Parallel evolutionary algorithms can achieve super-linear performance

Abstract

One of the main reasons for using parallel evolutionary algorithms (PEAs) is to obtain efficient algorithms with an execution time much lower than that of their sequential counterparts in order, e.g., to tackle more complex problems. This naturally leads to measuring the speedup of the PEA. PEAs have sometimes been reported to provide super-linear performances for different problems, parameterizations, and machines. Super-linear speedup means that using "m" processors leads to an algorithm that runs more than "m" times faster than the sequential version. However, reporting super-linear speedup is controversial, especially for the "traditional" research community, since some non-orthodox practices could be thought of being the cause for this result. Therefore, we begin by offering a taxonomy for speedup, in order to clarify what is being measured. Also, we analyze the sources for such a scenario in this paper. Finally, we study an assorted set of results. Our conclusion is that super-linear performance is possible for PEAs, theoretically and in practice, both in homogeneous and in heterogeneous parallel machines. 

Parallel Ant Colony Optimization: Algorithmic Models and Hardware Implementations

“…An effort is made to keep the algorithm and parameters as close as possible to the original MMAS. Following the guidelines of Barr and Hickman [36] and Alba [37], the relative speedup metric is computed on mean execution times to evaluate the performance of the proposed implementation. Speedups are calculated by dividing the sequential CPU time with the parallel time, which is obtained with the same CPU and the GPU acting as a co-processor.…”

Homogeneous and Heterogeneous Island Models for the Set Cover Problem

“…The speedup of an island model with µ islands is defined as the rate between the expected parallel running time of the island model and the expected running time of the same EA using only a single island. This kind of speedup is called weak orthodox speedup in Alba's taxonomy [17]. If the speedup is of order Θ(µ), we speak of a linear speedup.…”

A new initialization procedure for the distributed estimation of distribution algorithms

“…Because EAs are inherently parallel, the so-called parallel evolutionary algorithms (pEAs) have been studied as an alternative to tackle one of the aspects of this drawback (Alba 2002). A successful example of parallel evolutionary algorithms is the model called distributed evolutionary algorithms (dEAs) or island model.…”