Hello Richard,
Yes the formula is correct, it is
xi•(x1 + … + xn) - ai•(x1 + … + xn) + bi•xi == 0, i = 1 to n.
, so bi*xi and not just bi. One solution is xi = 0 for all i = 1...n.
I also found a formula for c[2]:
c[2] == a[n]^(n - 2) Product[b[i], {i, 1, n}] ((a[n] n)/b[n] + (b[n] - a[n]) Sum[1/b[i], {i, 1, n - 1}] - (n - 1) (1 - Sum[a[i]/b[i], {i, 1, n - 1}]) - b[n] (Sum[a[i]/b[i], {i, 1, n - 1}] Sum[1/b[i], {i, 1, n - 1}] - Sum[a[i]/b[i]^2, {i, 1, n - 1}]))
which is complicated, but the number of operations to calculate it is of order n and not n^2.
I assume that there are also simple formulas for the other c[i], because when I extract all powers of an, bn, ai, bi in the coefficients c[i], I see a pattern that is similar for all c[i], see attached notebook. It gives a table for all terms that appear in the coefficients.
The columns are the powers of ai, bi (all i including n), an, bn, and the coefficient in front of each term. The last column gives the multiplicity of how many terms of this type appear, which is in this case the number of permutations that the indices i different from n allow.
E.g. for n=5, c[5] is
+b[4]^3
-b[1] b[4]^2
-b[2] b[4]^2
-b[3] b[4]^2
+b[1] b[2] b[4]
+b[1] b[3] b[4]
+b[2] b[3] b[4]
-b[1] b[2] b[3]
which gives the table entries
0 3 0 3 1 1
0 3 0 2 -1 3
0 3 0 1 1 3
0 3 0 0 -1 1
It is easy to find a rule for how all these terms are constructed for any n; the binomial coefficients in column 5 and their sign are determined by the powers of the a's and b's in columns 1-4. But as discussed above, my goal is to find a simpler rule for constructing them, with a lower number of operations.
Best,
Max
Attachments: