1: \begin{abstract}

2: We investigate the application of Weak Poincaré Inequalities (WPI)

3: to Markov chains to study their rates of convergence and to derive

4: complexity bounds. At a theoretical level we investigate the necessity

5: of the existence of WPIs to ensure $\mathrm{L}^{2}$-convergence,

6: in particular by establishing equivalence with the Resolvent Uniform

7: Positivity-Improving (RUPI) condition and providing a counterexample.

8: From a more practical perspective, we extend the celebrated Cheeger's

9: inequalities to the subgeometric setting, and further apply these

10: techniques to study random-walk Metropolis algorithms for heavy-tailed

11: target distributions and to obtain lower bounds on pseudo-marginal

12: algorithms.

13: \end{abstract}

14: