Minimizing evolutionary algorithms energy consumption in the low-level language Zig

Abstract

Managing energy resources in scientific computing implies awareness of a wide range of software engineering techniques that, when applied, can minimize the energy footprint of experiments. In the case of evolutionary computation, we are talking about a specific workload that includes the generation of chromosomes and operations that change parts of them or access and operate on them to obtain a fitness value. In a low-level language such as Zig, we will show how different choices will affect the energy consumption of an experiment.