Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Performance Comparison of Surrogate-Assisted Evolutionary Algorithms on Computational Fluid Dynamics Problems

Jakub Kudela, Ladislav Dobrovsky

TL;DR

This paper uses two real-world computational fluid dynamics problems to compare the performance of eleven state-of-the-art single-objective SAEAs and suggests that the more recently published methods, as well as the techniques that utilize differential evolution as one of their optimization mechanisms, perform significantly better than the other considered methods.

Abstract

Surrogate-assisted evolutionary algorithms (SAEAs) are recently among the most widely studied methods for their capability to solve expensive real-world optimization problems. However, the development of new methods and benchmarking with other techniques still relies almost exclusively on artificially created problems. In this paper, we use two real-world computational fluid dynamics problems to compare the performance of eleven state-of-the-art single-objective SAEAs. We analyze the performance by investigating the quality and robustness of the obtained solutions and the convergence properties of the selected methods. Our findings suggest that the more recently published methods, as well as the techniques that utilize differential evolution as one of their optimization mechanisms, perform significantly better than the other considered methods.

Performance Comparison of Surrogate-Assisted Evolutionary Algorithms on Computational Fluid Dynamics Problems

TL;DR

This paper uses two real-world computational fluid dynamics problems to compare the performance of eleven state-of-the-art single-objective SAEAs and suggests that the more recently published methods, as well as the techniques that utilize differential evolution as one of their optimization mechanisms, perform significantly better than the other considered methods.

Abstract

Surrogate-assisted evolutionary algorithms (SAEAs) are recently among the most widely studied methods for their capability to solve expensive real-world optimization problems. However, the development of new methods and benchmarking with other techniques still relies almost exclusively on artificially created problems. In this paper, we use two real-world computational fluid dynamics problems to compare the performance of eleven state-of-the-art single-objective SAEAs. We analyze the performance by investigating the quality and robustness of the obtained solutions and the convergence properties of the selected methods. Our findings suggest that the more recently published methods, as well as the techniques that utilize differential evolution as one of their optimization mechanisms, perform significantly better than the other considered methods.
Paper Structure (12 sections, 2 equations, 7 figures, 5 tables, 1 algorithm)

This paper contains 12 sections, 2 equations, 7 figures, 5 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Surrogate-Assisted Optimization
  3. Selected Problems
  4. PitzDaily
  5. Electrostatic Precipitator
  6. Selected Methods and Experimental Setup
  7. Experimental Setup
  8. Results and Discussion
  9. PitzDaily Results
  10. ESP Results
  11. Aggregate Results
  12. Conclusions

Figures (7)

  • Figure 1: Example of the Catmull-Clark subdivision curve. $S^0$ are the original control points, $S^1$ is the approximation of $\mathcal{C}$ after one iteration, $S^5$ is the visually smooth curve after five iterations.
  • Figure 2: ESP with 3 separation zones (left) and GDS mounted in the inlet hood of an ESP (right) gecco2020.
  • Figure 3: Random design (top) and best-found design (bottom) for the PitzDaily problem (left), flowfield (CFD) simulations of these designs (right).
  • Figure 4: Convergence of the average best-found value of the selected SAEAs on the PitzDaily problem.
  • Figure 5: Boxplots of the solutions of the SAEAs on the PitzDaily problem with $B=500$ (top) and $B=1000$ (bottom).
  • ...and 2 more figures