Suppose in this question that $X$ additionally admits a finite variance $V$. Show then that $$\forall \varepsilon > 0, \forall n \geqslant 1, \quad \mathbb{P}\left(S_n \leqslant n(m-\varepsilon)\right) \leqslant \frac{V}{\varepsilon^2 n}.$$
Let $Y$ be a random variable taking values in $\mathbb{N}$ almost surely, and which admits an expectation. Show that $$\mathbb{E}(Y) = \sum_{k=1}^{+\infty} \mathbb{P}(Y \geqslant k)$$
Show that for all $n \in \mathbb{N}$ and $\ell \geqslant 0$, $$\mathbb{P}\left(S_n \leqslant \ell\right) \leqslant \mathbb{E}\left(\exp\left(\ell - S_n\right)\right)$$ then that $$\mathbb{P}\left(S_n \leqslant \ell\right) \leqslant e^{\ell} \mathbb{E}(\exp(-X))^n$$
Deduce that $\mathbb{P}\left(S_n \leqslant \ell\right)$ tends to 0 when $n \rightarrow +\infty$ and that $$\mathbb{E}(N(0,\ell)) \leqslant \frac{e^{\ell}}{1 - \mathbb{E}(\exp(-X))}$$
Show that for all $x \in \mathbb{R}, \ell \geqslant 0, k \in \mathbb{N}^*$ and $n \in \mathbb{N}^*$, $$\mathbb{P}\left(S_{n-1} < x \leqslant S_n, N(x, x+\ell) \geqslant k\right) \leqslant \mathbb{P}\left(S_{n-1} < x \leqslant S_n\right) \mathbb{P}(N(0,\ell) \geqslant k)$$ then that $$\mathbb{E}(N(x, x+\ell)) \leqslant \frac{e^{\ell}}{1 - \mathbb{E}(\exp(-X))}$$
Let $K > 0$ and $g : \mathbb{R} \rightarrow \mathbb{R}$ be a positive bounded function with support in $[0, K]$. The sequence of functions $f_n : \mathbb{R} \rightarrow \mathbb{R}$ is defined for $n \geqslant 0$ by $$f_n(x) = \sum_{k=0}^{n} \mathbb{E}\left(g\left(x - S_k\right)\right)$$ Show that for all $x \in \mathbb{R}$, the sequence $\left(f_n(x)\right)_{n \geqslant 0}$ is increasing. We denote by $f(x)$ its limit in $\mathbb{R} \cup \{+\infty\}$.
Let $K > 0$ and $g : \mathbb{R} \rightarrow \mathbb{R}$ be a positive bounded function with support in $[0, K]$. The sequence of functions $f_n : \mathbb{R} \rightarrow \mathbb{R}$ is defined for $n \geqslant 0$ by $$f_n(x) = \sum_{k=0}^{n} \mathbb{E}\left(g\left(x - S_k\right)\right)$$ Show that if $g = \mathbb{1}_{[0,K]}$, then $f(x) = \mathbb{E}(N(x-K, x))$.
Let $K > 0$ and $g : \mathbb{R} \rightarrow \mathbb{R}$ be a positive bounded function with support in $[0, K]$. The sequence of functions $f_n : \mathbb{R} \rightarrow \mathbb{R}$ is defined for $n \geqslant 0$ by $$f_n(x) = \sum_{k=0}^{n} \mathbb{E}\left(g\left(x - S_k\right)\right)$$ Deduce that for all $x \in \mathbb{R}$ and $n \in \mathbb{N}$, $$0 \leqslant f_n(x) \leqslant \|g\|_{\infty} \frac{e^K}{1 - \mathbb{E}(\exp(-X))}$$
Let $K > 0$ and $g : \mathbb{R} \rightarrow \mathbb{R}$ be a positive bounded function with support in $[0, K]$. The sequence of functions $f_n : \mathbb{R} \rightarrow \mathbb{R}$ is defined for $n \geqslant 0$ by $$f_n(x) = \sum_{k=0}^{n} \mathbb{E}\left(g\left(x - S_k\right)\right)$$ Conclude that the sequence of functions $f_n$ converges pointwise to a positive bounded function $f$ whose support is included in $\mathbb{R}^+$.
Let $Y$ be a discrete random variable, independent of $X$, and $\varphi : \mathbb{R}^2 \rightarrow \mathbb{R}$ be a bounded function. Show that $$\mathbb{E}(\varphi(X, Y)) = \sum_{i=0}^{+\infty} p_i \mathbb{E}\left(\varphi\left(x_i, Y\right)\right)$$
Let $K > 0$ and $g : \mathbb{R} \rightarrow \mathbb{R}$ be a positive bounded function with support in $[0, K]$. The sequence of functions $f_n : \mathbb{R} \rightarrow \mathbb{R}$ is defined for $n \geqslant 0$ by $$f_n(x) = \sum_{k=0}^{n} \mathbb{E}\left(g\left(x - S_k\right)\right)$$ Show that for all $n \in \mathbb{N}$ and $x \in \mathbb{R}$, $$f_{n+1}(x) = g(x) + \sum_{i=0}^{+\infty} p_i f_n\left(x - x_i\right)$$
Let $K > 0$ and $g : \mathbb{R} \rightarrow \mathbb{R}$ be a positive bounded function with support in $[0, K]$. The function $f$ is the pointwise limit of the sequence $f_n$. Show that the function $f$ satisfies the following equality on $\mathbb{R}$ $$f(x) = g(x) + \sum_{i=0}^{+\infty} p_i f\left(x - x_i\right) \tag{E}$$
Let $h : \mathbb{R} \rightarrow \mathbb{R}$ be a bounded function that satisfies $h(x) = \sum_{i=0}^{+\infty} p_i h\left(x - x_i\right)$ for all $x \in \mathbb{R}$. Show that for all $x \in \mathbb{R}$ and $n \in \mathbb{N}$, we have $h(x) = \mathbb{E}\left(h\left(x - S_n\right)\right)$.
Let $h : \mathbb{R} \rightarrow \mathbb{R}$ be a bounded function that satisfies $h(x) = \sum_{i=0}^{+\infty} p_i h\left(x - x_i\right)$ for all $x \in \mathbb{R}$, and such that for all $x \in \mathbb{R}$ and $n \in \mathbb{N}$, $h(x) = \mathbb{E}\left(h\left(x - S_n\right)\right)$. Deduce that if moreover the support of $h$ is included in $\mathbb{R}^+$, then for all $x \in \mathbb{R}$, $h(x) = 0$.
Conclude that there exists a unique bounded function with support in $\mathbb{R}^+$ solution of $$f(x) = g(x) + \sum_{i=0}^{+\infty} p_i f\left(x - x_i\right) \tag{E}$$
Show that the set $\Lambda_X := \bigcup_{n \in \mathbb{N}} \left\{y \in \mathbb{R} \mid \mathbb{P}\left(S_n = y\right) > 0\right\}$ is countable and included in $\mathbb{R}^+$.
We are given an enumeration of the set $\Lambda_X = \{y_i \mid i \in \mathbb{N}\}$. Show that for all $x \in \mathbb{R}$, $$f_n(x) = \sum_{k=0}^{n} \sum_{i=0}^{+\infty} \mathbb{P}\left(S_k = y_i\right) g\left(x - y_i\right)$$
Let $\Lambda$ be a non-empty subset of $\mathbb{R}_*^+$ such that $$\forall (x, y) \in \Lambda^2, \quad x + y \in \Lambda$$ We say that $\Lambda$ is closed under addition. Show that if $(x, y) \in \Lambda^2$, $(k, n) \in \mathbb{N} \times \mathbb{N}^*$ and $k \leqslant n$, then $nx + k(y-x) \in \Lambda$.
Let $\Lambda$ be a non-empty subset of $\mathbb{R}_*^+$ closed under addition. We define $$\Gamma = \left\{z \in \mathbb{R}_+^* \mid \exists (x, y) \in \Lambda,\, z = y - x\right\}, \quad \text{and} \quad r(\Lambda) = \inf \Gamma.$$ Give two examples of such sets $\Lambda$, one for which $r(\Lambda) > 0$ and another for which $r(\Lambda) = 0$.
Let $\Lambda$ be a non-empty subset of $\mathbb{R}_*^+$ closed under addition, with $$\Gamma = \left\{z \in \mathbb{R}_+^* \mid \exists (x, y) \in \Lambda,\, z = y - x\right\}, \quad r(\Lambda) = \inf \Gamma.$$ We assume that $r(\Lambda) > 0$. Show that there exist $(a, b) \in \Lambda^2$ such that $b - a \in [r(\Lambda), 2r(\Lambda)[$.
Let $\Lambda$ be a non-empty subset of $\mathbb{R}_*^+$ closed under addition, with $r(\Lambda) > 0$. Let $(a, b) \in \Lambda^2$ such that $b - a \in [r(\Lambda), 2r(\Lambda)[$ and denote $d = b - a$. Let $k, n \in \mathbb{N}$ such that $k \leqslant n-1$. Show that $$\Lambda \cap [na + kd,\, na + (k+1)d] = \{na + kd,\, na + (k+1)d\}$$
Let $\Lambda$ be a non-empty subset of $\mathbb{R}_*^+$ closed under addition, with $r(\Lambda) > 0$. Let $(a, b) \in \Lambda^2$ such that $b - a \in [r(\Lambda), 2r(\Lambda)[$ and denote $d = b - a$. Show that there exists $n_0 \in \mathbb{N}$ such that $n_0 a + n_0 d > (n_0 + 1)a$, then that there exists $k \in \mathbb{N}$ such that $a = kd$.
Let $\Lambda$ be a non-empty subset of $\mathbb{R}_*^+$ closed under addition, with $r(\Lambda) > 0$ and $d = b - a$ as defined above. Deduce that $\Lambda \subset d\mathbb{Z}$, where $d\mathbb{Z} = \{kd \mid k \in \mathbb{Z}\}$.
Let $\Lambda$ be a non-empty subset of $\mathbb{R}_*^+$ closed under addition, with $r(\Lambda) = 0$. Let $\eta > 0$. Show that there exists $A \geqslant 0$ such that for all $x > A$, $$\Lambda \cap [x, x + \eta] \neq \varnothing$$
Let $\Lambda$ be a non-empty subset of $\mathbb{R}_*^+$ closed under addition, with $r(\Lambda) = 0$. Let $f : \mathbb{R} \rightarrow \mathbb{R}$ be a uniformly continuous function. Suppose that for every sequence $\left(x_n\right)_{n \geqslant 0}$ with values in $\Lambda$ such that $x_n \rightarrow +\infty$, $f\left(x_n\right) \rightarrow 0$ when $n \rightarrow +\infty$. Show that $f(x) \rightarrow 0$ when $x \rightarrow +\infty$.
We assume that for all $d \geqslant 0$, $\mathbb{P}(X \in d\mathbb{Z}) < 1$. We consider a function $h$ uniformly continuous and bounded on $\mathbb{R}$ such that for all $x \in \mathbb{R}$, $h(x) \leqslant h(0)$ and $$h(x) = \sum_{i=0}^{+\infty} p_i h\left(x - x_i\right)$$ We recall that for all $x \in \mathbb{R}$ and $n \in \mathbb{N}$, $h(x) = \mathbb{E}\left(h\left(x - S_n\right)\right)$. Show that for all $n \in \mathbb{N}$ and $x \geqslant 0$ such that $\mathbb{P}\left(S_n = x\right) > 0$, we have $h(-x) = h(0)$.
We assume that for all $d \geqslant 0$, $\mathbb{P}(X \in d\mathbb{Z}) < 1$. Show that the set $\Lambda_X$ defined in question 8a is closed under addition and that $r\left(\Lambda_X\right) = 0$.
We assume that for all $d \geqslant 0$, $\mathbb{P}(X \in d\mathbb{Z}) < 1$. Using the results of questions 13a and 13b, deduce that $h(-x) \rightarrow h(0)$ when $x \rightarrow +\infty$.
We assume that for all $d \geqslant 0$, $\mathbb{P}(X \in d\mathbb{Z}) < 1$, and that $g$ is of class $\mathscr{C}^1$, with support in $[0, K]$ with $K > 0$. The function $f$ is the unique bounded and uniformly continuous solution of equation (E), and $f'$ is bounded and uniformly continuous. Prove that the function $x \mapsto \sup_{t \geqslant x} f'(t)$ admits a finite limit when $x \rightarrow +\infty$. We denote $$c := \lim_{x \rightarrow +\infty} \sup_{t \geqslant x} f'(t)$$
We assume that for all $d \geqslant 0$, $\mathbb{P}(X \in d\mathbb{Z}) < 1$, and that $g$ is of class $\mathscr{C}^1$, with support in $[0, K]$ with $K > 0$. With $c := \lim_{x \rightarrow +\infty} \sup_{t \geqslant x} f'(t)$, show that there exists a sequence $y_n \rightarrow +\infty$ such that $f'\left(y_n\right) \rightarrow c$ when $n \rightarrow +\infty$.