10 problems

Optimal-runtime conjecture. The true runtime for optimal parameters is $O(n\log n)$, and this runtime is achieved with the standard boundaries $1/n$ and $1-1/n$.

Binary Value hardness conjecture. Binary Value is the hardest linear function for optimization by evolutionary algorithms.

Let SDBV (Switching Dynamic Binary Value) be the dynamic monotone function defined in the paper, and let $T_f(n)$ denote the expected number of generations required by the $(1+1)$-…

Let a dynamic monotone function be a monotone fitness function whose fitness landscape may change between generations, and let Adversarial Dynamic BinVal (ADBV) denote the construc…

Let a linear pseudo-Boolean function be a function on $\{0,1\}^n$ of the form … with positive weights $w_i$. Let BinVal denote the linear function whose weights are exponentially d…

Consider the self-adjusting $(1,\lambda)$-EA with mutation probability $c/n$, success rate $s$, and update strength $F>1$, optimizing dynamic monotone functions. No-phase-transitio…

Let the self-adjusting $(1,\lambda)$-EA have success rate $s$ and update strength $F>1$. Middle-range dependence conjecture. There is a middle range of values of $s$ for which whet…

Let the self-adjusting $(1,\lambda)$-EA have success rate $s$ and update strength $F>1$. Non-universal threshold conjecture. There is no threshold $s_0$ such that the algorithm is…

Consider the self-adjusting $(1,\lambda)$-EA with mutation probability $c/n$, success rate $s$, and update strength $F>1$, optimizing dynamic monotone functions. Linear-evaluation…

Let the self-adjusting $(1,\lambda)$-EA use mutation probability $c/n$, success rate $s$, and update strength $F>1$. An efficiency threshold conjecture asserts that there does not…