Multiple Objective Evolution Strategies (MOES): A Users Guide to Running the Software

Report No. ARL-CR-0753
Authors: James Lill; Anthony Yau
Date/Pages: November 2014; 32 pages
Abstract: A users guide for the parallel Multiple Objective Evolution Strategies (MOES) software package is presented. MOES employs a sophisticated self-adaptive evolutionary algorithm known as Evolution Strategies. The software can perform single objective optimization (with and without constraints) as well as multiple objective optimization using a fitness function based on Pareto dominance. The novel multiple-objective fitness function is computed using the concept of efficiency from Data Envelopment Analysis (DEA), a specialized application of linear programming. MOES is unique in combining a very flexible self-adaptive algorithm with a novel multiple-objective algorithm to compute Pareto fitness, all within a package that has been efficiently parallelized.
Distribution: Approved for public release
  Download Report ( 0.247 MBytes )
If you are visually impaired or need a physical copy of this report, please visit and contact DTIC.

Last Update / Reviewed: November 1, 2014