=x-1y=x-1 from (1,0)(1,0) to (0,1)(0,1) is equal to
int_(0)^(1)sqrt((1+(-1)^(2))/(2g(1-1+x)))dx=int_(0)^(1)sqrt((2)/(2gx))dx=(2sqrtx)/(sqrtg)|_(0)^(1)=(2)/(sqrtg)=0.638877" seconds "\int_{0}^{1} \sqrt{\frac{1+(-1)^{2}}{2 g(1-1+x)}} d x=\int_{0}^{1} \sqrt{\frac{2}{2 g x}} d x=\left.\frac{2 \sqrt{x}}{\sqrt{g}}\right|_{0} ^{1}=\frac{2}{\sqrt{g}}=0.638877 \text { seconds }
(Taking the value of g as 9.8 ).
Figure 2.1: Graph of y=x^(2)y=x^{2} along with its tangent y=2x-1y=2 x-1 at x=1x=1
3 Time taken for the descent by various conic sections
In this section, we discuss the time taken by an object of any mass to travel from (1,0)(1,0) to (0,1)(0,1) through various conic sections listed below.
Figure 3.1: A graph summarising the curves, which are labelled in the order of increasing descent time.
Consider a parabola of the form y=ax^(2)+bx+c\mathrm{y}=a x^{2}+b x+c passing through the points (1,0)(1,0) and (0,1)(0,1). Let us substitute the coordinates and find the equation of such a parabola. We get:
The general equation for a parabola passing these points is
y=ax^(2)-(a+1)x+1y=a x^{2}-(a+1) x+1
We can now apply the general formula (2.1). First, we differentiate: (dy)/(dx)=2ax-a-1\frac{d y}{d x}=2 a x-a-1. Substituting, we get
t(a)=int_(0)^(1)sqrt((1+(2ax-a-1)^(2))/(2g(1-ax^(2)+(a-1)x-1)))dxt(a)=\int_{0}^{1} \sqrt{\frac{1+(2 a x-a-1)^{2}}{2 g\left(1-a x^{2}+(a-1) x-1\right)}} d x
Note that the time tt is dependent on the value of the parameter aa, and we can try to find the minimum of t(a)t(a) to minimize the descent time.
We calculate the value of the above integral by plotting it into Desmos [5]. As aa is a variable, we add a slider to change the value of aa. The observations are that the value of the integral is the least when a=1a=1 with time taken for descent equal to 0.595227 dots0.595227 \ldots seconds. At a=0.9a=0.9, the value comes out be 0.5958 and at 1.1 , it is 0.59543 . From a=0a=0 to a=1a=1, the time taken decreases and after 1 , it keeps on increasing. The curve is represented by y=x^(2)-2x+1=(x-1)^(2)y=x^{2}-2 x+1=(x-1)^{2}
Now, let us discuss the time taken by a parabola of the form x=ay^(2)+by+cx=a y^{2}+b y+c. Substituting the coordinates in the equation, c=1c=1 and a+b+1=0a+b+1=0 so that b=-1-ab=-1-a. The parabola is of the form x=ay^(2)-(a+1)y+1x=a y^{2}-(a+1) y+1. Differentiating the equation with respect to y,(dx)/(dy)=2ay-a-1y, \frac{d x}{d y}=2 a y-a-1. Our second step is to express yy in terms of xx. For this, we will express xx in terms of a perfectly squared equation, which will help us isolate yy. We have
+-sqrt((4ax-4a+(a+1)^(2))/(4a))=sqrtay-(a+1)/(2sqrta)\pm \sqrt{\frac{4 a x-4 a+(a+1)^{2}}{4 a}}=\sqrt{a} y-\frac{a+1}{2 \sqrt{a}}
so that
y=(+-sqrt((4ax-4a+(a+1)^(2))/(4a))+(a+1)/(2sqrta))/(sqrta)y=\frac{ \pm \sqrt{\frac{4 a x-4 a+(a+1)^{2}}{4 a}}+\frac{a+1}{2 \sqrt{a}}}{\sqrt{a}}
Clearly, the graph of the descent curve cannot pass through the second quadrant. In addition, it has to pass through the points (0,1)(0,1) and (1,0)(1,0). This implies that in the equation above we need to choose the minus sign and restict aa to the interval a in[0,1]a \in[0,1]. So we have
y=(-sqrt((4ax-4a+(a+1)^(2))/(4a))+(a+1)/(2sqrta))/(sqrta)=(a+1-sqrt(4ax-4a+(a+1)^(2)))/(2a)y=\frac{-\sqrt{\frac{4 a x-4 a+(a+1)^{2}}{4 a}}+\frac{a+1}{2 \sqrt{a}}}{\sqrt{a}}=\frac{a+1-\sqrt{4 a x-4 a+(a+1)^{2}}}{2 a}
Now, we substitute the values of (dy)/(dx)\frac{d y}{d x} and yy in the equation (2.1), we obtain an integral depending on aa. The least value of this integral is 0.58377 dots0.58377 \ldots. which occurs at a=0.91\mathrm{a}=0.91.
3.2 Circle
Consider a circle of (x-a)^(2)+(y-b)^(2)=r^(2)(x-a)^{2}+(y-b)^{2}=r^{2} passing through the points (1,0)(1,0) and (0,1)(0,1). Here, aa stands for the xx-coordinate of the centre and bb refers to the yy coordinate of the centre; rr is the radius of the circle. Substituting the points in the equation we get
We use the negative sign because it represents the lower part of the circle, which is used to traverse from (1,0)(1,0) to (0,1)(0,1). As for the parabola case, we substitute (dy)/(dx)\frac{d y}{d x} and yy in equation (2.1). The least value of the integral is 0.58512 dots0.58512 \ldots which appears when a=1.3a=1.3. An important point to consider here is that a >= 1a \geq 1. This is due to the fact that when a < 1a<1< latex>1<> the circle no longer uses its lower part to pass through those points. Instead, the graph will be represented by {y=sqrt(2a^(2)-2a+1-(x-a)^(2))+a:}\left\{y=\sqrt{2 a^{2}-2 a+1-(x-a)^{2}}+a\right., which is the upper part of the circle. Figure 3.2 graphically demonstrates the above idea.
Figure 3.2: Curves represented graphically demonstrating why the negative sign was chosen for positive values of aa
Now, for the case a < 1a<1< latex>1<>, we find the value of the integral by using the corresponding yy value 3.2 . The integral's value keeps on decreasing as x becomes more negative until it reaches a limiting value of 0.6388 dots0.6388 \ldots seconds. This is same as the least time taken by a line as when x rarr-oox \rightarrow-\infty, the part of the circle joining (1,0)(1,0) and (0,1)(0,1) becomes increasingly flatter, eventually taking the shape of a line.
3.3 Ellipse
Consider an ellipse with x -axis as its major axis and x,y\mathrm{x}, \mathrm{y} coordinate of its center as x_(0)x_{0}, 1 . i.e. ((x-x_(0))^(2))/(a^(2))+((y-1)^(2))/(b^(2))=1\frac{\left(x-x_{0}\right)^{2}}{a^{2}}+\frac{(y-1)^{2}}{b^{2}}=1 passing through the points (1,0)(1,0) and (0,1)(0,1). Substituting the coordinates in the equation, we get
. Inserting the value of x_(0)x_{0} as +a+a in the first equation, ((1-a)^(2))/(a^(2))+(1)/(b^(2))=1\frac{(1-a)^{2}}{a^{2}}+\frac{1}{b^{2}}=1 So, (1)/(b^(2))=1-((1-a)^(2))/(a^(2))=(-1+2a)/(a^(2))\frac{1}{b^{2}}=1-\frac{(1-a)^{2}}{a^{2}}=\frac{-1+2 a}{a^{2}}:.b^(2)=(a^(2))/(2a-1)\therefore b^{2}=\frac{a^{2}}{2 a-1}. So , the equation of the ellipse in this case is
We take the value of xx as +a+a as the ellipse is defined only for positive aa values. . Differentiating the above equation with respect to x,2(x-a)+2(2a-1)(y-1)(dy)/(dx)=0x, 2(x-a)+2(2 a-1)(y-1) \frac{d y}{d x}=0
. The next step will be to isolate y. (y-1)^(2)=(a^(2)-(x-a)^(2))/(2a-1)(y-1)^{2}=\frac{a^{2}-(x-a)^{2}}{2 a-1}.
{:(3.2)y=-sqrt((2ax-x^(2))/(2a-1))+1:}\begin{equation*}
y=-\sqrt{\frac{2 a x-x^{2}}{2 a-1}}+1 \tag{3.2}
\end{equation*}
. Due to the same reason in the curves discussed, we choose the negative sign. Plugging the respective values of (dy)/(dx)\frac{d y}{d x} and yy into the time period equation 2.1 , we find that value of this integral varies inconsistently for values of aa. However, gradually, with a huge increase in the value of a, there is a decreasing trend in the value of this integral. Hence, as a rarr ooa \rightarrow \infty, the value of the integral reaches a limiting value. Let's go back to the original equation we derived 3.1. Expanding it, we obtain x^(2)-2ax+a^(2)+(2a-1)(y-1)^(2)=a^(2)x^{2}-2 a x+a^{2}+(2 a-1)(y-1)^{2}=a^{2}. As our aim is to examine the curve near x=0x=0 and x=1,x^(2)x=1, x^{2} can be neglected. :.2ax=(2a-1)(y-1)^(2)\therefore 2 a x=(2 a-1)(y-1)^{2}. Now, 2a=2a-12 a=2 a-1 as we are finding the equation in which a rarr ooa \rightarrow \infty
*(dy)/(dx)=-(1)/(2sqrtx)\cdot \frac{d y}{d x}=-\frac{1}{2 \sqrt{x}}. Hence, the least time taken by an object to reach from (1,0)(1,0) to (0,1)(0,1) through 3.3=3.3= int_(0)^(1)sqrt((1+(1)/(4x))/(2**9.8(1+sqrtx-1)))dx=0.58436 dots\int_{0}^{1} \sqrt{\frac{1+\frac{1}{4 x}}{2 * 9.8(1+\sqrt{x}-1)}} d x=0.58436 \ldots. seconds.
Let us take the case of an ellipse with y axis as its major axis and center as (1,y_(0))\left(1, y_{0}\right). The equation of such an ellipse would be ((y-y_(0))^(2))/(a^(2))+((x-1)^(2))/(b^(2))=1\frac{\left(y-y_{0}\right)^{2}}{a^{2}}+\frac{(x-1)^{2}}{b^{2}}=1. Applying the former strategy,
. Plugging these values in the equation, ((y-a)^(2))/(a^(2))+((x-1)^(2)(2a-1))/(a^(2))=1\frac{(y-a)^{2}}{a^{2}}+\frac{(x-1)^{2}(2 a-1)}{a^{2}}=1 Hence, {(y-a)^(2)+(2a-1)(x-1)^(2)=a^(2):}\left\{(y-a)^{2}+(2 a-1)(x-1)^{2}=a^{2}\right..
Differentiating with respect to x,2(y-a)(dy)/(dx)+2(2a-1)(x-1)=0x, 2(y-a) \frac{d y}{d x}+2(2 a-1)(x-1)=0
. Using the equation 2.1 T to find the time formula, we find that the least time is observed to occur at a=1.375\mathrm{a}=1.375 which is 0.588349 dots0.588349 \ldots seconds. For a > 2.2a>2.2, the value of the integral is seen to have values slightly more than 0.59 , and for a in[1,2.2]a \in[1,2.2], the value decreases until a=1.375a=1.375 and then starts to increase.
Figure 3.3: Graph of ellipse of type 3.1 with a=100\mathrm{a}=100 and parabola x=(y-1)^(2)x=(y-1)^{2}
Figure 3.4: Graph of ellipse of type (3.1) with a =500=500 and parabola x=(y-1)^(2)x=(y-1)^{2}
3.4 Hyperbola
In the previous section, while deriving the equation of the ellipse 3.1 , we equated xx to +a+a because when aa becomes negative, the ellipse takes the form of a hyperbola. Hence, for a hyperbola with x axis as its major axis and center as (x_(0),1)\left(x_{0}, 1\right), the time taken by an object to pass through it can be defined as 3.3 with negative values of aa.
Figure 3.5: Graph of equation 3.1 with positive value of aa
Figure 3.6: Graph of equation 3.1 with negative value of aa
This integral also has a gradually decreasing trend with a large decrease in the value of aa. Hence , as a rarr-ooa \rightarrow-\infty, the integral approaches a limiting value. In the equation 3.2,x^(2)3.2, x^{2} can be neglected as it is drastically smaller than 2ax2 a x. Also, 2a2 a can be equated to 2a-12 a-1 as a rarr-ooa \rightarrow-\infty. Hence, the equation becomes
y=1-sqrtxy=1-\sqrt{x}
This is same as the equation we derived for the ellipse 3.4. Hence, the least time will be same as that of the ellipse with xx axis as major axis i.e. 0.58436 seconds.
Now, for an object to pass through a hyperbola with y axis as major axis and centre at (1,y_(0))\left(1, y_{0}\right) from (1,0)(1,0) to (0,1)(0,1), we will again choose the negative values of aa in the equation for the ellipse 3.3 with y axis as major axis. However, the same curve used for the ellipse won't be applicable to the hyperbola. This is due to the fact that the curve 3.5 which passes through (1,0)(1,0) and (0,1)(0,1) does not through pass through those points for negative values of aa.
Figure 3.7: Graph of equation3.5 for positive values of aa
Figure 3.8: Graph of equation 3.5
for negative values of aa
Instead, the upper part i.e.
will be the equation for the hyperbola. Finding the value of the time integral 2.1 by plugging the values of (dy)/(dx)\frac{d y}{d x} and yy, we find that for all negative values of aa, we find that there is also a decreasing trend in the its value. As, aa approaches -oo-\infty, the integral reaches a limiting value. Rearranging equation 3.6, (y-a)^(2)=a^(2)-(x-1)^(2)(2a-1)rarry^(2)-2ay+a^(2)=a^(2)-(x-1)^(2)(2a-1)(y-a)^{2}=a^{2}-(x-1)^{2}(2 a-1) \rightarrow y^{2}-2 a y+a^{2}=a^{2}-(x-1)^{2}(2 a-1). We want to work with the curve near the points (0,1)(0,1) and (1,0)(1,0).Hence, y^(2)y^{2} can be considered negligible. :.2ay=(x-1)^(2)(2a-1).2 a=2a-1\therefore 2 a y=(x-1)^{2}(2 a-1) .2 a=2 a-1 as a rarr-ooa \rightarrow-\infty. Hence, the curve will be y=(x-1)^(2)*(dy)/(dx)=2(x-1)y=(x-1)^{2} \cdot \frac{d y}{d x}=2(x-1). So, the time integral ==
int_(0)^(1)sqrt((1+4(x-1)^(2))/(2**9.8(1-(x-1)^(2))))dx\int_{0}^{1} \sqrt{\frac{1+4(x-1)^{2}}{2 * 9.8\left(1-(x-1)^{2}\right)}} d x
The least time taken by an object of any mass to travel from the point (1,0)(1,0) to (0,1)(0,1) through hyperbola 3.6 , =0.595227=0.595227 seconds.
4 Conclusion
Comparing the time taken by an object to pass through a line and various conic sections, we observe that a parabola of the form x=f(y)x=f(y) is the curve with the quickest descent. However, we have restricted the number of parameters in the cases of ellipse and hyperbola by assuming them to be centered at (x_(0),1)\left(x_{0}, 1\right) and (1,y_(0))\left(1, y_{0}\right) for those with xx and yy axes as major axis respectively. This problem can be applied in fields such as architecture and engineering, so as to design a quicker route. Further studies could be done to validate this paper's findings and expand upon the research by finding the least time to traverse any ellipse and hyperbola. This Additionally, using the parametric equations of a general cycloid, a similar method can be employed to prove why the cycloid is the curve of quickest descent.
References
[1] Mohammed Lufti, H. E. "The Effect of Gravitational Field on Brachistochrone Problem " . Phys: Conf. Ser. 1028012060
[2] Kurilin, Alexander"Brachistochrone and Sliding with Friction". Reseach Gate : 10.48550/arXiv. 2305.18345
[3] Esam H. Abdul-Hafidh "A new approach to solve the Brachistochrone problem by constructing a lattice unit cell" Heliyon, Volume 8, Issue 12, 2022, e11994, ISSN 2405-8440
[4] "Brachistochrone curve", Wikipedia, 23 June 2024
[5] Desmos, Amherst, 2011