Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

An accelerate Prediction Strategy for Dynamic Multi-Objective Optimization

Ru Lei, Lin Li, Rustam Stolkin, Bin Feng

TL;DR

This paper proposes an adaptive prediction strategy that incorporates second-order derivatives to predict and adjust the algorithms search behavior, enhancing the algorithm's ability to anticipate changes in the environment, allowing for more efficient population re-initialization.

Abstract

This paper addresses the challenge of dynamic multi-objective optimization problems (DMOPs) by introducing novel approaches for accelerating prediction strategies within the evolutionary algorithm framework. Since the objectives of DMOPs evolve over time, both the Pareto optimal set (PS) and the Pareto optimal front (PF) are dynamic. To effectively track the changes in the PS and PF in both decision and objective spaces, we propose an adaptive prediction strategy that incorporates second-order derivatives to predict and adjust the algorithms search behavior. This strategy enhances the algorithm's ability to anticipate changes in the environment, allowing for more efficient population re-initialization. We evaluate the performance of the proposed method against four state-of-the-art algorithms using standard DMOPs benchmark problems. Experimental results demonstrate that the proposed approach significantly outperforms the other algorithms across most test problems.

An accelerate Prediction Strategy for Dynamic Multi-Objective Optimization

TL;DR

This paper proposes an adaptive prediction strategy that incorporates second-order derivatives to predict and adjust the algorithms search behavior, enhancing the algorithm's ability to anticipate changes in the environment, allowing for more efficient population re-initialization.

Abstract

This paper addresses the challenge of dynamic multi-objective optimization problems (DMOPs) by introducing novel approaches for accelerating prediction strategies within the evolutionary algorithm framework. Since the objectives of DMOPs evolve over time, both the Pareto optimal set (PS) and the Pareto optimal front (PF) are dynamic. To effectively track the changes in the PS and PF in both decision and objective spaces, we propose an adaptive prediction strategy that incorporates second-order derivatives to predict and adjust the algorithms search behavior. This strategy enhances the algorithm's ability to anticipate changes in the environment, allowing for more efficient population re-initialization. We evaluate the performance of the proposed method against four state-of-the-art algorithms using standard DMOPs benchmark problems. Experimental results demonstrate that the proposed approach significantly outperforms the other algorithms across most test problems.
Paper Structure (17 sections, 15 equations, 4 figures, 3 tables, 3 algorithms)

This paper contains 17 sections, 15 equations, 4 figures, 3 tables, 3 algorithms.

Table of Contents

  1. Introduction
  2. Related Work
  3. Dynamic Multi-objective Problems Definition
  4. Related Dynamic Multi-Objective Evolutionary Algorithms
  5. Motivation for Conducting Second-order Derivative and Adaptive Prediction from Dual-Domain Spaces
  6. THE PROPOSED METHOD
  7. Dynamic Detection
  8. Adaptive Dual-Domain prediction Algorithm
  9. Second-order Derivative Prediction Strategy
  10. EXPERIMENTS
  11. Test Instances
  12. Performance Metrics
  13. Parameter Settings
  14. Experimental Results
  15. MIGD and MHV Values of Final Populations
  16. ...and 2 more sections

Figures (4)

  • Figure 1: Illustrations of the Complex Changes in PS and PF for Two Representative Type II DMOPs:(a) Changing PS of DF12 and DF13. (b) Changing PF of DF12 and DF13.
  • Figure 2: Second-order Derivative-based prediction strategy for DMOPs
  • Figure 3: Flow diagram of the proposed prediction method.
  • Figure 5: The average running times of ADPS-MOEA/D and its four competitors on DF test problems with dynamic configuration C3.