The book I'm concerned with is "Integer Points in Polyhedra" by A. Barvinok, which, I must say, is turning out to be highly fascinating.

A real finite-dimensional vector space $V$ defines the following two spaces, also real. $\mathcal{P}(V)$ generated by indicator functions of closed convex polyhedra in $V$ (not necessarily bounded) and $\mathcal{C}_b(V)$ generated by indicator functions of convex compacts in $V$. Let then $$\mathcal{P}_b(V)=\mathcal{P}(V)\cap\mathcal{C}_b(V).$$ This seems to obviously be generated by indicators of bounded closed polyhedra but, just in case, I present the definitions as in the book.

Each of these three spaces can be equipped with a nontrivial multiplicative structure by setting $$\chi_X*\chi_Y=\chi_{X+Y}$$ (Minkowski sum) for two generating indicator functions $\chi_X$ and $\chi_Y$. (That this is a well-defined algebra structure is far from obvious but is shown in the book.)

Now, problem 10 in chapter 3 states: $\mathcal{P_b}(V)$ has no zero-divisors with respect to $*$.

This seems to be wrong in dimension 1:$$(\chi_{[0;1]}-\chi_{\{0\}})*(\chi_{[0;1]}-\chi_{\{1\}})=\chi_{[0;2]}-\chi_{[0;1]}-\chi_{[1;2]}+\chi_{\{1\}}=0.$$

So... Am I missing something? If not, I would suppose that there exists some correct version which was supposed to be here instead. Any idea what that could look like?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.