Open Access

ARTICLE

Comparative Analysis of Genetic and Quantum-Inspired Optimization for Zero-Trust Microsegmentation in Brownfield Networks

Reen-Cheng Wang¹, Hong-Sheng Wang², Kuo-Chun Tseng^2,*

1 Department of Computer Science and Information Engineering, National Taitung University, Taitung, Taiwan
2 Department of Computer Science and Information Engineering, National Ilan University, Ilan, Taiwan

* Corresponding Author: Kuo-Chun Tseng. Email:

(This article belongs to the Special Issue: Next-Generation Optimization: Quantum and Hybrid Classical Computing for Real-World Applications)

Computers, Materials & Continua 2026, 88(3), 59 https://doi.org/10.32604/cmc.2026.083124

Received 29 March 2026; Accepted 25 May 2026; Issue published 23 July 2026

Abstract

Network microsegmentation has become a key mechanism for enforcing zero-trust architecture in enterprise environments, yet its effectiveness remains closely tied to initialization quality. This study formulates network microsegmentation as a state-dependent combinatorial optimization problem in which optimization behavior depends on the availability of structural guidance. A comparative analysis is conducted across four representative optimization paradigms, including genetic algorithms (GA), differential evolution (DE), particle swarm optimization (PSO), and amplitude-ensemble quantum-inspired tabu search (AE-QTS), under both structured and unstructured conditions. Experiments are conducted on a representative brownfield enterprise network using 30 independent runs per configuration. In addition to cost-based evaluation, a fragmentation metric is used to assess the structural quality and manageability of segmentation outcomes. The results indicate that under structured conditions, GA and AE-QTS achieve the best overall performance, with AE-QTS obtaining the best average objective value of −61.29 and GA demonstrating rapid convergence under limited optimization time. Under unstructured conditions, AE-QTS consistently outperforms all other methods, reducing the average objective value from 344.28 (GA) and 1214.60 (DE) to 4.54 under uniform initialization. Moreover, PSO demonstrates comparatively stable and robust behavior, although its performance remains below that of AE-QTS. These findings suggest that microsegmentation can be more appropriately viewed as a condition-dependent optimization problem, in which different optimization methods exhibit different strengths across operational scenarios. The results provide empirical evidence and practical insights that may support the future development of adaptive or hybrid optimization strategies for real-world deployment environments.

---

Keywords

---