Recover the unknown function by undoing a logarithm

Functions f:x↦1+2x−1 f:x\mapsto1+\frac2{x-1}, x∈Rx\in\mathbb R, x<1 x<1; g:x↦ln⁡xg:x\mapsto\ln x, x∈Rx\in\mathbb R, 0<x<10<x<1.

(i) Explain why composite gfgf does not exist. mark

(ii) Find an expression for fg(x) fg(x). Hence or otherwise find (fg)−1(0)(fg)^{-1}(0). marks

(iii) Find h(x) h(x) for each case: (a) gh(x)=xgh(x)=x, mark (b) hg(x)=x2+1 hg(x)=x^2+1. marks

[2018/DHS/Prelim/I/5]

Total marks: 8

Solution

(i)

Existence of composites

f(x)=1+2x−1f(x)=1+\dfrac2{x-1} for x<1x<1 and g(x)=ln⁡xg(x)=\ln x for 0<x<10<x<1. Explain why gfgf does not exist.

Explore this idea

Hint

  1. gfgf exists only if Rf⊆DgR_f\subseteq D_g.

  2. Find RfR_f from the graph: horizontal asymptote y=1y=1 and vertical asymptote x=1x=1.

Answer · (i)

gfgf does not exist: Rf=(−∞,1)⊈(0,1)=DgR_f=(-\infty,1)\nsubseteq(0,1)=D_g.

Explore the idea

f's outputs spill outside the domain of g

Drag x. Whenever f(x) ≤ 0, it is not an allowed input of g = ln on (0, 1).

Input x (in the domain of f)x = 0.00
Try an example
input
0
↓ f
f(x)
-1
↓ g
gf(x)
undefined

-1 is outside the domain of g, so gf(x) does not exist for this x.

D_gR_f-4-12

D_g: (0, 1); R_f: (−∞, 1). Finite bounds are shown to 6 significant figures. Hollow ends are excluded; arrows continue beyond this window.

Selected output: -1.

The composite does not exist as defined

Rf=(−∞,1)R_f=(-\infty,1) is not inside Dg=(0,1)D_g=(0,1): all the negative outputs are the problem.

  • F(x)
  • Gf(x), where it exists
  • Lower end of D_g
Key idea

A single bad output is enough: Rf⊈DgR_f\nsubseteq D_g, so gfgf does not exist.

Ideas to try
  1. Drag xx to −1-1: what is f(−1)f(-1)? Is it allowed?
  2. For which xx is f(x)f(x) allowed in gg?
(ii) composite

Defining composites

Find an expression for fg(x)fg(x).

Hint

  1. Substitute g(x)=ln⁡xg(x)=\ln x into ff and combine the fractions.

Answer · (ii) composite

fg(x)=(ln⁡x+1)/(ln⁡x−1)fg(x)=(\ln x+1)/(\ln x-1), 0<x<10<x<1.

(ii) inverse value

Inverse values

Hence or otherwise find (fg)−1(0)(fg)^{-1}(0).

Explore this idea

Hint

  1. (fg)−1(0)(fg)^{-1}(0) is the input xx with fg(x)=0fg(x)=0.

Answer · (ii) inverse value

(fg)−1(0)=e−1(fg)^{-1}(0)=e^{-1}.

Explore the idea

Which input gives fg(x) = 0?

fg is strictly decreasing on (0, 1) from 1 to −1, so each height is hit once, and k = 0 is hit at x = e⁻¹.

Height of the horizontal linek = 0
Try an example
Test boundary heights

Buttons use full-precision boundary values; numbers shown here are rounded.

Number of distinct intersections
1

Between −1 and 1 there is exactly one intersection. The function is one-one.

All listed intersections are within the original drawing window. A touching point counts once; hollow endpoints are excluded.

Coordinates are rounded to 6 significant figures. Inspect a row to highlight that intersection and bring it into view.
Pointxy
0.3678790
  • Fg
  • Y = k
  • Input x
  • Intersection selected in the coordinate table
Key idea

(fg)−1(0)(fg)^{-1}(0) means 'which input gives output 0?'. Reading off the graph or solving fg(x)=0fg(x)=0 gives the same answer, e−1e^{-1}.

Ideas to try
Set k=0k=0 and read the intersection: is it near e−1≈0.37e^{-1}\approx0.37?
(iii)(a)

Recovering a function

Find h(x)h(x) such that gh(x)=xgh(x)=x.

Explore this idea

Hint

  1. gh=identitygh=\text{identity} means hh undoes gg: h=g−1h=g^{-1}.

Answer · (iii)(a)

h(x)=exh(x)=e^x, x<0x<0.

Explore the idea

ln x on (0, 1) and its inverse eˣ on x < 0

gh(x) = x means h is the mirror image of g in y = x.

Point on f: input xx = 0.40
Reset the controls

Reflection swaps the coordinates. The segment PQ is perpendicular to y = x, and its blue midpoint M lies on that line.

P = (x, f(x))
(0.4, -0.91629)
Q = (f(x), x)
(-0.91629, 0.4)
on the inverse
Domain of f
0<x<10<x<1
Range of f = domain of f⁻¹
y<0y<0
Midpoint M on y = x
(-0.25815, -0.25815)
  • G(x) = ln x on (0, 1)
  • H(x) = eˣ for x < 0
  • Y = x
  • PQ is perpendicular to y = x; its blue midpoint M lies on that line
Key idea

h=g−1h=g^{-1}: domain and range swap, so hh lives on x<0x<0 and takes values in (0,1)(0,1).

Ideas to try
Why does hh only exist for x<0x<0?
(iii)(b)

Recovering a function

Find h(x)h(x) such that hg(x)=x2+1hg(x)=x^2+1.

Explore this idea

Hint

  1. hg(x)=x2+1hg(x)=x^2+1 gives hh evaluated at ln⁡x\ln x.

  2. Replace xx by g−1(x)=exg^{-1}(x)=e^x.

Answer · (iii)(b)

h(x)=e2x+1h(x)=e^{2x}+1, x<0x<0.

Values outside this interval are not fixed by the stated condition.

Explore the idea

h(x) = eˣ and h(x) = e²ˣ + 1 on x < 0

Part (a) gives h = eˣ (from 0 up to 1). Part (b) gives h = e²ˣ + 1 (from 1 up to 2).

Switch on a feature to add it to the sketch and see how it is found. A full-mark sketch shows all of them.

  • (a)
  • (b)
Ideas to try
Check: hg(x)hg(x) for (b) at x=12x=\frac12: h(ln⁡12)=e2ln⁡12+1=14+1h(\ln\frac12)=e^{2\ln\frac12}+1=\frac14+1, which equals x2+1x^2+1.