A fresh approach to evaluate performance in distributed parallel genetic algorithms

This work proposes a novel approach to evaluate and analyze the behavior of multi-population parallel genetic algorithms (PGAs) when running on cluster multi-core processors. In particular, we deeply study their numerical computational by proposing mathematical model representing observed performance curves. them, discuss emerging descriptions PGA instead of, e.g., individual isolated results subject visual inspection, for better understanding effects number cores used (scalability), migration policy (the gap, in this paper), features solved problem (type encoding size). The conclusions based real figures models fitting them represent fresh way speed-up, time, effort, allowing comparison few meaningful numeric parameters. represents set beyond usual textual lessons found past works PGAs. It can be as an estimation tool future finding out upper limit if increases.