Find all of the fixed points of the function \(f(x) = x^2 - 2\).
Set \(f(\bar{x}) = \bar{x}\) and solve. \[\begin{align*} \rightarrow \bar{x}^2 - 2 &= \bar{x}\\ \bar{x}^2 - \bar{x} - 2 &= 0\\ (\bar{x}+1)(\bar{x}-2) &= 0\end{align*}\] This has two solutions: \(\bar{x}_1 = -1\) and \(\bar{x}_2 = 2\), both of which must be fixed points of \(f(x) = x^2 - 2\).
Let’s check!
\(f(-1) = (-1)^2 - 2 = 1 - 2 =
-1\)
\(f(2) = 2^2 - 2 = 4 - 2 = 2\)
Given that \(f(x) = \frac{1}{x}\), find the periodic points of period two of \(f(x)\).
Hint: You may want to find the fixed points of \(f(x)\) first.
First, find the fixed points of \(f(x)\) by solving \(f(\bar{x}) = \bar{x}\). \[\begin{align*} \frac{1}{\bar{x}} &= \bar{x}\\ \bar{x}^2 = 1\end{align*}\] This has solutions \(\bar{x}_1 = -1\) and \(\bar{x}_2 = 1\). so these must be our fixed points.
Now we want to solve for the fixed points of \(f^{[2]}(x)\). \[\begin{align*} f^{[2]}(x) &= f(f(x))\\ &= \frac{1}{1/x}\\ &= x\end{align*}\] Setting \(f^{[2]}(\bar{x})=\bar{x}\) gives \(\bar{x} = \bar{x}\). This is true for all values of \(\bar{x}\). Does this mean that all values of \(\bar{x}\) are periodic points of period two of \(f(x)\)? Almost!
Since the points \(\bar{x}_1 = -1\) and \(\bar{x}_2 = 1\) are fixed points of \(f(x)\), they cannot also be periodic points of period two. We also need to be careful and consider the domain of \(f(x)\), which excludes the point \(x=0\) (since \(f(x) = \frac{1}{x}\) is undefined when \(x=0\)). This means that \(x=0\) cannot be a periodic point of \(f(x)\). What we are left with is that all \(x\in\mathbb{R}\) except for \(x=-1, 1\), and \(0\) are periodic points of period two of \(f(x)\). We could write the set of periodic points of period two of \(f(x)\) as \(\left\{x\in\mathbb{R}| x\neq -1,1,0\right\}\).
Consider the function \(f(x) = x^2\). Sketch this function and plot the first few points of its orbit \(\left\{ x_0,x_1,x_2,x_3,\ldots\right\}\), i.e. plot the points \(\left(x_0,x_1 = f(x_0)\right)\), \(\left(x_1,x_2 = f(x_1)\right)\), etc..., for the starting values \(x_0 = 0, \frac{1}{2}\), and \(2\). Describe what is happening to the orbit of \(f(x)\) for each of these starting values.
Solution:
\(x_0 = 0\):
The orbit of \(x_0 = 0\) under \(f(x) = x^2\) is \(\left\{0,0,0,0,\ldots\right\}\). The point \(x_0 = 0\) is a fixed point of \(f(x)\).
\(x_0 = \frac{1}{2}\):
The orbit of \(x_0 = \frac{1}{2}\) under \(f(x) = x^2\) is \(\left\{\frac{1}{2},\frac{1}{4},\frac{1}{16},\frac{1}{256},\ldots\right\}\). The iterates of \(x_0 = \frac{1}{2}\) are getting smaller in magnitude on each iteration and approaching zero.
\(x_0 = 2\):
The orbit of \(x_0 = 2\) under \(f(x) = x^2\) is \(\left\{2,4,16,256,\ldots\right\}\). The iterates of \(x_0 = 2\) are getting larger in magnitude on each iteration and approaching infinity.
Let \(f(x) = x^2 + 3x + 1\). Find all of the fixed points of \(f(x)\).
Solution: To find the fixed points we need to solve \(f(\bar{x}) = \bar{x}\). \[\begin{align*} \bar{x}^2 + 3\bar{x} + 1 &= \bar{x}\\ \bar{x}^2 + 2\bar{x} + 1 &= 0\\ (\bar{x}+1)^2 &= 0\end{align*}\] This has one solution, \(\bar{x} = -1\), so \(f(x)\) has just one fixed point at \(\bar{x} = -1\).
Consider the family of functions defined by \(f_c(x) = cx\) where \(c\) is a constant and \(c\neq 0\). Determine all of the fixed points of \(f_c(x)\).
Hint: You may end up with different fixed points depending on the value of \(c\).
Solution: To find the fixed points of \(f_c(x)\) we solve \(f_c(\bar{x}) = \bar{x}\), treating \(c\) as a constant. \[\begin{align*} c\bar{x} &= \bar{x}\\ (c-1)\bar{x} &= 0\end{align*}\] For most values of \(c\), this has one solution, \(\bar{x} = 0\), however we can see that when \(c=1\), then all \(\bar{x}\in\mathbb{R}\) are solutions. Thus, the fixed points of \(f_c(x)\) are \(\bar{x} = 0\), \(c\neq 1\) and \(\bar{x}\in\mathbb{R}\), \(c=1\). Notice that when \(c=1\), \(f_1(x) = x\), for which all \(x\in\mathbb{R}\) are clearly fixed points.
Consider the function \(f(x) = x^2 - \frac{1}{2}\). Sketch \(f(x)\) and \(y = x\) on the same set of axes and show graphically that \(f(x)\) has two fixed points. Label these fixed points on your sketch as \(\bar{x}_1\) and \(\bar{x}_2\) such that \(\bar{x}_1 < \bar{x}_2\).
Use a graphical method (i.e. cobweb diagram) to help determine the behaviour of various orbits starting near both \(\bar{x}_1\) and \(\bar{x}_2\). Use your diagram to make an educated guess as to the nature (attractive, repelling, or neither) of each fixed point.
Now consider the family of functions \(f_c(x) = x^2 + c\) where \(c\) is a constant. For what values of \(c\) do fixed points of \(f_c(x)\) exist? Some sketches of the graphs of \(f_c(x)\) for various values of \(c\) may help, but they are not necessary.
Solution:
From the graph, we can see that \(f(x)\) intersects the line \(y=x\) twice, and thus has two fixed points.
Cobweb diagram:
From our cobweb diagram we see that the iterates of \(f(x)\) are attracted towards the fixed point \(\bar{x}_1\), so we can guess that this is an attractive fixed point. On the other hand, the iterates of \(f(x)\) move away from the fixed point \(\bar{x}_2\), so this is likely to be a repelling fixed point.
To find the fixed points of \(f_c(x)\) we need to solve \(f_c(\bar{x}) = \bar{x}\). \[\begin{align*} \bar{x}^2 + c &= \bar{x}\\ \bar{x}^2 - \bar{x} + c &= 0\end{align*}\] Using the quadratic formula, this has solutions \(\bar{x} = \frac{1}{2} \pm \frac{1}{2}\sqrt{1 - 4c}\). This has (real) solutions only when the argument of the square root is greater than (or equal to) zero, i.e. \[\begin{align*} 1 - 4c &\geq 0\\ c &\leq \frac{1}{4}.\end{align*}\] Thus, \(f_c(x)\) has fixed points when \(c\leq \frac{1}{4}\). Notice that when \(c = \frac{1}{4}\), \(f_c(x)\) has one fixed point (only one solution to the quadratic formula) and when \(c < \frac{1}{4}\), \(f_c(x)\) has two fixed points (two solutions to the quadratic formula).
Let \(f(x) = -x^3\). Find all fixed points and periodic points of period two of \(f(x)\).
Solution: First, let’s find any fixed points by solving \(f(\bar{x}) = \bar{x}\). \[\begin{align*} -\bar{x}^3 &= \bar{x}\\ \bar{x} + \bar{x}^3 &= 0\\ \bar{x}(1 + \bar{x}^2) & = 0\\\end{align*}\] This has just one solution, \(\bar{x} = 0\), so \(f(x)\) has one fixed point at \(\bar{x} = 0\).
Now, let’s find any periodic points of period two. We need to solve \(f^{[2]}(\bar{x}) = \bar{x}\). \[\begin{align*} f^{[2]}(x) &= f(f(x))\\ &= -\left(-x^3\right)^3\\ &= x^9\end{align*}\] So we need to solve \(\bar{x}^9 = \bar{x}\). Rearranging, this gives \(\bar{x}(\bar{x}^8-1) = 0\), which has three solutions \(\bar{x} = 0, -1,\) and \(1\). Since \(\bar{x}=0\) is a fixed point, the remaining two points must form a two cycle. Thus, \(\bar{x} = -1\) and \(1\) are the points of period two of \(f(x)\).
CHALLENGE Let \(f(x) = 1 - x^2\). Find all fixed points and periodic points of period two of \(f(x)\).
Solution: First, let’s find the fixed points. \[\begin{align*} 1 - \bar{x}^2 &= \bar{x}\\ \bar{x}^2 + \bar{x} - 1 &= 0\\\end{align*}\] Using the quadratic formula, this gives two solutions, \(\bar{x} = \frac{-1}{2}\pm\frac{\sqrt{5}}{2}\), which are the two fixed points of \(f(x)\).
Next, we want to solve for any periodic points of period two. First we find \(f^{[2]}(x)\), \[\begin{align*} f^{[2]}(x) &= 1 - \left(1-x^2\right)^2\\ &= 1 - 1 + 2x^2 - x^4\\ &= 2x^2 - x^4\end{align*}\] and then solve \(f^{[2]}(\bar{x}) = \bar{x}\). \[\begin{align*} 2\bar{x}^2 - \bar{x}^4 &= \bar{x}\\ \bar{x}^4 - 2\bar{x}^2 + \bar{x} &= 0\\ \bar{x}\left(\bar{x}^3 - 2\bar{x} + 1\right) &= 0\end{align*}\] Factoring the cubic part of this expression could be difficult, but luckily we know that the fixed points of \(f(x)\) are also solutions to \(f^{[2]}(\bar{x}) = \bar{x}\). This means that \((\bar{x}^2 + \bar{x} - 1)\) must be a factor. Thus we have \[\begin{align*} \bar{x}\left(\bar{x}^2 + \bar{x} - 1\right)\left(\bar{x}-1\right) &= 0\end{align*}\] The two new solutions (which are not fixed points of the original function) are \(\bar{x} = 0\) and \(1\). Thus we must have the two cycle \(\left\{0,1\right\}\).
CHALLENGE Consider the function \(f(x) = x + \cos(x)\). Show that \(f(x)\) has an infinite number of fixed points.
Solution: To find the fixed points of \(f(x)\) we must solve \[\begin{align*} \bar{x} + \cos\left(\bar{x}\right) &= \bar{x}\\ \cos\left(\bar{x}\right) &= 0.\end{align*}\] Considering the unit circle (or a graph of \(\cos(x)\)), we see that this is true when \(\bar{x}\) is an odd multiple of \(90\) degrees, i.e. \(\bar{x} = 90\degree, 270\degree, 450\degree, -90\degree,\ldots\)
We can write this as \(\bar{x}\) is a fixed point of \(f(x)\) when \(\bar{x} = (2k+1)\cdot 90\degree\) for any integer \(k\). Since the set of all integers is an infinite set, this results in an infinite number of fixed points for \(f(x)\).