How can I derive closed form of polynomial recursive sequence? Saying we have \[\huge a_n = a_{n-1} + n^2 + 4n,\text{ where }a_0=0\] What is its closed form?
solve the homogenous, then solve for the particular ... difference equations are similar to differential equations
or, define a few terms and see if we get to a constant difference tier
Well, I'm not familiar with differential equations yet. Can you show me exactly how to do second approach?
I'm self learning, so I may need to learn something beforehand, but I dunno.
the most basic approach is to list your first few terms, say 5 up to 10? then see if we can develop the difference teirs
Basically like this?\[a_n = a_{n-1} + n^2+4n = a_{n-2}+(n-1)^2+4(n-1)+ n^2+4n=\cdots\] And sorry, but exactly what is "difference teirs?"
the other way is to define the homogenous \[ a_n - a_{n-1} = 0\] and use it in determining the particular let an = x^n \[ x^n - x^{n-1} = 0\] \[ x^{n-1}( x - 1) = 0~:~x=1\] \[c_0(1)^{1}=c_1(1)^{0}+(1)^2+4(1)\] \[c_0(1)^{2}=c_1(1)^{1}+(2)^2+4(2)\] seems familiar
Um let's just stick to most basic approach because I'm lost already.
well, lets define the terms a0, a1, a2, a3, a4, a5, a6, .. as many as you can muster
By define, you mean find the value of a0, a1, a2, a3, etc?
yes 0 1 2 3 4 ... 0 5 17 38 70 ...
Ok thanks for doing calculations for me :) so what's next?
How many terms should I define up to?
0 | 5 | 17 | 38 | 70 5 12 21 32 7 9 11 2 2 take the differences, and then the differences of the differecnes, and then .... continue doing this in teirs until you come to a constant tier, one where the differences are all the same
ah I see
define a5 and see if we get to a constant of 2 11+2 = 13 +32 = 45 + 70 = 115 is a5 = 115?
70 + 5^2 + 4(5) 70 + 45, it is :)
Yeah I checked
now we can take the values on the front of each row and work it into a formula to find the(n-1)th term on the top
one way is to do a formula, im curious if we can get it this way tho notice that if we were to represent this by a polnomial f(x) = a + bx + cx^2 + dx^3 + ex^4 + ... we have 3 tiers, so the thrid derivative is constant f'''(x) = 2 f''(x) = 2x + k f'(x) = x^2 + kx + j f(x) = 1/3 x^3 + 1/2 kx^2 + jx + m ------------------------------- f(0) = 0, so m=0 ------------------------ f(1) = 5 = 1/3 + 1/2 k + j ------------------------ f(2) = 17 = 1/3 (2^3) + 1/2 k (2^2) + 2j ----------------------- think thisll work? 2 equations in 2 unknowns?
Hmm interesting... Give me couple mins to solve for unknowns.
I got \(k=5\) and \(j=\dfrac{13}{6}\) So closed form would be \(f(n)=\boxed{a_n=\dfrac{1}{3}n^3+\dfrac{5}{2}n^2+\dfrac{13}{6}n}\) Checking: from closed form, \(a_6 = 175\); from recursive form, \(a_6 = 115 + 6^2 + 4(6) = 175\) It seems to work. Awesome!
I'm surprised to see that coefficients are ugly fraction, yet it will always result as integer for \(n\in\mathbb N\)
yeah, i even impressed myself :) i know that differences are related to derivatives, so i took a stab at it
Awesome. Leveled up: math skill: +1 thanks to you! :)
youre welcome :)
Join our real-time social learning platform and learn together with your friends!