A characterization of the eigenvalue of a general (min, max, +)-system (Q5943165): Difference between revisions

General \((\min,\max,+)\)-systems are introduced as described by \[ z(k+1)= F(z(k)),\quad k=0,1,2,\dots, \] where the function \(F:\mathbb{R}^n \to\mathbb{R}^n\) is composed with three kinds of operations, namely addition, maximization and minimization. A system belonging to this class is said to have an eigenvalue \(\lambda\in \mathbb{R}\), if there exists a vector \(\nu\in \mathbb{R}^n\) such that \(F(\nu)= \lambda+\nu\), where the expression on the right hand side stands for the vector obtained by adding \(\lambda\) to each component of \(\nu\). The main result of the paper consists in the characterization of the eigenvalue \(\lambda\) (provided it exists) as the smallest number ensuring the convergence for \(k\to \infty\) of the iteration \(z(k+1)= (F(z(k))- \lambda) \oplus 0\), \(z(0)=0\), where \(\oplus\) denotes the componentwise maximization for vectors of the same size. Conditions for the structural existence of the eigenvalue of a general \((\min,\max,+)\)-system are also presented.