1: \begin{abstract}

2:      \cite{GOMEZ201953} proposes a formal and systematic approach for characterizing stochastic global optimization algorithms. Using it, Gomez formalizes algorithms with a fixed next-population stochastic method, i.e., algorithms defined as stationary Markov processes. These are the cases of standard versions of hill-climbing, parallel hill-climbing, generational genetic, steady-state genetic, and differential evolution algorithms. This paper continues such a systematic formal approach. First, we generalize the sufficient conditions convergence lemma from stationary to non-stationary Markov processes. Second, we develop Markov kernels for some selection schemes. Finally, we formalize both simulated-annealing and evolutionary-strategies using the systematic formal approach.

3:     \keywords{ Evolutionary Algorithms \and Non-stationary Markov Kernel \and Convergence Analysis \and Evolutionary Strategies \and Simulated Annealing \and Selection Mechanism}

4:     \subclass{MSC 68T20 \and MSC 65K10}

5: \end{abstract}

6: