Hybrid Metaheuristic Frameworks for Multi-Objective Engineering Optimization Problems

Abstract

Hybrid metaheuristic frameworks have emerged as a dominant paradigm in addressing the complexities of multi-objective engineering optimization. Modern engineering design often demands the simultaneous optimization of conflicting objectives—such as minimizing cost while maximizing performance and reliability—under uncertain and nonlinear conditions. Traditional single-objective or standalone metaheuristics often exhibit limitations in exploration–exploitation balance, convergence stability, and robustness against uncertainty. To overcome these challenges, hybrid metaheuristics integrate multiple algorithmic strategies, combining the global exploration power of methods like the Gravitational Search Algorithm (GSA) with the local exploitation capability of techniques such as the Bat Algorithm (BAT), as exemplified by the MOGSABAT framework. This study provides a comprehensive examination of hybrid metaheuristic models for multi-objective optimization, discussing their theoretical underpinnings, mathematical formulations under uncertainty, and empirical performance. A systematic review of algorithmic architectures—including parallel, sequential, and machine-learning-assisted hybrids—is conducted, supported by rigorous statistical evaluation using Wilcoxon signed-rank tests and convergence-diversity metrics. Furthermore, the paper presents a detailed catalogue of metaheuristic algorithms and their hybridization potential for engineering applications. The findings demonstrate that hybrid metaheuristics not only outperform conventional algorithms in convergence speed and solution diversity but also offer enhanced scalability and resilience to data uncertainty. Finally, emerging trends such as adaptive hybridization, integration with machine learning, and parallelized implementations are identified as key directions for advancing future research in robust multi-objective optimization.