Ergodicity of a class of nonlinear time series models in random environment domain (Q2267287)

The authors study the problem of a variety of nonlinear time series models, \[ X_{n+1}= T_{Z_{n+1}}(X(n),\dots,X(n- Z_{n+1}), e_{n+1}(Z_{n+1})), \] where \(\{Z_n\}\) is a Markov chain with finite state space, and for every state \(i\) of the Markov chain, \(\{e_n(i)\}\) is a sequence of independent and identically distributed (shortly, i.i.d.) random variables. It is known that various problems of the theory of nonlinear time series and their applications are considered in many papers and books. The authors observe that for adequate modeling of real phenomena (as, for example, in engineering and finance) nonlinear time series with shifts of the arguments (or with delays) and white noise are used. Moreover, the interference with a system will change when the environment changes. The authors introduce a general model with interference from the environments taken into consideration to a certain extent. An irreducible, aperiodic and time-homogeneous Markov chain \(\{Z_n\geq 1\}\), defined on a probability space \((\Omega,{\mathcal F},P)\) with values in a state space \((\mathbb{H},\mathbb{D})\), where \(\mathbb{H}=\{1, 2,\dots,r\}\) is a finite set and \(\mathbb{D}\) is the \(\sigma\)-algebra consisting of all subsets of \(\mathbb{H}\), is considered. Also a measurable space \((\mathbb{R}^m,\mathbb{B}_m)\), where \(\mathbb{R}^m\) is an \(m\)-dimensional real space and \(\mathbb{B}_m\) is the \(\sigma\)-algebra generated by all Borel subsets of \(\mathbb{R}^m\), is considered. Under these conditions, for a sequence \(\{e_n(1)\},\dots,\{e_n(r)\}\) of i.i.d. random vectors, defined on \((\Omega,{\mathcal F},P)\), with state space \((\mathbb{R}^m,\mathbb{B}_m)\), \(e_n(Z_n):= \sum^r_{i=1}\{e_n(i)\} I_{\{i\}}(Z_n),\) where \(I_{\{i\}}(\cdot)\) is an indicator function for a single element set \(\{i\}\) is considered. Then, the model to be investigated is introduced, and it is supposed that the sequence \(\{X(n)\}\) is defined by this model. The authors are interested to study the limit behaviour of the sequence \(\{X(n)\}\), generated iteratively by the model. Finally, the main results are presented and an example is given. This is a useful and interesting paper.