MathLabs

Problem 6

Let a1,a2,a3,…a_1,a_2,a_3,\ldots be a sequence of positive real numbers. Suppose that for some positive integer ss, we have an=max⁡{ak+an−k∣1≤k≤n−1}a_n=\max\{a_k+a_{n-k}\mid 1\le k\le n-1\} for all n>sn>s. Prove that there exist positive integers ℓ\ell and NN, with ℓ≤s\ell\le s, such that an=aℓ+an−ℓa_n=a_\ell+a_{n-\ell} for all n≥Nn\ge N.
Step 5 of 9: The normalized sequence obeys the same recurrence
bn=max⁡1≤k≤n−1(bk+bn−k)b_n=\max_{1\le k\le n-1}(b_k+b_{n-k})
Detailed analysis

For n>sn>s, subtracting qnqn from the original recurrence gives bn=max⁡1≤k≤n−1(bk+bn−k)b_n=\max_{1\le k\le n-1}(b_k+b_{n-k}), because qk+q(n−k)=qnqk+q(n-k)=qn. Together with the initial values, this recurrence also gives an expansion of every bnb_n as a sum of terms bib_i with i≤si\le s.