OCR H240/01 2023 June — Question 10 9 marks

Exam BoardOCR
ModuleH240/01 (Pure Mathematics)
Year2023
SessionJune
Marks9
PaperDownload PDF ↗
TopicSign Change & Interval Methods
TypeExplain Sign Change Method Failure
DifficultyStandard +0.3 This is a multi-part question on numerical methods that is slightly easier than average. Part (a) requires recognizing that a discontinuity prevents sign change method from working—a standard conceptual point. Parts (b) and (c) involve routine algebraic rearrangement and iteration with a calculator. Part (d) tests understanding that |F'(x)| > 1 causes divergence, which is a standard A-level criterion. While it covers multiple techniques, each part is straightforward application of taught material without requiring novel insight or complex problem-solving.
Spec1.09a Sign change methods: locate roots1.09c Simple iterative methods: x_{n+1} = g(x_n), cobweb and staircase diagrams1.09e Iterative method failure: convergence conditions

10
The diagram shows part of the curve \(\mathrm { f } ( x ) = \frac { \mathrm { e } ^ { x } } { 4 x ^ { 2 } - 1 } + 2\). The equation \(\mathrm { f } ( x ) = 0\) has a positive root \(\alpha\) close to \(x = 0.3\).
  1. Explain why using the sign change method with \(x = 0\) and \(x = 1\) will fail to locate \(\alpha\).
  2. Show that the equation \(\mathrm { f } ( x ) = 0\) can be written as \(x = \frac { 1 } { 4 } \sqrt { \left( 4 - 2 \mathrm { e } ^ { x } \right) }\).
  3. Use the iterative formula \(x _ { n + 1 } = \frac { 1 } { 4 } \sqrt { \left( 4 - 2 \mathrm { e } ^ { x _ { n } } \right) }\) with a starting value of \(x _ { 1 } = 0.3\) to find the value of \(\alpha\) correct to \(\mathbf { 4 }\) significant figures, showing the result of each iteration.
  4. An alternative iterative formula is \(x _ { n + 1 } = \mathrm { F } \left( x _ { n } \right)\), where \(\mathrm { F } \left( x _ { n } \right) = \ln \left( 2 - 8 x _ { n } ^ { 2 } \right)\). By considering \(\mathrm { F } ^ { \prime } ( 0.3 )\) explain why this iterative formula will not find \(\alpha\).

Question 10:
Part (a):
AnswerMarks Guidance
AnswerMarks Guidance
Both \(f(0)\) and \(f(1)\) are positive so no sign change will be seenB1 Identify both \(y\)-values being positive and state 'no sign change'. Could also refer to asymptote/discontinuity within range (\(x=0\) to \(x=1\)). B0 for no reference to interval
Part (b):
AnswerMarks Guidance
AnswerMarks Guidance
\(\frac{e^x}{4x^2-1} = -2\); \(e^x = -8x^2+2\); \(8x^2 = 2-e^x\)M1 Attempt rearrangement, as far as \(kx^2 = \ldots\) Allow sign errors only
\(16x^2 = 4-2e^x\); \(4x = \sqrt{4-2e^x}\); \(x = \frac{1}{4}\sqrt{4-2e^x}\) A.G.A1 Obtain given answer convincingly. If \(x = \sqrt{\frac{1}{4}-\frac{1}{8}e^x}\) then additional line of working needed
Part (c):
AnswerMarks Guidance
AnswerMarks Guidance
\(x_2 = 0.285074813\ldots\)B1 Correct first iterate (at least 4sf). State 0.2851 or better
\(0.28943, 0.28817, 0.28853, 0.28843, 0.28846, 0.28845\ldots\)M1 Correct iterative process (at least 3 more values). Allow M1 for 3sf
\(\alpha = 0.2885\)A1 Correct root to 4sf, following 2 iterates that agree to 4sf. At least 7 iterations needed
Part (d):
AnswerMarks Guidance
AnswerMarks Guidance
\(F'(x) = \frac{-16x}{2-8x^2}\)M1 Attempt differentiation using chain rule. Obtain derivative of form \(\frac{kx}{2-8x^2}\)
\(F'(0.3) = -3.75\)M1 Attempt \(F'(0.3)\) — not dependent on previous M1, but must follow some attempt at differentiation
For convergence \(\F'(\alpha)\ < 1\), but \(-3.75 < -1\), so iteration will not find root
# Question 10:

## Part (a):
| Answer | Marks | Guidance |
|--------|-------|----------|
| Both $f(0)$ and $f(1)$ are positive so no sign change will be seen | **B1** | Identify both $y$-values being positive and state 'no sign change'. Could also refer to asymptote/discontinuity within range ($x=0$ to $x=1$). B0 for no reference to interval |

## Part (b):
| Answer | Marks | Guidance |
|--------|-------|----------|
| $\frac{e^x}{4x^2-1} = -2$; $e^x = -8x^2+2$; $8x^2 = 2-e^x$ | **M1** | Attempt rearrangement, as far as $kx^2 = \ldots$ Allow sign errors only |
| $16x^2 = 4-2e^x$; $4x = \sqrt{4-2e^x}$; $x = \frac{1}{4}\sqrt{4-2e^x}$ **A.G.** | **A1** | Obtain given answer convincingly. If $x = \sqrt{\frac{1}{4}-\frac{1}{8}e^x}$ then additional line of working needed |

## Part (c):
| Answer | Marks | Guidance |
|--------|-------|----------|
| $x_2 = 0.285074813\ldots$ | **B1** | Correct first iterate (at least 4sf). State 0.2851 or better |
| $0.28943, 0.28817, 0.28853, 0.28843, 0.28846, 0.28845\ldots$ | **M1** | Correct iterative process (at least 3 more values). Allow M1 for 3sf |
| $\alpha = 0.2885$ | **A1** | Correct root to 4sf, following 2 iterates that agree to 4sf. At least 7 iterations needed |

## Part (d):
| Answer | Marks | Guidance |
|--------|-------|----------|
| $F'(x) = \frac{-16x}{2-8x^2}$ | **M1** | Attempt differentiation using chain rule. Obtain derivative of form $\frac{kx}{2-8x^2}$ |
| $F'(0.3) = -3.75$ | **M1** | Attempt $F'(0.3)$ — not dependent on previous M1, but must follow some attempt at differentiation |
| For convergence $\|F'(\alpha)\| < 1$, but $-3.75 < -1$, so iteration will not find root | **A1** | Correct reasoning following correct $F'(0.3)$. Allow $F'(\alpha) < -1$, hence will not converge. No credit for just testing the given iterative formula |

---
10\\
\begin{tikzpicture}[x=1.8cm, y=1.2cm, >=latex]

    % Draw axes
    \draw[->, semithick] (-0.5, 0) -- (4.4, 0) node[right] {$x$};
    \draw[->, semithick] (0, -2.6) -- (0, 6) node[left] {f($x$)};

    % Draw x-axis ticks
    \foreach \x in {1, 2, 3, 4} {
        \draw[semithick] (\x, 0) -- (\x, -0.08) node[below] {\x};
    }
    % Origin label
    \node[below] at (0, -0.08) {0};

    % Draw the function curve
    \begin{scope}
        % Clip the area to cleanly cut off the lines heading towards infinity
        \clip (-0.5, -2.6) rectangle (4.2, 6);
        
        % Left branch of the curve (before the x = 0.5 asymptote)
        \draw[thick, samples=200, domain=-0.35:0.47] plot (\x, {exp(\x)/(4*\x*\x - 1) + 2});
        
        % Right branch of the curve (after the x = 0.5 asymptote)
        \draw[thick, samples=200, domain=0.53:4.2] plot (\x, {exp(\x)/(4*\x*\x - 1) + 2});
    \end{scope}

\end{tikzpicture}

The diagram shows part of the curve $\mathrm { f } ( x ) = \frac { \mathrm { e } ^ { x } } { 4 x ^ { 2 } - 1 } + 2$. The equation $\mathrm { f } ( x ) = 0$ has a positive root $\alpha$ close to $x = 0.3$.
\begin{enumerate}[label=(\alph*)]
\item Explain why using the sign change method with $x = 0$ and $x = 1$ will fail to locate $\alpha$.
\item Show that the equation $\mathrm { f } ( x ) = 0$ can be written as $x = \frac { 1 } { 4 } \sqrt { \left( 4 - 2 \mathrm { e } ^ { x } \right) }$.
\item Use the iterative formula $x _ { n + 1 } = \frac { 1 } { 4 } \sqrt { \left( 4 - 2 \mathrm { e } ^ { x _ { n } } \right) }$ with a starting value of $x _ { 1 } = 0.3$ to find the value of $\alpha$ correct to $\mathbf { 4 }$ significant figures, showing the result of each iteration.
\item An alternative iterative formula is $x _ { n + 1 } = \mathrm { F } \left( x _ { n } \right)$, where $\mathrm { F } \left( x _ { n } \right) = \ln \left( 2 - 8 x _ { n } ^ { 2 } \right)$.

By considering $\mathrm { F } ^ { \prime } ( 0.3 )$ explain why this iterative formula will not find $\alpha$.
\end{enumerate}

\hfill \mbox{\textit{OCR H240/01 2023 Q10 [9]}}