Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Novel Virus Diffusion Optimization (VDO) Algorithm for Global Optimization

Zhaoqi Sun, Qingsong Wang

TL;DR

The paper addresses global optimization in high-dimensional spaces by introducing the Virus Diffusion Optimizer (VDO), an HSV-inspired metaheuristic that balances exploration and exploitation through Tropism Exploration, Burst Replication, Virion Diffusion, and Latency Reactivation. The algorithm integrates adaptive step-size control and archive-driven memory to sustain diversity and prevent premature convergence. Through extensive experiments on the CEC2017 and CEC2022 benchmarks and on engineering design problems (PVD, TTD, WBD), VDO demonstrates superior convergence speed, solution quality, and robustness compared to 11 baseline methods, with publicly available code for reproducibility. Overall, VDO provides a scalable, biology-inspired approach to tackling complex, large-scale optimization tasks with practical impact in engineering and computational intelligence.

Abstract

Meta-heuristic algorithms are widely used to tackle complex optimization problems, including nonlinear, multimodal, and high-dimensional tasks. However, many existing methods suffer from premature convergence, limited exploration, and performance degradation in large-scale search spaces. To overcome these limitations, this paper introduces a novel Virus Diffusion Optimizer (VDO), inspired by the life-cycle and propagation dynamics of herpes-type viruses. VDO integrates four biologically motivated strategies, including viral tropism exploration, viral replication step regulation, virion diffusion propagation, and latency reactivation mechanism, to achieve a balanced trade-off between global exploration and local exploitation. Experiments on standard benchmark problems, including CEC 2017 and CEC 2022, demonstrate that VDO consistently surpasses state-of-the-art metaheuristics in terms of convergence speed, solution quality, and scalability. These results highlight the effectiveness of viral-inspired strategies in optimization and position VDO as a promising tool for addressing large-scale, complex problems in engineering and computational intelligence.To ensure reproducibility and foster further research, the source code of VDO is made publicly available.

A Novel Virus Diffusion Optimization (VDO) Algorithm for Global Optimization

TL;DR

The paper addresses global optimization in high-dimensional spaces by introducing the Virus Diffusion Optimizer (VDO), an HSV-inspired metaheuristic that balances exploration and exploitation through Tropism Exploration, Burst Replication, Virion Diffusion, and Latency Reactivation. The algorithm integrates adaptive step-size control and archive-driven memory to sustain diversity and prevent premature convergence. Through extensive experiments on the CEC2017 and CEC2022 benchmarks and on engineering design problems (PVD, TTD, WBD), VDO demonstrates superior convergence speed, solution quality, and robustness compared to 11 baseline methods, with publicly available code for reproducibility. Overall, VDO provides a scalable, biology-inspired approach to tackling complex, large-scale optimization tasks with practical impact in engineering and computational intelligence.

Abstract

Meta-heuristic algorithms are widely used to tackle complex optimization problems, including nonlinear, multimodal, and high-dimensional tasks. However, many existing methods suffer from premature convergence, limited exploration, and performance degradation in large-scale search spaces. To overcome these limitations, this paper introduces a novel Virus Diffusion Optimizer (VDO), inspired by the life-cycle and propagation dynamics of herpes-type viruses. VDO integrates four biologically motivated strategies, including viral tropism exploration, viral replication step regulation, virion diffusion propagation, and latency reactivation mechanism, to achieve a balanced trade-off between global exploration and local exploitation. Experiments on standard benchmark problems, including CEC 2017 and CEC 2022, demonstrate that VDO consistently surpasses state-of-the-art metaheuristics in terms of convergence speed, solution quality, and scalability. These results highlight the effectiveness of viral-inspired strategies in optimization and position VDO as a promising tool for addressing large-scale, complex problems in engineering and computational intelligence.To ensure reproducibility and foster further research, the source code of VDO is made publicly available.

Paper Structure

This paper contains 21 sections, 34 equations, 13 figures, 6 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Virus Diffusion Optimization
  3. Inspired by HSV
  4. Mathematical model and optimization algorithm
  5. Viral Tropism Exploration
  6. Viral Replication Step Regulation
  7. Virion Diffusion Propagation
  8. (a) Budding-like update.
  9. (b) Fusion or receptor-targeted jump.
  10. (c) Multidirectional release and Lévy-based reinfection.
  11. (d) Differential recombination and boundary reflection.
  12. Latency Reactivation Mechanism
  13. Algorithm framework
  14. Experiments and Results
  15. Experimental method on CEC2017
  16. ...and 6 more sections

Figures (13)

  • Figure 1: Viral Tropism Exploration.
  • Figure 2: A comparison between the proposed adaptive method (red curve) and the conventional adaptive method (blue curve) for $\beta_t$.
  • Figure 3: Virion Diffusion Propagation.
  • Figure 4: Latency Reactivation Mechanism.
  • Figure 5: The overall flowchart of the Virus Diffusion Optimizer (VDO) algorithm.
  • ...and 8 more figures