On nonlinear extensions of the Perron-Frobenius theorem (Q1079340)

The main result of the paper is the following (Theorem 2): If H is a continuous mapping: \(R^ n\to R^ n\), homogeneous, non-negative and primitive (the last means that there exists an integer \(\ell >0\) such that \(H^{\ell}(X)<H^{\ell}(y)\) for \(X\leq y)\) then: (a) there exist \(\lambda >0\), \(X^*\in R^ n\) such that \(H(X^*)=\lambda^*X^*\); \(X^*\) is unique up to a scalar multiple. (b) If \(\lambda =\lambda^*\) there is no \(X\geq 0\) such that \(H(X)=\lambda X\). (c) For any \(X\geq 0\), \(H^ k(X)\) converges to \(X^*\) as \(k\to \infty\). (d) \(\min_ i(H_ i(X)/X_ i)\leq \lambda^*\leq \max_ i(H_ i(X)/X_ i)\). Equality holds if and only if \(X=X^*\). (e) \(\lambda^*>| \lambda |\) for any other eigenvalue of H.