[Paper Review] A Novel Meta-Heuristic Optimization Algorithm Inspired by the Spread of Viruses

Zhixi Li, Vincent Tam|arXiv (Cornell University)|Jun 11, 2020

Metaheuristic Optimization Algorithms Research27 references7 citations

TL;DR

This paper proposes Virus Spread Optimization (VSO), a novel nature-inspired meta-heuristic algorithm that models viral infection dynamics to solve continuous optimization problems. VSO uses viral RNA as solution representation and employs infection, mutation, and recovery operations to explore solution space adaptively, while an imported infection mechanism enhances global search and avoids premature convergence, outperforming state-of-the-art algorithms on benchmark functions and real-world problems like portfolio optimization and SVM hyperparameter tuning.

ABSTRACT

According to the no-free-lunch theorem, there is no single meta-heuristic algorithm that can optimally solve all optimization problems. This motivates many researchers to continuously develop new optimization algorithms. In this paper, a novel nature-inspired meta-heuristic optimization algorithm called virus spread optimization (VSO) is proposed. VSO loosely mimics the spread of viruses among hosts, and can be effectively applied to solving many challenging and continuous optimization problems. We devise a new representation scheme and viral operations that are radically different from previously proposed virus-based optimization algorithms. First, the viral RNA of each host in VSO denotes a potential solution for which different viral operations will help to diversify the searching strategies in order to largely enhance the solution quality. In addition, an imported infection mechanism, inheriting the searched optima from another colony, is introduced to possibly avoid the prematuration of any potential solution in solving complex problems. VSO has an excellent capability to conduct adaptive neighborhood searches around the discovered optima for achieving better solutions. Furthermore, with a flexible infection mechanism, VSO can quickly escape from local optima. To clearly demonstrate both its effectiveness and efficiency, VSO is critically evaluated on a series of well-known benchmark functions. Moreover, VSO is validated on its applicability through two real-world examples including the financial portfolio optimization and optimization of hyper-parameters of support vector machines for classification problems. The results show that VSO has attained superior performance in terms of solution fitness, convergence rate, scalability, reliability, and flexibility when compared to those results of the conventional as well as state-of-the-art meta-heuristic optimization algorithms.

Motivation & Objective

• To address the limitations of existing meta-heuristic algorithms in solving complex, continuous optimization problems with high scalability and reliability.
• To develop a novel nature-inspired optimization framework that mimics the spread of viruses among hosts to enhance global search and avoid premature convergence.
• To improve solution quality and convergence speed through adaptive neighborhood search around local and global optima.
• To validate the algorithm's effectiveness on standard benchmark functions and real-world applications such as financial portfolio optimization and SVM hyperparameter tuning.
• To demonstrate VSO's superiority over conventional and state-of-the-art meta-heuristic algorithms in terms of fitness, convergence rate, scalability, and flexibility.

Proposed method

• VSO models each host as a potential solution, with viral ribonucleic acid (RNA) representing the solution vector in the search space.
• The algorithm employs three core operations: infection (solution update), mutation (diversity enhancement), and recovery (solution preservation) to explore and refine solutions.
• An imported infection mechanism allows colonies to inherit optima from other colonies, promoting global exploration and preventing premature convergence.
• VSO dynamically conducts adaptive neighborhood searches around discovered local and global optima to refine solutions and improve convergence.
• The algorithm uses a flexible infection mechanism that enables rapid escape from local optima to explore alternative regions of the search space.
• A new representation scheme and viral operations are introduced that are fundamentally different from prior virus-based optimization approaches.

Experimental results

Research questions

• RQ1 Can a virus-inspired meta-heuristic algorithm effectively solve complex continuous optimization problems with high convergence speed and solution quality?
• RQ2 How does the imported infection mechanism in VSO enhance global search and prevent premature convergence compared to standard meta-heuristics?
• RQ3 To what extent does VSO outperform established algorithms like PSO, GA, and DE on standard CEC benchmark functions?
• RQ4 Can VSO achieve superior performance in real-world applications such as financial portfolio optimization and SVM hyperparameter tuning?
• RQ5 How does VSO’s adaptive neighborhood search strategy contribute to improved convergence and solution reliability?

Key findings

• VSO achieved superior performance on 30-dimensional CEC17 and CEC18 benchmark functions, outperforming state-of-the-art algorithms in terms of solution fitness and convergence rate.
• On the CEC2018 benchmark suite, VSO demonstrated strong scalability and reliability across all 30 test functions, including hybrid and composition functions.
• In financial portfolio optimization, VSO achieved a higher Sharpe ratio and lower risk-adjusted return compared to conventional and meta-heuristic methods.
• For SVM hyperparameter tuning, VSO reduced classification error rates and improved model generalization compared to PSO and DE on multiple UCI datasets.
• VSO showed faster convergence and better solution diversity, particularly in escaping local optima due to its flexible infection and imported infection mechanisms.
• The algorithm maintained high reliability and flexibility across diverse problem types, including unimodal, multimodal, and shifted functions.