Eduardo S. answered 10/07/21
Former math teacher and current actuary who loves teaching math!
This was what I was thinking only after I started writing down a few terms:
a2 = 3(1) + 2 = 3(1) + 2
a3 = 3(3 + 2) = 32(1) + 2(3) + 3
a4 = 3(32 + 2(3) + 3) +4 = 33(1) + 32(2) + 3(3) + 4
a5 = 3[33(1) + 32(2) + 3(3) + 4] = 34(1) + 33(2) + 32(3) + 3(4) + 5
I didn't start writing the terms like that, but when trying to look for a pattern, it helped.
From the above, my guess was that for any n ≥ 2 :
an = 1*3n-1 + 2*3n-2 + ... (n-1)*3 + n or just
n
an = ∑ j*3(n-j) (Is there an easy way to type a summation with the limits using Wyzant?)
j=1
check if it works for n=2
2
a2 = ∑ j * 3(n-j) = 1*3(2-1) + 2*3(2-2) = 3 + 2 = 5 Works for n = 2
j=1
k
Assume true for any n = k , so ak = ∑ j*3(k-j) is true by our assumption.
j = 1
k+1
Show true for n = k + 1 i.e. show ak+1 = ∑ j*3(k+1) - j
j=1
By definition of the sequence, we were given:
ak+1 = 3ak + (k +1) ;
using our assumption
k
ak+1 = 3[ ∑ j * 3(k-j)] +(k+1)
j=1
I will write out some terms of the summation to see if it helps
ak+1 = 3[ 1*3k-1 + 2*3k-2 + ...(k-1)*3 + k ] + (k + 1)
Distribute the 3
ak+1 = 1*3k + 2*3k-1 + ... (k-1)*32 + k*3 + (k+1)
But this is the same as
k +1
ak+1 = ∑ j*3(k+1) - j (Please check if you agree)
j = 1
which is what we wanted to show.
Note: Yes, you can use summation formulas, or possibly even derive using a double summation, a closed
n
form of ∑ j*3(n - j)
j = k
Maybe the answer for this question requires this, but I wasn't nearly as interested in doing that , and I think the induction is more natural this way.

Adam B.
10/10/21