Functions: Star problems
Ten optional challenges to stretch your reasoning. Work on paper, use hints when you need them, and check the answer or full solution when you are ready. You can skip these problems and continue the course.
- 1 of 3 stars: Stretch
- 2 of 3 stars: Challenge
- 3 of 3 stars: Deep challenge
Stars indicate difficulty within this set.
0 of 10 completed · 0 skipped
Progress saved in this browser.
Progress can't be saved in this browser, so your choices last for this visit only.
-
Problem 1 Two names for one input
Difficulty: 1 of 3 stars, Stretch
A student proposes a function satisfying for every real .
(a) Prove that no such function exists.
(b) Find all real constants for which a function can satisfy for every real . Describe every possible , including any freedom in its definition.
Builds on What Is a Function?, Completing the Square
- Hint 1
Two different choices of can send the same number into .
- Hint 2
Compare with , then use the full range of .
Answer
(a) Impossible. (b) Exactly , with arbitrary real ; then for , and values for are arbitrary.
Full solution
In part (a), gives , whereas gives .
One input cannot have two outputs, so the proposed function cannot exist.
For part (b), the same two substitutions require , hence .
This is necessary even without assuming a formula for .
With this condition the right side becomes .
The possible inputs are exactly all real numbers : each is obtained by taking
Therefore the condition forces on that entire interval.
Conversely, this rule gives the desired identity for every real whenever .
No number less than ever appears as the input .
The identity consequently puts no restriction on there.
Assigning any real value separately at each such input completes a valid function.
Thus the displayed formula need not hold on all of .
Answer
(a) Impossible. (b) Exactly , with arbitrary real ; then for , and values for are arbitrary.
Key idea
Before solving a functional equation, check whether different substitutions name the same input and which inputs are ever reached.
- Hint 1
-
Problem 2 The missing inputs remain missing
Difficulty: 1 of 3 stars, Stretch
Let on its natural real domain, and let on its natural real domain.
Find the exact domain and range of each function and the exact domain of each composition and . Are and inverse functions? Explain why simplifying both compositions to does not make either composition defined for every real input.
Builds on Algebraic Fractions
- Hint 1
A composition needs both a valid first input and a valid intermediate output.
- Hint 2
To establish a range, solve or for and determine exactly which values are possible.
Answer
and are inverses. The composition domains are respectively and .
Full solution
The formula for requires , and its output cannot be .
Every nonzero real is obtained at , which is allowed.
Thus its range is exactly .
Similarly, requires and never gives .
Every is obtained at , so its range is exactly .
For , the inner function requires .
Its output never equals , so the outer function adds no restriction.
On this domain,
For , the inner restriction is , and its nonzero output always lies in the domain of .
On this domain,
Both compositions are the appropriate identity functions on the original domains, so these are inverse functions.
Cancellation has simplified their values where they exist; it cannot supply the undefined first step at a missing input.
Answer
and are inverses. The composition domains are respectively and .
Key idea
A function includes its domain; an algebraically simplified formula alone does not specify the same function.
- Hint 1
-
Problem 3 The largest reversible interval
Difficulty: 1 of 3 stars, Stretch
Consider . Find the largest interval containing on which the restriction of has an inverse function. Include or exclude boundary points as appropriate, prove that no larger interval containing works, and give the inverse with its domain and range.
Builds on Completing the Square
- Hint 1
The same output occurs at inputs reflected across the vertex.
- Hint 2
Write . The inverse must choose a root consistent with the chosen interval.
Answer
The interval is . Its inverse is , with domain and range .
Full solution
Completing the square gives
On , increasing the input strictly increases the nonnegative number , and hence its square.
The restriction is one-to-one and contains the input .
Its range is , since the square takes every nonnegative value.
Any interval containing and any point must also contain two distinct points and for a sufficiently small positive .
Those points give the same output , so such an interval is not one-to-one.
Therefore every interval containing that works is contained in , establishing the claimed largest interval.
For a given output , the equation has two formal root choices, but selects
This gives exactly one allowed input for every allowed output.
Its domain is and its range is , the reverse of those of the restricted function.
Answer
The interval is . Its inverse is , with domain and range .
Key idea
Choosing an inverse branch is a domain decision, and maximality must be justified using repeated outputs.
- Hint 1
-
Problem 4 One-way undoing on a finite set
Difficulty: 2 of 3 stars, Challenge
Let and . Define by .
(a) How many functions satisfy for every ?
(b) Find all such for which the sum of their five output values is .
(c) Can any function satisfy for every ? Prove your answer.
Builds on What Is a Function?
- Hint 1
For each square in , choose one input that maps to it under .
- Hint 2
For part (b), choosing signs for gives sum zero exactly when the positive magnitudes sum to half of .
Answer
(a) . (b) In input order , the outputs are or . (c) No.
Full solution
The condition forces .
For each of , the output of can independently be either square root in : respectively .
Each choice works because squaring recovers the input.
Thus there are functions.
For part (b), the magnitudes sum to .
The signed sum is zero precisely when the magnitudes assigned positive signs sum to .
A single magnitude cannot do this, and three magnitudes already sum to at least .
The only two-element subsets summing to are and .
These give exactly the two output lists stated.
For part (c),
The proposed condition would require both and , which is impossible.
Although each of the functions in part (a) chooses a valid way back from an output, none can recover both original inputs when the same square has two sources.
A one-way undoing condition does not guarantee an inverse.
Answer
(a) . (b) In input order , the outputs are or . (c) No.
Key idea
Choosing one preimage of every output can undo a function in one direction even when information has been lost.
- Hint 1
-
Problem 5 Returning after two steps
Difficulty: 2 of 3 stars, Challenge
Let for all real . Find every real such that . Separate the inputs that return after one step from those that return only after two steps, and identify the two-step pairs. Prove your list is complete without expanding a quartic.
- Hint 1
Name the intermediate output . Then the condition says .
- Hint 2
Subtract and . One factor distinguishes a fixed point from a two-step pair.
Answer
Fixed inputs: . The remaining inputs are and , which map to each other.
Full solution
Put .
The condition is equivalent to the pair and .
Subtracting gives , hence
Every solution therefore belongs to one of two cases.
If , then , so
The inputs and are fixed points and satisfy the original condition.
If , then .
Substituting into yields , whose roots are and
Each satisfies , and , so and
They are distinct and neither is a fixed point.
The two cases exhaust the factorized equation, and every candidate has been checked in the original two-step process.
Thus these four inputs are the complete list.
Answer
Fixed inputs: . The remaining inputs are and , which map to each other.
Key idea
For a two-step return, introduce the intermediate state and use the symmetry between the two equations.
- Hint 1
-
Problem 6 Which invented operations associate?
Difficulty: 2 of 3 stars, Challenge
Fix real constants and define an operation on all real numbers by .
(a) Find every triple for which for all real .
(b) Among these operations, identify every one possessing a two-sided identity: a real with for all real . Give the identity in each case.
- Hint 1
Associativity is an identity in three independent inputs, not an equation at a few test values.
- Hint 2
Expand both parenthesizations and compare the coefficients of and , then compare constants.
Answer
(a) or with arbitrary ; or . (b) Exactly , with identity .
Full solution
Expanding the left parenthesization gives ; expanding the right gives .
Their difference must vanish for every independent choice of .
Therefore
The first two conditions force .
If they are equal, the third condition allows any , giving the constant operation and the shifted sum
If they differ, , giving the projections and
The expansion also proves each listed family is associative, so the classification is complete.
A constant operation cannot have an identity for every real .
For the first projection, fails to equal arbitrary ; the second projection fails on the other side.
Finally, for the shifted sum, and hold exactly when .
This identity is unique.
Answer
(a) or with arbitrary ; or . (b) Exactly , with identity .
Key idea
Universal operation laws become coefficient conditions, but an identity must work on both sides.
- Hint 1
-
Problem 7 A fraction operation with hidden multiplication
Difficulty: 2 of 3 stars, Challenge
For real numbers with , define .
(a) Prove that the output is again between and , and prove that the operation is associative.
(b) Find all such that .
An associativity proof must apply to all permitted inputs, not only the numbers in part (b).
Builds on Algebraic Fractions
- Hint 1
First inspect and .
- Hint 2
Try the new coordinate . What is in terms of and ?
Answer
(a) The operation is closed on and associative. (b) The unique solution is .
Full solution
Since , the denominator is positive.
Also , and
These prove , so repeated operations are defined.
Define
This sends one-to-one onto the positive real numbers: solving gives for every .
The two factorizations above give
Therefore both and have the same -value, namely .
Since is one-to-one, they are equal.
This proves associativity for every allowed triple.
For part (b), the equation becomes
Thus , and
This lies in the domain.
Every step was reversible, so it is the unique solution.
Answer
(a) The operation is closed on and associative. (b) The unique solution is .
Key idea
A change of coordinates can turn an unfamiliar operation into an ordinary one and transfer its laws.
- Hint 1
-
Problem 8 Can a finite function have a square root?
Difficulty: 3 of 3 stars, Deep challenge
Let . A function is called a square root of here if for every .
(a) Does a square root exist when the outputs of , in input order , are ?
(b) Find every square root when the outputs of are . Prove completeness in both cases. You do not need any theorems about permutations.
- Hint 1
Because each proposed takes distinct inputs to distinct outputs, the same must be true of .
- Hint 2
Show . Thus must send inputs fixed by to inputs fixed by . In part (b), begin with the possibilities for .
Answer
(a) None. (b) Exactly two: output lists and .
Full solution
In either case, is one-to-one.
If , applying gives and hence .
Thus is one-to-one, and on the finite set it uses every output exactly once.
Furthermore
Consequently sends the set of fixed inputs of into itself, and, by its one-to-one property, onto that same set.
In part (a), the fixed set is , so must also map onto itself.
On two inputs it either leaves both in place or swaps them.
Doing either operation twice leaves both in place, whereas swaps them.
This is impossible.
In part (b), only is fixed, so .
The value cannot be , since .
It cannot be , since that would force and violate one-to-one behavior.
It is not .
Thus it is or .
If , then , then , then , using successively.
If , the same reasoning gives , , .
Directly composing each output list with itself gives the required , proving existence and completeness.
Answer
(a) None. (b) Exactly two: output lists and .
Key idea
Iteration constrains how a function moves whole cycles and fixed sets, not only individual values.
- Hint 1
-
Problem 9 Three substitutions close the loop
Difficulty: 3 of 3 stars, Deep challenge
Let and define for . Find every function satisfying
Your solution must justify that all compositions used stay in and must prove both existence and uniqueness. Then evaluate .
Builds on Algebraic Fractions
- Hint 1
Compute and before guessing a formula for .
- Hint 2
Write the given equation at the three inputs in this loop. Combine the three equations to isolate .
Answer
The unique function is on , and .
Full solution
For , is defined and nonzero.
It equals only if , which is excluded.
Thus maps into itself.
Direct calculation gives and , with all denominators valid on .
Here and mean repeated composition.
Write and .
Applying the identity at gives , , and
Add the first and third equations and subtract the second to obtain
This forces the stated formula at every allowed input, so there can be at most one solution.
To check existence, define
Then because .
Adding these gives , exactly as required.
Thus the formula does define a solution on the whole domain.
Finally, and , so
No continuity or polynomial assumption was needed.
Answer
The unique function is on , and .
Key idea
When substitutions return to their starting input, a functional equation can become a finite linear system.
- Hint 1
-
Problem 10 A square root of a shift
Difficulty: 3 of 3 stars, Deep challenge
Find every function such that for every integer . Give a complete description that constructs each function, prove that every function described works, and exhibit one that is not .
- Hint 1
Compute in two ways to relate and .
- Hint 2
Once and are known, the relation determines every even and odd input. Examine whether could be even.
Answer
For every integer there is one solution: and for all integers . These are all solutions; gives a nontranslation example.
Full solution
Associating the threefold composition in two ways gives
Repeatedly applying this relation, forwards and backwards, shows that if and , then and for every integer .
If were even, the even-input formula would give
But the original condition at requires , forcing , contrary to its being even.
Hence for some integer .
The odd-input formula now gives , so .
These are exactly the formulas in the answer, and every solution has been forced into this family.
Conversely, choose any integer and use those formulas to define .
An even input maps to , which then maps to
An odd input maps to , which then maps to
Thus for both parities and all integers.
For , even inputs map to and odd inputs to .
This function satisfies the condition but is not the uniform shift .
Answer
For every integer there is one solution: and for all integers . These are all solutions; gives a nontranslation example.
Key idea
Commuting an unknown function with its own iterate can reveal a recurrence that reduces an infinite problem to a few values.
- Hint 1