Proof.
Let and to be determined. We will compute .
As before, for any point we are considering, we can always do a unitary coordinate transform which makes .
Under this coordinate, we can compute:
| (6.30) |
|
|
|
Similar to the calculation in Theorem 2.1, we can find:
| (6.31) |
|
|
|
|
| (6.32) |
|
|
|
|
Hence we obtain
| (6.33) |
|
|
|
Now we choose , and we choose sufficiently large so that . Hence we obtain from (6.33):
| (6.34) |
|
|
|
Here depends only on and . Define for . We show that .Indeed,
|
|
|
From this the claim follows easily.
Denote . Suppose the function achieves maximum at . There are two possibilities:
Suppose , then we immediately conclude that
|
|
|
Then we are done.
Suppose otherwise , then we know at :
| (6.35) |
|
|
|
In the third inequality above, we used that , which is true only in dimension 2.
Also we used that at , .
Suppose at , we have , this immediately gives a bound for , hence at . Then we are done.
Suppose otherwise, then we have at
| (6.36) |
|
|
|
Then we also get an estimate for at . So we are done as well.
∎