Runtime Analysis of Fitness-Proportionate Selection on Linear Functions
This work provides theoretical insights into evolutionary algorithm performance, which is incremental for algorithm design and biological applications like directed evolution.
The paper analyzes the runtime of non-elitist evolutionary algorithms with fitness-proportionate selection on linear functions, showing that with standard mutation rates, optimization takes exponential time, but reducing the mutation rate or scaling fitness allows polynomial expected runtime.
This paper extends the runtime analysis of non-elitist evolutionary algorithms (EAs) with fitness-proportionate selection from the simple OneMax function to the linear functions. Not only does our analysis cover a larger class of fitness functions, it also holds for a wider range of mutation rates. We show that with overwhelmingly high probability, no linear function can be optimised in less than exponential time, assuming bitwise mutation rate $Θ(1/n)$ and population size $λ=n^k$ for any constant $k>2$. In contrast to this negative result, we also show that for any linear function with polynomially bounded weights, the EA achieves a polynomial expected runtime if the mutation rate is reduced to $Θ(1/n^2)$ and the population size is sufficiently large. Furthermore, the EA with mutation rate $χ/n=Θ(1/n)$ and modest population size $λ=Ω(\ln n)$ optimises the scaled fitness function $e^{(χ+\varepsilon)f(x)}$ for any linear function $f$ and any $\varepsilon>0$ in expected time $O(nλ\lnλ+n^2)$. These upper bounds also extend to some additively decomposed fitness functions, such as the Royal Road functions. We expect that the obtained results may be useful not only for the development of the theory of evolutionary algorithms, but also for biological applications, such as the directed evolution.