Find all function f:R→R such that for all x,y∈R the following equality holds f(⌊x⌋y)=f(x)⌊f(y)⌋ where ⌊a⌋ is greatest integer not greater than a. Proposed by Pierre Bornsztein, France
Problem
Source:
Tags: algebra, functional equation, IMO
07.07.2010 20:27
Put x=y=0. Then f(0)=0 or ⌊f(0)⌋=1. ∙ If ⌊f(0)⌋=1, putting y=0 we get f(x)=f(0), that is f is constant. Substituing in the original equation we find f(x)=0, ∀x∈R or f(x)=a, ∀x∈R, where a∈[1,2). ∙ If f(0)=0, putting x=y=1 we get f(1)=0 or ⌊f(1)⌋=1. For f(1)=0, we set x=1 to find f(y)=0 ∀y, which is a solution. For ⌊f(1)⌋=1, setting y=1 yields f(⌊x⌋)=f(x), (∗). Putting x=2,y=12 to the original we get f(1)=f(2)⌊f(12)⌋. However, from (∗) we have f(12)=f(0)=0, so f(1)=0 which contradicts the fact ⌊f(1)⌋=1. So, f(x)=0, ∀x or f(x)=a, ∀x, a∈[1,2). Hope I'm right.
07.07.2010 20:35
Let P(x,y) be the assertation of f([x]y)=f(x)[f(y)] P(0,0):f(0)=f(0)[f(0)] There are 2 cases: 1. [f(0)]=1. Then, P(x,0):f(0)=f(x) Making f(x)=c for some constant c. Substituting back to P(x,y), we get: c=c[c] or [c]=1. Then, 1≤c<2. 2. f(0)=0. P(1,1):f(1)=f(1)[f(1)] Again, there are 2 subcases: a. f(1)=0. P(1,y):f(y)=0. After checking back to the f.q., we get f(x)=0 is also a solution. b. [f(1)]=1. P(x,1):f(x)=f([x]). Consider: P(2010,12010):f(1)=f(2010)[f(12010)]=f(2010)[f(0)]=0. Contradiction, because 1=[f(1)]=[0]=0. Then the function which satisfies P(x,y) is: f(x)=0∀x∈R f(x)=c∀x∈R,1≤c<2. @above: we almost have the same solution.. hope we're right..
07.07.2010 20:47
Solution to Problem 1: f([x]y)=f(x)[f(y)] (1) Substituting y=0 we have f(0)=f(x)[f(0)]. If [f0)]≠0 then f(x)=f(0)[f(0)]. Then f(x) is constant. Let f(x)=c. Then substituting that in (1) we have c=c[c]⇒c(1−[c])=0⇒c=0, or [c]=1. Therefore f(x)=c where c=0 or c∈[1,2) If [f(0)]=0 then f(0)=0. Now substituting x=1 we have f(y)=f(1)[f(y)]. If f(1)≠0 then [f(y)]=f(y)f(1) and substituting this in (1) we have f([x]y)=f(x)f(y)f(1). Then f([x]y)=f(x[y]). Substituting x=1/2,y=2 we get f(0)=f(1). Then f(1)=0, which is a contradiction Therefore f(1)=0. and then f(y)=0 for all R Then the only solutions are f(x)=0 or f(x)=c where c∈[1,2)
07.07.2010 22:04
08.07.2010 00:10
The answers are f(x)=c for all x and constant c, where c∈{0}∪[1,2). Put in x=0 to get f(0)=f(0)⌊f(y)⌋. This implies either (a) ⌊f(y)⌋=1 for all y, or (b) f(0)=0. If (a) is true, then the original equation reduces to f(⌊x⌋y)=f(x). Set y=0 and we get f(0)=f(x), or rather that f is constant. We know ⌊f(0)⌋=1, and we can verify for any a∈[1,2), f(x)=a for all x is a solution. Otherwise we have (b), f(0)=0. Let 0<x<1. Then ⌊x⌋=0 and the equation becomes f(x)⌊f(y)⌋=0. Either (c) ⌊f(y)⌋=0 for all y, or (d) f(x)=0 for all 0<x<1. If (c) is true, then the original equation becomes f(⌊x⌋y)=0, and x=1 gives f(y)=0 for all y. If (d) is true, then let 0<y<1 in the original equation. We get f(⌊x⌋y)=0. For any real a with ⌊a⌋≠0, take x=2a and y=a⌊2a⌋, and note that 0<y<1. Then f(⌊x⌋y)=f(a)=0, and so f is constant at 0. EDIT: Alternate finish since the end is a bit sloppy. Use x=2,y=1/2 to get f(1)=0, then plug in x=1 to the original equation.
08.07.2010 01:02
good problem..hope all the indian members could solve it
08.07.2010 03:19
Ohhhh i'm sure that the most of the bolivian team didn't do this problem, i think just Arran did it ^_^ the bolivian team was not tried tried
08.07.2010 09:54
Clearly f(⌊x⌋y)=f(⌊⌊x⌋⌋y)=f(⌊x⌋)⌊f(y)⌋, so (f(x)−f(⌊x⌋))⌊f(y)⌋=0 for all x,y∈R. If ⌊f(y)⌋=0 for all y∈R, then by taking x=1 we get f(y)=f(1)⌊f(y)⌋=0, so f is identically null (which checks). If, contrariwise, ⌊f(y0)⌋≠0 for some y0∈R, it follows f(x)=f(⌊x⌋) for all x∈R. Now it immediately follows f(x)=f(⌊x⌋⋅1)=f(x)⌊f(1)⌋, hence f(x)(1−⌊f(1)⌋)=0. For x=y0 this implies ⌊f(1)⌋=1. Assume ⌊f(0)⌋=0; then 1≤f(1)=f(2⋅12)=f(2)⌊f(12)⌋=f(2)⌊f(⌊12⌋)⌋=f(2)⌊f(0)⌋=0, absurd. Therefore ⌊f(0)⌋≠0, and now y=0 in the given functional equation yields f(0)=f(x)⌊f(0)⌋ for all x∈R, therefore f(x)=c≠0 constant, with ⌊c⌋=⌊f(1)⌋=1, i.e. c∈[1,2) (which obviously checks).
08.07.2010 10:03
That is a nice problem for number 1 of day 1. If i could solve it in 40 minutes, I hope everyone from Uzbekistan team managed to solve it =p
08.07.2010 17:03
08.07.2010 20:30
Yay, I solved an IMO problem. Is it just me or does it seem as though this year's #1 and #4 are much easier than normal? Solution We claim that either f(x)=k such that k∈[1,2) or f(x)=0. We can assume that f(1)≠0. Because otherwise, by the equality, we have that f(y)=f(1)⌊f(y)⌋=0 and the claim is satsfied. Let f(1)=k≠0 where k∈[1,2). Now we will establish several results about f(x). (1) Let x=y=1. Then since f(1)≠0, it follows that f(1)⌊f(1)⌋⇒⌊f(1)⌋=1. (2) Let x=1. Then f(y)=f(1)⌊f(y)⌋=k⌊f(y)⌋. (3) Let y=1. Then by (1), f(⌊x⌋)=f(x)⌊f(1)⌋=f(x). Now we will divide into two cases. Case 1: f(0)≠0. Let x=0. Then it follows that f(0)=f(0)⌊f(y)⌋⇒⌊f(y)⌋=1. Combining this with (2) yields that f(x)=k⌊f(x)⌋=k where k∈[1,2) and the claim is satisfied. Case 2: f(0)=0. Let y be such that y∈[0,1). Then it follows that f(⌊x⌋y)=f(x)⌊f(y)⌋. Now note that by (3), we have that f(y)=f(⌊y⌋)=f(0)=0. Hence f(⌊x⌋y)=f(x)⌊0⌋=0. Now we will prove that for each real number a, we have x and y such that y∈[0,1) and a=⌊x⌋y. If a>0, then let x=a+1 and 0≤y=a/⌊a+1⌋<1. It is clear that from here, we have that a=⌊x⌋y. If a<0, then let x=a−1 and 0≤y=a/⌊a−1⌋<1. Hence a=⌊x⌋y This yields that for all real a, it follows that f(a)=0 and the claim is satisfied. Now we will show that both of the claimed functions satisfy the equation for all real x and y. (1) f(x)=k where k∈[1,2)⇒f(⌊x⌋y)=k=f(x)⌊f(y)⌋ (2) f(x)=0⇒f(⌊x⌋y)=0=f(x)⌊f(y)⌋ Therefore the claim has been proven.
09.07.2010 12:33
My Solution : For : x=0 we get : f(0)=f(0)⌊f(y)⌋ So either f(0)=0 or ⌊f(y)⌋=1 for all y ⌊f(y)⌋=1 for all y is indeed a solution . Now if f(0)=0 : Letting y=1 and x=n∈N∗ we get : f(n)=f(n)⌊f(1)⌋ if f(n)=0 for all n then for x=n we must have f(yn)=0 for all y so f(y)=0which is indeed a solution . If f(0)=0 and ⌊f(1)⌋=1 and for x=n : f(ny)=f(n)⌊f(y)⌋ for x=n+p where n∈Z and p∈(0,1): f(ny)=f(n+p)⌊f(y)⌋ so either : f(n)=f(n+p) for p∈(0,1) or ⌊f(y)⌋=0 for all y but ⌊f(1)⌋=1 so : f(n)=f(n+p)forp∈(0,1) Let n be an integer n<0 we have : f(n+1)=f(⌊n⌋n+1n)=f(n)⌊f(1+1n)⌋=f(n).⌊f(0)⌋=0 So f(x)=0 for all x≤−2 Now let n be an integer n>1 : We have : f(n+1)=f(⌊n⌋n+1n)=f(n)⌊f(1+1n)⌋=f(n).⌊f(1)⌋=f(n)=c so f(x)=cfor all x≥2 forx=−2 and y=−1 we get c=0 Using the fact that f(n)=f(n+p) forp∈(0,1) we will have : f(x)=0for x∈(−∞,−1)∪[0,1)∪(2,+∞) f(x)=f(−1)for x∈[−1,0) f(x)=f(1)forx∈[1,2) but we have : f(1)=f(2.12)=f(2).⌊f(12)⌋=0 whish gives a contradiction . So the solutions are : f(x)=0 f(x)=c,c∈[1,2[ Hope it's correct !
10.07.2010 11:04
f(⌊x⌋y)=f(x)⌊f(y)⌋- Q case 1 : there is a s.t. ⌊a⌋ is not zero and f(a)=0. f(⌊a⌋y)=f(a)⌊f(y)⌋. so f(⌊a⌋y)=0. put y=z⌊a⌋. so we get f(z)=0 for all z. case 2 : take the negation of case 1. so if f(x)=0 then x belongs to [0,1). case 2.1 : f(0) is not equal to 0. put x=0 to equation Q. so we get ⌊f(y)⌋=1. thus equation Q gives f(⌊x⌋y)=f(x). put x=1. then f(y)=f(1) for each y. thus f(x)=b for each x where b is a number in [1,2). case 2.2 : f(0)=0. case 2.2.1 : there is b in [1,0) s.t. f(b) is not zero. so f(⌊b⌋y)=f(b)⌊f(y)⌋. thus ⌊f(y)⌋=0. by Q we get f(⌊x⌋y)=0. putting x=1 gives f(y)=0 for each y. case 2.2.2 : negation of case 2.2.1. put y=0.5 to Q. then since f(0.5)=0 we get f(⌊x⌋2)=0. putting x=2 will give a contradiction since f(1) is not zero. thus the solutions are f(x)=0 for each x, f(x)=c for each x where c is a constsnt in [1,2). we can easily verify them. Q.E.D
10.07.2010 18:59
I will try to do it differently, maybe it's wrong I dont know.
11.07.2010 15:43
Dr N0 wrote: Let k be a positive integer, and let ϵ>0 and let ϵ be very small number. Then because of f is continous we have: f is not assumed to be continuous. However, you were able to end up your proof differently : Case [f(1)]=1 Taking x=2 and y=12 gives f(1)=f(2)[f(12)]. (1) Taking y=1 gives f([x])=f(x)[f(1)], therefore, f([x])=f(x) because we are in case [f(1)]=1. Hence for all real x in ]0,1[,f(x)=f(0). (2) (1) and (2) implies that f(1)=f(2)[f(0)]. And because [f(1)]=1≠0, then [f(0)]≠0, hence f(0)≠0. Now let x=0. Then f(0)=f(0)[f(y)]. Thus, f(0)=0 or [f(y)]=1. But f(0)≠0. So [f(y)]=1. In particular, [f(0)]=1. Let y=0. Then f(0)=f(x)[f(0)], which implies f(0)=f(x) because [f(0)]=1. Therefore, f is constant. At that step, the conclusion follows easily.
12.07.2010 04:02
If f is constant, say f(x)=c for all x and some constant c; we get c=c⌊c⌋, so c=0 or ⌊c⌋=1, and c=0 or 1≤c<2. Plugging in y=0 gives f(0)=f(x)⌊f(0)⌋. If f(0)≠0, we also have that the RHS is not 0, so f(x)=f(0)⌊f(0)⌋=c, contradiction. So f(0)=0. Now, let x be such that f(x)=0. Then, for all y, f(⌊x⌋y)=0 for all y∈R. If ⌊x⌋≠0, this will give us f(x)=0 for all x, contradiction. Thus, ⌊x⌋=0. Now, switching x and y gives f(⌊y⌋x)=f(y)⌊f(x)⌋=0. But if x≠0, we can choose y appropriately so that ⌊y⌋x>1, which is a contradiction. Thus, x=0. Finally, plugging in x=1/2 gives f(0)=f(1/2)⌊f(y)⌋ for all y, giving ⌊f(y)⌋=0 for all y. So plugging in x=1 into the original equation, f(y)=0 for all y, again a contradiction. Thus, f is constant, and our only solutions are those above.
13.07.2010 10:06
Here is my solution (The same idea as many solutions given here): Consider the two possible cases: 1-If f(12)=0: Then for any x∈R, f(x)=f(⌊−12⌋(−x))=f(−12)⌊f(−x)⌋=f(−1)⌊f(12)⌋⌊f(−x)⌋=0. Clearly, f=0 satisfies the equation. 2-If f(12)≠0: Then for any x∈R, f(0)=f(12)⌊f(x)⌋ and f(x)=f(1)⌊f(x)⌋. Therefore f(x)=f(0)f(1)f(12)=c=f(12)≠0 is constant. Using c=f(1)=f(1)⌊f(1)⌋=c⌊c⌋ we deduce that ⌊c⌋=1. Clearly, f=c with c∈[1,2[ satisfy the equation.
13.02.2011 05:41
Let x=0 So f(0)=f(0)⌊f(y)⌋ So f(0)(1-⌊f(y) ⌋ )=0 F(0)=0 OR ⌊f(y) ⌋=1 Case 1: f(0)=0 Substituting y=1, the equation become ⌊f(x) ⌋=f(x)⌊f(1) ⌋ Substituting x=1 and y=1, we get ⌊f(1) ⌋=0 or 1. If ⌊f(1) ⌋=1 it’s on the second case. So ⌊f(1) ⌋=0 Substituting x=1, we get f(y)=f(1) ⌊f(y) ⌋ If f(y) positive number except 0 f(1) must be greater of same with 1, contradiction So f(y) must be negative or 0 for all y If f(y) is negative, so f(1) is positive (contradiction) So f(y)=0 Case 2: f(y)=x which 1≤x<2 The answer: f(y)=0 or f(y)=c which 1≤x<2 Please help me if this answer is true or not. Thank you
09.06.2011 11:42
i agree that it's a nice functional equation.but wasn't it too easy?i hope that France wont provide such water problems!France has ox of god like Gabriel Dospinescu,why dont they provide some problems for imo?
17.12.2023 22:38
Substituting (0,0), we find f(0)=f(0)⋅⌊f(0)⌋, so f(0)=0 or ⌊f(0)⌋=1. Case 1: f(0)=0. Substituting (1,1) tells us f(1)=f(1)⋅⌊f(1)⌋, so f(1)=0 or ⌊f(1)⌋=1. f(1)=0: Substitute (1,t) to get f(t)=f(1)⋅⌊f(t)⌋=0, so we get the solution f(x)=0. ⌊f(1)⌋=1: Suppose we have a real number α∈(0,1). Then substituting (α,1) tells us f(0)=f(α)⌊f(1)⌋⟹f(α)=0.However, if we substitute (t,α), we find f(⌊t⌋α)=f(t)⋅⌊f(α)⌋=0.Since ⌊t⌋α can take on any real value, this means f(x)=0 for all x, which contradicts ⌊f(1)⌋=1. ◻ Case 2: ⌊f(0)⌋=1. Substitute (0,t) to get f(0)=f(0)⋅⌊f(t)⌋⟹⌊f(t)⌋=1.Thus the values of f in this case are bounded from [1,2). We can prove they all must be the same value by substituting (t,0) to find f(0)=f(t)⋅⌊f(0)⌋=f(t).◻ Hence our final solution is f(x)=c,c∈{0}∪[1,2), which we can test to work. ◼
30.12.2023 07:22
canada wrote: Find all function f:R→R such that for all x,y∈R the following equality holds f(⌊x⌋y)=f(x)⌊f(y)⌋where ⌊a⌋ is greatest integer not greater than a. Proposed by Pierre Bornsztein, France A Very easy question for international mathematics olimpiad.
31.12.2023 10:53
f≡0 and f≡c where c∈[1,2) is a constant. P(x,0)⟹f(0)=f(x)⌊f(0)⌋. If f(0)∉[0,1), then ⌊f(0)⌋≠0⟹f(x)=f(0)⌊f(0)⌋=c. So putting x=0, we get c=c⋅⌊c⌋⟹c∈[1,2) because c≠0 as f(0)∉[0,1). Otherwise f(0)=[0,1)⟹⌊f(0)⌋=0⟹f(0)=0. If there exists α∈[0,1) such that f(α)≠0, then P(α,y)⟹0=f(α)⌊f(y)⌋⟹0≤f(y)<1 for all y∈R. So f(⌊x⌋y)=f(x)⌊f(y)⌋=0. Now x=1⟹f(y)=0 for all y∈R. Now putting y=α gives a contradiction. Thus we must have that f(x)=0 for all x∈[0,1). If f(1)=0, then P(1,y)⟹f(y)=0 for all y. Otherwise P(1,1)⟹⌊f(1)⌋=1⟹1≤f(1)≤2. But at the same time, P(2,12)⟹f(1)=0, contradiction and we are done.
11.01.2024 19:21
P(0,0) to get that f(0)=f(0)⌊f(0)⌋, so f(0)=0 or ⌊f(0)⌋=1 If ⌊f(0)⌋=1, then P(x,0) gives f(0)=f(x), so f(x)=c works, where 1≤c<2 If f(0)=0, then P(1,1) gives f(1)=f(1)⌊f(1)⌋, so f(1)=0 or ⌊f(1)⌋=1 If f(1)=0, then P(1,x) gives f(x)=0 If ⌊f(1)⌋=1, then P(x,1) gives f(⌊x⌋)=f(x). P(x,1⌊x⌋) gives f(1)=f(x)⌊f(1⌊x⌋)⌋=f(x)⌊f(0)⌋=0. This is a contradiction.
07.03.2024 17:50
Firstly, let y=0. If ⌊f(0)⌋ is not equal to 0,then it is obvious that f(x) is constant. f(x)=c. Substituting into the original equation gives us that ⌊f(y)⌋ is equal to 1.( If c is not 0. That case is trivial anyways). So ⌊c⌋ is 1 .As a result, 1≤c<2. Now, assume that ⌊f(0)⌋ is 0. So, f(0)=0 by the first substitution. Putting x=y=1 gives us that f(1)=0 or ⌊f(1)⌋=1. The former case is trivial; we can finish it easily, as it yields that f(x)=0 for all x. Now look at the latter case. Let y=1 to get that f(x)=f(⌊x⌋). However, P(x,y), where x is a large real number and y=1⌊x⌋ gives us that f(1)=0 since f(y)=f(⌊y⌋)=0 for that y. Contradiction.
12.03.2024 02:22
The answer is f≡c for c=0 or any c such that ⌊c⌋=1. Denote the given assertion as P(x,y). From P(0,y) we have f(0)=f(0)⌊y⌋ for any real y. This rearranges to f(0)(⌊f(y)⌋−1)=0, so we have two cases to consider: Case 1: f(0)=0. Take any t such that ⌊t⌋=0. From P(t,t)=0, we get 0=f(t)⌊f(t)⌋. This is enough to imply that f(t)=0, as if ⌊f(t)⌋=0 then P(1,t) yields f(t)=0. Now, let k be any real number. Then we can pick reals r0,t0 such that ⌊r0⌋t0=k and ⌊t0⌋=1 (choose any r0 such that ⌊r0⌋ has the same sign as k and |⌊r0⌋|>|k|). From P(r0,t0) we conclude that f(k)=0, so f≡0. Case 2: ⌊f(y)⌋=1 for all y. From P(1,y) we then get f(y)=f(1). This is enough to imply that f≡c for some constant c such that ⌊c⌋=1.
16.05.2024 06:52
Our two solutions are f(x)=0 and f(x)=c∀c∈[1,2]. We can plug in (0,0) to get f(0)=f(0)⌊f(0)⌋ which implies f(0)=0 or ⌊f(0)⌋=1. If the latter is true then plug in (t,0) then plug in f(0)=f(t)=c∈[1,2]. If f(0)=0 we can plug in (1,1) to get f(1)=f(1)⌊f(1)⌋⟹f(1)=0⊻. If f(1) = 0 then (1, 0) gives f(y) = 0, done. If \left\lfloor f(1)\right\rfloor = 1 then (x, 1) \implies f(\left\lfloor x\right\rfloor) = f(x). Now plug in (1434, \frac{1}{1434}) to get f(1) = f(1434)f\left(\left\lfloor \frac{1}{1434} \right\rfloor\right) = 0, contradiction. So our solutions are f(x) = 0 and \left\lfloor f(x)\right\rfloor = 1.
16.06.2024 05:25
We claim the only solutions are f\equiv c, where c\in \{0\}\cup [1,2). It is easy to see all such functions work; we know prove they are the only ones. Let P(x,y) denote the assertion. From P(0,0), we have that f(0) = f(0)\lfloor f(0) \rfloor. Thus, f(0) = 0 or 1\le f(0) < 2. We proceed by casework. Case 1: f(0)=0. Fix x\in (0,1), and we have f(0) = 0 = f(x)\lfloor f(y) \rfloor. Suppose there exists some real y such that f(y) \notin [0,1). Then, f(x)=0 for all x\in (0,1). Then, let x vary over the reals and y over (0,1) to find that f(\lfloor x \rfloor y) = 0. But since \lfloor x \rfloor y can be any real number, f\equiv 0, contradiction. Thus, we always have f(x)\in [0,1). Thus, we conclude f(\lfloor x \rfloor y) = 0, which implies f\equiv 0, as desired. Case 2: f(0)\in [1,2). If this is the case, then by P(x,0), we have f(0) = f(x), as desired, and we are done. \blacksquare
21.07.2024 02:40
Put \displaystyle x=y=0. We have \displaystyle f(0)=0 or \left\lfloor f(0)\right\rfloor=1. We split into two cases. (i)\left\lfloor f(0)\right\rfloor=1 Putting \displaystyle y=0 gives f\equiv f(0),so f\equiv c where 1\le c \le2 \displaystyle (ii)f(0)=0. Put \displaystyle x=y=1. We have \displaystyle f(1)=0 or \left\lfloor f(1)\right\rfloor=1. Claim. \displaystyle f(1)=0 \emph{Proof.} We use proof by contradiction. Assume that \left\lfloor f(1)\right\rfloor=1. Puttting \displaystyle y=1 gives f(\left\lfloor x\right\rfloor)=f(x) Now f(1)=f(2)\left\lfloor f(\frac{1}{2})\right\rfloor=f(2)\left\lfloor f(0)\right\rfloor=0 since \left\lfloor \frac{1}{2}\right\rfloor=0.\blacksquare Put \displaystyle x=1,y=t. \displaystyle f(t)=0 Combining these cases together, we have f\equiv c, where c\in \{0\}\cup [1,2)
25.07.2024 01:58
Let P(x,y) denote the assertion. First, we have that P(0,y) implies f(0)=f(0)\lfloor f(y)\rfloor, so either f(0)=0 or \lfloor f(1)\rfloor = 1 for all y. In the latter case, we have f(\lfloor x\rfloor y)=f(x)\cdot 1, so setting x=1, we get f(y)=f(1). Plugging in f(x)=c for all x into the equation, with c=f(1), we see that either c=0 or \lfloor c\rfloor = 1. Now, we deal with the case where f(0)=0. P(1,1) gives us f(1) = f(1)\lfloor f(1)\rfloor, so either f(1)=0 or \lfloor f(1)\rfloor = 1. If f(1)=0, then P(1,y) gives f(y)=f(1)\lfloor f(y)\rfloor = 0 for all y. Otherwise if \lfloor f(1)\rfloor = 1, then P(x,1) gives f(\lfloor x\rfloor) = f(x)\lfloor f(1)\rfloor = f(x).Then, for any 0\leq y<1, we have f(\lfloor x\rfloor y) = f(x)\lfloor f(y)\rfloor = f(x) \lfloor f(\lfloor y\rfloor)\rfloor = f(x)\cdot 0 = 0,so since for any a, we can select some x>a+1 and let 0\leq y=\frac{a}{\lfloor a+1\rfloor } < 1, we have f(a)=0 for all a. Since we have addressed all our cases, the solutions are f(x)=c for all x where c\in \{0\}\cup [1, 2), and it's easy to check that all such functions satisfy our equation.
14.08.2024 17:19
Warideeb wrote: Let P(x,y) be the assertion of f(\lfloor x \rfloor y)=f(x) \lfloor f(y) \rfloor Our Claim is that f \equiv 0 or f \equiv c where 1\leq c < 2 Now f \equiv 0 is clearly a solution. Now we assume f \not\equiv 0 Now P(0,x) gives us f(0)=f(0) \lfloor f(x) \rfloor Now if f(0) \ne 0 then this gives us f \equiv c for some 1\leq c < 2 Now lets assume f(0)=0 P(1,1) gives us f(1)=c where 1 \leq c < 2 now P(x,1) gives us f(x)=f(\lfloor x \rfloor) define a number 0 \leq A <1 Now P(x,A) and subsituting f(A) with f(\lfloor A \rfloor) gives us f(\lfloor x \rfloor A)=0 and give is f \equiv 0 and here is our solution
31.08.2024 21:24
bad formatting because i copied my solution from mathdash vary x along [n, n+1), then we see for all x in this interval we must either have f(y) in [0,1) (to make the floor zero), or if the floor is not zero we are forced that all x in the interval have the same output when f is applied. we can now divide into two cases: f(y) lies in [0,1) for all y. then we get f(floor(x)y) = 0, so letting x = 1 gives f(y) = 0, which trivially satisfies the fe. otherwise, take some value of y for which this is not true. then we get for all integers n, if a,b in [n, n+1), then f(a) = f(b). now e = n, 0 < f - e < 1. then consider varying y over [e,f], then the right side is constant, the left side must be as well, so f(floor(x)y) is constant for all e <= y <= f. with e = 0, f = (m - 1)/m for some large integer m, set x = m, we then get f(my) is constant for all 0 <= y <= m - 1, take m to infinity and we see f is constant over positives. then set x = -1, vary y over all positives and see f is constant over negatives. we show the constant value for postiives and negatives must be the same. firstly, if the positive value is zero set x = 1 and vary y to get negative constant is also zero, so we can dsicrard this, same for other case except do x = -1. now neither constant value is 0, so set the value for x to 0 and we then get 1 = floor(f(y)), so floor(f(0)) = 1. then the positive constant value is in [1,2), then substite negative x, y = 0, then we instantly get the negative constant is also the same as f(0), so the other solution is f = k for k in [1,2), which obviously satisfies the fe.
07.01.2025 09:41
Let P(x, y) denote the assertion. Note that if there is some k such that f(k)=0, P(k, y) \implies f(x) = 0 for all x. Thus assume that f(x) has no zeros. Thus P(0, 0) \implies \lfloor f(0) \rfloor = 1, so P(x, 0) \implies f(x) = f(0)=c. Clearly 1 \leq c < 2 and we see easily that this solution works. Therefore, the only solutions are f(x) = 0, c \in [1, 2).
12.01.2025 17:55
Let P(x,y) be the assertion for f([x]y) = f(x)[f(y)]. By P(0,0) we get: f(0) = f(0)[f(0)]. Here there are two cases: Case 1: f(0) \neq 0 So we can cross out f(0) and have [f(0)] = 1 Now by P(x,0) we have: f(0) = f(x) so f is constant. Also by [f(0)] = 1 we have that if f(x)=c then [c] should equal to 1. We can easily check that for any such c the equality holds. Case 2: f(0)=0 Now by P(1,1) we have f(1) = f(1)[f(1)]. If f(1) = 0 then P(1,x) gives f(x) = 0 which we can check that it is a solution. So we continue by supposing f(1) \neq 0 to check the other case. As a result, we can cross out f(1) and have that [f(1)]=1. Then by P(x,1) we have f([x])=f(x). Now \forall x>2: P(x,\frac{1}{[x]}) gives us f(1) = f(x).[f(\frac{1}{[x]}] and as we know [f(\frac{1}{[x]})] = [f([\frac{1}{[x]}])] = [f(0)] = 0. so f(1) = 0, contradiction. To sum up, we proved that the only functions in which the given equality holds are f(x) = c, \forall x \in \mathbb{R}, \forall c \in [1,2) and f(x) = 0, \forall x \in \mathbb{R}. And we can easily check that these two functions work.
14.01.2025 13:07
ily >_< Lets take x=0. Then, f(0)=f(0)\lfloor f(y)\rfloor. If f(0)\neq 0, then \lfloor f(y) \rfloor=1. If so, f(\lfloor x \rfloor y)=f(x). Fix x>1434. Then y times a scalar not zero gives all reals, hence f is constant. Now, if f(0)=0, then take 0\leq x<1, so 0=f(x)\lfloor f(y)\rfloor. We vary y over all reals now. If at any point \lfloor f(y)\rfloor\neq 0, then pick that y, and thus f(x)=0 for all 0\leq x<1. Now, in the initial equation let 0\leq y<1, so f(y)=0, and f(\lfloor x\rfloor y)=0. Now say we want to achieve z with \lfloor x\rfloor y. If z>0, then make \lfloor x\rfloor>z, and scale down by y. If z<0 then make \lfloor x \rfloor < z and scale down by y. Of course f(0)=0 already so nothing is to be done at 0. Hence f is simply 0. Else, \lfloor f(y)\rfloor=0 for all y. Then, f(\lfloor x\rfloor y)=0. Now, fix x>1+4+3+4, and vary y to get all reals, so f is zero again. Now, if f was just constant, then we have c=c\lfloor c\rfloor. Thus, c=0 or \lfloor c\rfloor =1. This means that f(x)=0 or f(x)=c, where 1\leq c<2.