1. UNDERSTANDING THE PROBLEM
Prove or disprove: 15n2 is in Ω (3x 2n).
I know that f(n) in Ω(g(n)) means that after a certain n, f(n) is always larger than g(n).
First I need to figure out if this statement is true or not, so for help, I decided to graph the two lines.
It was obvious that while 15x2 started off outputting bigger, 3* 2x eclipsed it later on. Which means that that the statement is wrong.
2. DEVISING A PLAN
I knew that in order to prove that the statement was wrong, I needed to prove the contrapositive of the statement, which is: ∀c ∊ ℝ, ∀B∊ ℕ, ∃n∊ ℕ, n ≥ B ⋀ g(n) < c f(n).
With a little help from the hint from the question, I figured out that I need to use limits and l’Hopital’s rule to prove using the definition of a limit that g(n) < c f(n).
3. CARRYING OUT THE PLAN
I start out my proof as usual, assuming the ∀s.
Assume
c ∊ ℝ+
and B ∊ ℕ.
I want to show
that the limit as n → ∞ (15n2/3 x 2n) is zero which
would prove that the denominator, 3 x 2n is exponentially larger
than the numerator (15n2), thus making the claim that 15n2
is bigger than 3 x 2n false after a certain n.
Let L = limn →∞
15n2
/ 3 x 2n
Then L = limn →∞
30n
/ 3n x 2n-1
# L’Hopital’s Rule
Then L = limn →∞
30
/ 3n2 - 3n x 2n-2 # L’Hopital’s Rule Again
Then L = 0 # 1/∞ = 0
Now I will write
out the definition of a limit, and note that because c and ε are both ∊ ℝ+, we can
make c = ε.
Then ∃n0
∊
ℕ,
∀n
∊
ℕ,
n ≥
n0 ⇒
L - c < 15n2 / 3 x 2n < L + c
Since L = 0,
Then ∃n0
∊
ℕ,
∀n
∊
ℕ,
n ≥
n0 ⇒ 15n2 < c (3
x 2n)
Let n0 be such that ∀n
∊
ℕ,
n ≥
n0 ⇒ 15n2 < c (3
x 2n), and n’ = max (B, n0)
Since B and n0 are both ∊
ℕ:
Then n’ ∊
ℕ
Then n’ ≥
B #By Definition of max
Then 15n’2 < c3 x 2n’ #By the lines above, since n’ ≥
n0
Now let 2 functions g(n’) = 15n’2, and f(n’) = 3 x 2n’
Then n’ ≥
B ⋀
g(n’) < c f(n’) # ⋀ intro, g(n’) =
15n’2, and f(n’) = 3 x 2n’
Then ∃n∊
ℕ,
n ≥
B ⋀
g(n) < c f(n) # ∃ intro
Conclude ∀c
∊
ℝ,
∀B∊
ℕ,
∃n∊
ℕ,
n ≥
B ⋀
g(n) < c f(n) # ∀ intro
I have now proved
the contrapositive, thus disproving the original statement.
4. LOOKING BACK
Let’s check the
answer by observing the 2 functions at xs with high expected y, say x = 10:
15(10) 2 =
1500.
3(2)10 =
3072.
And x = 11:
15(11) 2 =
1815.
3(2)10 =
6144.
So it is obvious
that 3(2)10 is NOT the lower bound for 15(10) 2 because
3(2)10 is obviously growing at a much faster rate than 15(10) 2.
Although there may
be other solutions to solve this problem, I think this method, using limits is
the easiest and cleanest method.