Let $\Bbb F_q$ be a finite field with $\text{ char}\,\Bbb F_q=p$ and $n>0$ an integer with $\text{ gcd}(n, \log_pq)=1$. Let $(\ )^*:\Bbb F_q({ \tt x}_0,\dots,{ \tt x}_{ n-1})\to\Bbb F_q({ \tt x}_0,\dots,{ \tt x}_{ n-1})$ be the $\Bbb F_q$-monomorphism defined by ${ \tt x}_i^*={ \tt x}_{ i+1}$ for $0\le i< n-1$ and ${ \tt x}_{ n-1}^*={ \tt x}_0^q$. For $f,g\in\Bbb F_q({ \tt x}_0,\dots,{ \tt x}_{ n-1})\setminus\Bbb F_q$, define $f\circ g=f(g,g^*,\dots,g^{ (n-1)*})$. Then $(\Bbb F_q({ \tt x}_0,\dots,{ \tt x}_{ n-1})\setminus\Bbb F_q,\,\circ)$ is a monoid whose invertible elements are called global $\mathcal P$-forms. Global $\mathcal P$-forms were first introduced by H. Dobbertin in 2001 with $q=2$ to study certain type of permutation polynomials of $\Bbb F_{ 2^m}$ with $\text{ gcd}(m,n)=1$; global $\mathcal P$-forms with $q=p$ for an arbitrary prime $p$ were considered by W. More in 2005. In this paper, we discuss some fundamental questions about global $\mathcal P$-forms, some of which are answered and others remain open.
Categories: math.NT math.GR