On Tenfold Execution Time in Real World Optimization Problems with Differential Evolution in Perspective of Algorithm Design
Aleš Zamuda, Janez Brest · 2018
This paper presents an algorithm design perspective on tenfold execution time in Real World Industry Challenges (RWIC) optimized using Differential Evolution (DE). The DE is a branch of optimization algorithms that include a set of design decisions to make before committing a finally proposed DE algorithm for a specific RWIC at hand. Therefore, analysis of a well known DE example variant algorithm on RWIC benchmark is reported, based on Competition on Testing Evolutionary Algorithms on Real World Optimization at Congress on Evolutionary Computation (CEC) 2011. The number of fitness functions allowed to execute the optimization runs is reduced or expanded, i.e. tenfolded in ten times of execution limit, and then the corresponding performance of the respective algorithms is compared to performance with the original execution limit. Discussion on aggregated performance with this runtime is then provided in the perspective of new algorithm design.