Existence of approximate fixed point

33 Views Asked by At

In K.Urai's paper "Fixed point theorems and the existence of economic equilibria based on conditions for local directions of mappings", he claimed the following in Lemma 17:

Statement: Let $X\subset \mathbb{R}^n$ be a nonempty compact convex subset, and $f:X\to X$ be any function. Then $0$ is in the convex hull of the set $\{f(x)-x|x\in X\}$.

However, his proof is false. He used the following wrong argument: if $C\subset \mathbb{R}^n$ is a nonempty convex subset and $0\notin C$, then there is $p\in \mathbb{R}^n$ such that the inner product $(p,c)>0$ for any $c\in C$.

But the Statement may be true, through. For example, I think I can prove it when $X$ is the closed unit ball or a simplex of $\mathbb{R}^n$. Is the Statement really true for general $X$?

1

There are 1 best solutions below

0
On BEST ANSWER

Here is a proof via induction over the dimension of (the affine hull of) $X$: Suppose that $0$ does not belong to the convex hull of $\{f(x) - x \mid x \in X\}$. Then, there exists $p \in \mathbb R^n \setminus \{0\}$ with $$ p^\top f(x) \ge p^\top x \qquad\forall x \in X. $$ We set $$ \hat X = \{ x \in X \mid p^\top x = \max_{y \in X} p^\top y\}.$$ Clearly, $\hat X$ is nonempty, closed and convex and the dimension is less than the dimension of $X$. Moreover, the above inequality ensures $$ p^\top f(x) \ge p^\top x = \max_{y \in X} p^\top x \qquad\forall x \in \hat X,$$ i.e., $f \colon \hat X \to \hat X$. By the induction hypothesis, $0$ belongs to the convex hull of $$\{f(x) - x \mid x \in \hat X\}$$ and this is a subset of the convex hull of $$\{f(x) - x \mid x \in X\}.$$ This finishes the induction step.