Exercise 7.9

(a)
The formula h(1) = 1, () h(2) = 2, h(n) = [h(n + 1)]2 [h(n 1)]2 for n 2

is not one to which the principle of recursive definition applies. Show that nevertheless there does exist a function h : + satisfying this formula. [Hint: Reformulate (∗) so that the principle will apply and require h to be positive.]

(b)
Show that the formula (∗) of part (a) does not determine h uniquely. [Hint: If h is a positive function satisfying (∗), let f(i) = h(i) for i3, and let f(3) = h(3).]
(c)
Show that there is no function h : + satisfying the formula h(1) = 1, h(2) = 2, h(n) = [h(n + 1)]2 + [h(n 1)]2 for n 2.

Answers

(a) First, notice that (∗) does not satisfy the principle of recursive definition because, for n 2, h(n) is not defined strictly in terms of values of h for positive integers less than n, since its definition depends on h(n + 1). Now we show that there does exist a function satisfying (∗).

Proof. Consider the following reformulation:

h(1) = 1, h(2) = 2, h(n) = h(n 1) + [h(n 2)]2 for n > 2,

where as a convention we take the positive square root for h(n). Clearly for n {1,2} we have that h(n) is positive. Now suppose n > 2 and that h(k) is positive for k < n so that h(n 1) and h(n 2) are both positive. Then clearly h(n 1) + [h(n 2)]2 is positive so that h(n) = h(n 1) + [h(n 2)]2 is defined and is positive. Hence h(n) is positive and well-defined for all n + by induction.

Thus, since h(n) depends only on values of h for integers less than n, this satisfies the recursion principle so that a unique h satisfying the above exists. We also claim that this h satisfies (∗). Clearly, the explicitly defined values of h(1) and h(2) are satisfied. For n 2, we have that n + 1 > 2 so that, by definition,

h(n + 1) = h((n + 1) 1) + [h((n + 1) 2)]2 = h(n) + [h(n 1)]2 [h(n + 1)]2 = h(n) + [h(n 1)]2 h(n) = [h(n + 1)]2 [h(n 1)]2,

which is the final constraint of (∗) so that it is also satisfied since n 2 was arbitrary. □

(b) First note that for the recursively defined function h from part (a),

h(3) = h(2) + [h(1)]2 = 2 + 12 = 3 h(4) = h(3) + [h(2)]2 = 3 + 22 = 3 + 4.

Now define the function f as in the hint, that is f(i) = h(i) for i3 and f(3) = h(3). Then we clearly have f(3) = h(3) = 3 while

[f(4)]2 [f(2)]2 = [h(4)]2 [h(2)]2 = (3 + 4)2 22 = 3 + 4 4 = 3

so that f(n) = 33 = [f(n + 1)]2 [f(n 1)]2 for n = 3, and hence (∗) is violated. So it would seem that the hint as given does not exactly work.

Now we show that the function satisfying (∗) is not unique, taking inspiration from the hint.

Proof. We construct a function f, different from h from part (a), that also satisfies (∗). We define f using recursion:

f(1) = 1, f(2) = 2, f(3) = 3, f(n) = f(n 1) + [f(n 2)]2 for n > 3.

Clearly, since each f(n) is defined only in terms of f(k) for k < n (or without dependence on any values of f), f exists uniquely by the recursion principle so long as each f(n) is well-defined. We show this presently by induction.

Clearly f(n) is defined for n {1,2,3}. For n = 4 we have f(n) = f(4) = f(3) + [f(2)]2 = 3 + 223 + 4. Now, since 1 < 3, we have that 3 < 3 < 4 so that 3 + 4 = 4 3 > 0 and hence the square root, and therefore f(4), is defined and positive. Now consider any n > 4 and suppose that f(n 1) is positive. Then clearly f(n) = f(n 1) + [f(n 2)]2 is defined and positive since f(n 1) > 0, noting that even if f(n 2) 0, its square is non-negative.. This completes the induction that shows that f is uniquely defined.

Clearly fh since f(3) = 33 = h(3). Also obviously f(n) satisfies (∗) explicitly for n {1,2}. For n = 2 we have

[f(n + 1)]2 [f(n 1)]2 = [f(3)]2 [f(1)]2 = [3]2 12 = 3 1 = 2 = f(n).

Then, for n > 2 we have n + 1 > 3 so that, by definition,

f(n + 1) = f((n + 1) 1) + [f((n + 1) 2)]2 = f(n) + [f(n 1)]2 [f(n + 1)]2 = f(n) + [f(n 1)]2 f(n) = [f(n + 1)]2 [f(n 1)]2.

Thus the recursive part of (∗) holds for n 2 so that (∗) holds over the whole domain of f as desired. □

(c)

Proof. Suppose that such a function h does exist. Since the recursive property holds for n 2, we have

h(2) = [h(3)]2 + [h(1)]2 2 = [h(3)]2 + 12 [h(3)]2 = 2 12 = 1 h(3) = ±1.

Similarly, we have

h(3) = [h(4)]2 + [h(2)]2 ± 1 = [h(4)]2 + 22 [h(4)]2 = ±1 22 = ±1 4

so that either [h(4)]2 = 1 4 = 3 or [h(4)]2 = 1 4 = 5. In either case, we have [h(4)]2 < 0, which is of course impossible since the square of a real number is always non-negative! So it must be that such a function does not exist. □

User profile picture
2019-12-01 00:00
Comments