Porisma

Worked solution

Choose a valid substitution for a square-root estimate

Expand 1−x/(1+2x)\sqrt{1-x}/(1+2x) through x2x^2. marks

(i) A student substitutes x=−1x=-1 to estimate 2\sqrt2. Explain why this is wrong. mark

(ii) Put x=1/9x=1/9 in the expansion to show 2≈p/1296\sqrt2\approx p/1296, and find the positive integer pp. marks

[2001/RJC/I/16(a)]

CJC · Review problems · Question 3

Answer

1−5x/2+39x2/8+O(x3)1-5x/2+39x^2/8+O(x^3); x=−1x=-1 is invalid; p=1859p=1859.

Multiply both expansions through quadratic order

Use (1−x)1/2=1−x/2−x2/8+⋯(1-x)^{1/2}=1-x/2-x^2/8+\cdots and (1+2x)−1=1−2x+4x2+⋯(1+2x)^{-1}=1-2x+4x^2+\cdots. Thus

1−x1+2x=1−52x+398x2+O(x3).\begin{aligned}&\frac{\sqrt{1-x}}{1+2x}\\ & =1-\frac52x+\frac{39}{8}x^2+O(x^3).\end{aligned}

How did this part go?

(i) Intersect the convergence conditions

The two source binomial/geometric conditions are ∣x∣<1|x|<1 and ∣2x∣<1|2x|<1, so −1/2<x<1/2-1/2<x<1/2. The proposed x=−1x=-1 is outside this interval and cannot justify the estimate.

How did this part go?

(ii) Isolate the required radical after substitution

At x=1/9x=1/9, the function is 8/9/(11/9)=62/11\sqrt{8/9}/(11/9)=6\sqrt2/11, while the quadratic is 1−5/18+39/648=169/2161-5/18+39/648=169/216. Hence 62/11≈169/2166\sqrt2/11\approx169/216, giving 2≈1859/1296\sqrt2\approx1859/1296 and p=1859p=1859. This is the printed low-order estimate, not an exact identity.

How did this part go?

Common mistakes

A real value of the original expression does not guarantee convergence of its Maclaurin series there.

Read the hint