Strong law and central limit theorem for a process between maxima and sums (Q803638)

We prove an invariance principle for the random process \((X_ n)_{n\geq 1}\) given by \[ X_ 1=x\in {\mathbb{R}},\quad X_{n+1}=\max (X_ n,\alpha_ nX_ n+Y_ n),\quad n\geq 1, \] where \((Y_ n)_{n\geq 1}\) are i.i.d. random variables and \((\alpha_ n)_{n\geq 1}\) are nonrandom numbers tending upward to 1 (both in \({\mathbb{R}})\). This process interpolates between maxima \((\alpha_ n\equiv 0)\) and sums \((\alpha_ n\equiv 1)\). Depending on the distribution of \(Y_ n\) and on the rate at which \(\alpha_ n\to 1\) the scaling behaviour exhibits different regimes. Our techniques are flexible and are applicable to more general types of iterative schemes.