Porisma

Worked solution

A self-inverse map and a restricted modulus inverse

A function ff is said to be self-inverse if f(x)=f−1(x)f(x)=f^{-1}(x) for all xx in its domain. The functions are defined by f:x↦7−3x3−xf:x\mapsto\dfrac{7-3x}{3-x}, x∈Rx\in\mathbb R, x≠3x\ne3, and g:x↦∣(2−x)(1+x)∣g:x\mapsto|(2-x)(1+x)|, x∈Rx\in\mathbb R, x≤−1x\le-1. (i) Explain why f−1f^{-1} exists and show that ff is self-inverse. Hence, or otherwise, evaluate f2003(5)f^{2003}(5). (ii) Find an expression for g−1(x)g^{-1}(x). (iii) Sketch gg and g−1g^{-1} on the same diagram, illustrating their relationship and labelling axial intercepts, if any. Write down the set of values of xx satisfying gg−1(x)=xgg^{-1}(x)=x. (iv) Show that fg−1fg^{-1} exists and find its exact range.

2017/NJC/II/2

CJC · Functions · Example 13

(i) Establish the inverse and self-inverse property

Answer

f−1(x)=7−3x3−xf^{-1}(x)=\dfrac{7-3x}{3-x} on R∖{3}\mathbb R\setminus\{3\}, so f=f−1f=f^{-1}.

The source graph has asymptotes x=3x=3, y=3y=3, and axial intercepts (7/3,0)(7/3,0) and (0,7/3)(0,7/3). Each horizontal line meets it at most once: rewriting f(x)=3+2/(x−3)f(x)=3+2/(x-3) shows that the left branch has outputs below 3, the right branch outputs above 3, and each branch decreases strictly. Thus ff is one-to-one. Let y=f(x)y=f(x), so f−1(y)=xf^{-1}(y)=x. Then (3−x)y=7−3x(3-x)y=7-3x, 3y−xy=7−3x3y-xy=7-3x, x(3−y)=7−3yx(3-y)=7-3y, and x=(7−3y)/(3−y)x=(7-3y)/(3-y). Rename the inverse input to obtain the stated rule. Its domain is Rf=(−∞,3)∪(3,∞)=DfR_f=(-\infty,3)\cup(3,\infty)=D_f. Rule and domain agree, proving self-inverse behaviour.

Expand diagram

Use the scrollbars or arrow keys to inspect details.

Rational f=3+2/(x−3), asymptotes x3,y3, intercepts (0,7/3),(7/3,0). Left and right branch ranges are disjoint, each horizontal line meets at most once.
Rational f=3+2/(x−3), asymptotes x3,y3, intercepts (0,7/3),(7/3,0). Left and right branch ranges are disjoint, each horizontal line meets at most once.
Rational f=3+2/(x−3), asymptotes x3,y3, intercepts (0,7/3),(7/3,0). Left and right branch ranges are disjoint, each horizontal line meets at most once.
How did this part go?

(i) Evaluate the 2003rd iterate

Answer

f2003(5)=4f^{2003}(5)=4.

Since f=f−1f=f^{-1}, ff=ff−1ff=ff^{-1} gives f2(x)=xf^2(x)=x. Then ff2(x)=f(x)ff^2(x)=f(x) gives f3(x)=f(x)f^3(x)=f(x), and ff3(x)=ff(x)ff^3(x)=ff(x) gives f4(x)=xf^4(x)=x. Continuing, f2002(x)=xf^{2002}(x)=x, so f2003(x)=ff2002(x)=f(x)f^{2003}(x)=ff^{2002}(x)=f(x). Hence f2003(5)=f(5)=(7−3(5))/(3−5)=4f^{2003}(5)=f(5)=(7-3(5))/(3-5)=4. The superscript counts compositions; it is not a pointwise power. Independently 5 and 4 form a two-cycle and neither hits the excluded input 3.

How did this part go?

(ii) Select and invert the permitted branch

Answer

g−1(x)=12−x+94g^{-1}(x)=\dfrac12-\sqrt{x+\dfrac94}, x≥0x\ge0.

Choose the equation before rearranging, as the Teacher note stresses. On x≤−1x\le-1, (2−x)(1+x)≤0(2-x)(1+x)\le0, so y=−(2−x)(1+x)=−(2+x−x2)=x2−x−2y=-(2-x)(1+x)=-(2+x-x^2)=x^2-x-2. The unrestricted modulus illustration has roots −1,2 and central maximum (1/2,9/4)(1/2,9/4); only its left branch belongs to gg. Set y=g(x)y=g(x), so g−1(y)=xg^{-1}(y)=x. Complete the square: y=[(x−1/2)2−1/4]−2y=[(x-1/2)^2-1/4]-2, hence (x−1/2)2=y+9/4(x-1/2)^2=y+9/4, x−1/2=±y+9/4x-1/2=\pm\sqrt{y+9/4}, and x=1/2±y+9/4x=1/2\pm\sqrt{y+9/4}. Reject the positive branch because the original inputs satisfy x≤−1x\le-1. Thus g−1(y)=1/2−y+9/4g^{-1}(y)=1/2-\sqrt{y+9/4}; rename its input to xx. The original left branch has range [0,∞)[0,\infty), so that is the inverse domain, not merely the larger algebraic radical domain [−9/4,∞)[-9/4,\infty).

Expand diagram

Use the scrollbars or arrow keys to inspect details.

Unrestricted absolute(2−x)(1+x) has roots −1,2, y-intercept 2, central maximum (1/2,9/4); negative original tails dashed. g retains only the left branch x≤−1.
Unrestricted absolute(2−x)(1+x) has roots −1,2, y-intercept 2, central maximum (1/2,9/4); negative original tails dashed. g retains only the left branch x≤−1.
Unrestricted absolute(2−x)(1+x) has roots −1,2, y-intercept 2, central maximum (1/2,9/4); negative original tails dashed. g retains only the left branch x≤−1.
How did this part go?

(iii) Sketch the inverse relationship

Answer

Reflect in y=xy=x: gg has intercept (−1,0)(-1,0); g−1g^{-1} has intercept (0,−1)(0,-1).

Sketch g(x)=x2−x−2g(x)=x^2-x-2 only for x≤−1x\le-1, with closed endpoint (−1,0)(-1,0) and no y-intercept. Reflect every point across y=xy=x to obtain g−1g^{-1} on x≥0x\ge0, starting at closed (0,−1)(0,-1) and decreasing. It has no x-intercept because all outputs are at most −1. Draw y=xy=x and use equal axis scales to make the reflection explicit. A vertical axis marked x in the source sketch is a label slip; the reconstruction uses y.

Expand diagram

Use the scrollbars or arrow keys to inspect details.

Equal-scale g=x²−x−2 for x≤−1 and inverse 1/2−sqrt(x+9/4) for x≥0 reflect in y=x. Closed endpoints (−1,0),(0,−1) are their only axial intercepts.
Equal-scale g=x²−x−2 for x≤−1 and inverse 1/2−sqrt(x+9/4) for x≥0 reflect in y=x. Closed endpoints (−1,0),(0,−1) are their only axial intercepts.
Equal-scale g=x²−x−2 for x≤−1 and inverse 1/2−sqrt(x+9/4) for x≥0 reflect in y=x. Closed endpoints (−1,0),(0,−1) are their only axial intercepts.
How did this part go?

(iii) State where the identity holds

Answer

{x∈R:x≥0}\{x\in\mathbb R:x\ge0\}.

gg−1(x)=xgg^{-1}(x)=x on the entire original domain of the inner inverse: Dgg−1=Dg−1=[0,∞)D_{gg^{-1}}=D_{g^{-1}}=[0,\infty). The inverse outputs are accepted by gg. Do not extend the identity beyond this domain.

How did this part go?

(iv) Show that the composite exists

Answer

fg−1fg^{-1} exists on [0,∞)[0,\infty).

Rg−1=(−∞,−1]⊆Df=(−∞,3)∪(3,∞)R_{g^{-1}}=(-\infty,-1]\subseteq D_f=(-\infty,3)\cup(3,\infty). Every inverse output avoids the excluded input 3, so fg−1fg^{-1} exists on all of Dg−1=[0,∞)D_{g^{-1}}=[0,\infty).

How did this part go?

(iv) Map the exact range

Answer

Rfg−1=[5/2,3)R_{fg^{-1}}=[5/2,3).

Use the source mapping method: Dg−1=[0,∞)D_{g^{-1}}=[0,\infty) maps through g−1g^{-1} to (−∞,−1](-\infty,-1], then through ff to [5/2,3)[5/2,3). The question asks for the range, so no composite rule is needed. At the included endpoint, f(−1)=(7−3(−1))/(3−(−1))=10/4=5/2f(-1)=(7-3(-1))/(3-(-1))=10/4=5/2. As the intermediate input tends to −∞-\infty, f(t)=3+2/(t−3)f(t)=3+2/(t-3) approaches 3 from below without reaching it. On t≤−1t\le-1 the function decreases as t increases; it reaches every output between those bounds. Include 5/2 and exclude 3.

Expand diagram

Use the scrollbars or arrow keys to inspect details.

Mapping[0, infinity) through g inverse reaches (−infinity,−1], then f reaches[5/2,3). Restricted f graph has closed endpoint(−1,5/2) and approaches horizontal asymptote y3 from below as input tends to−infinity.
Mapping[0, infinity) through g inverse reaches (−infinity,−1], then f reaches[5/2,3). Restricted f graph has closed endpoint(−1,5/2) and approaches horizontal asymptote y3 from below as input tends to−infinity.
Mapping[0, infinity) through g inverse reaches (−infinity,−1], then f reaches[5/2,3). Restricted f graph has closed endpoint(−1,5/2) and approaches horizontal asymptote y3 from below as input tends to−infinity.
How did this part go?