I was studying the sequence A249665, which I will call $a_n$, and came up with a second sequence $b_n$ for which I could prove that the following recurrence holds. \begin{align} a_n&=a_{n-1}+a_{n-4}+a_{n-5}+b_{n-2}+b_{n-3} \\b_n&=a_{n-1}+a_{n-2}+a_{n-3}+a_{n-4}+b_{n-2}+b_{n-3}+b_{n-4} \end{align} Here are some initial values. \begin{align} (a_0,...)&=(0,1,1,1,2,6,14,28,56,118,254,541,1140,...) \\(b_0,...)&=(0,0,1,2,4,8,17,37,79,166,349,738,1563,...) \end{align}
On the OEIS they give a linear recurrence in three sequence, so when I saw that I felt good about myself for finding one in only two sequences. However a bit further on, they give a recurrence in $a_n$ only, which caught me by surprise. $$a_n=2a_{n-1}-a_{n-2}+2a_{n-3}+a_{n-4}+a_{n-5}-a_{n-7}-a_{n-8}$$ The companion matrix of this recurrence has the same dimension as the one for my recurrence, which leads me to believe there should be some method to find this recurrence out of mine.
So is there a general method to combine two simultaneous linear recurrences into one? Or can you at least find a way to deduce the one from OEIS from the simultaneous one I found?