OPTIMIZATION TECHNIQUES IN STRUCTURAL ENGINEERING DESIGN: A REVIEW OF METAHEURISTIC ALGORITHMS AND MACHINE LEARNING APPROACHES

Abstract

Structural engineering design requires the simultaneous optimization of conflicting objectives such as cost, safety, weight, and environmental impact. Traditional design methodologies often rely on empirical approaches and trial-and-error procedures that are computationally inefficient and may not yield globally optimal solutions. In recent decades, metaheuristic algorithms and machine learning techniques have emerged as powerful tools for addressing these complex optimization challenges. This review examines the state-of-the-art in optimization techniques for structural engineering, synthesizing findings from 45 peer-reviewed empirical studies and review articles. The paper presents a comprehensive analysis of genetic algorithms, particle swarm optimization, ant colony optimization, harmony search, and emerging algorithms such as grey wolf optimization and modified tornado optimizers. Additionally, machine learning approaches including artificial neural networks, support vector machines, and surrogate modeling techniques are evaluated. Results demonstrate that hybrid metaheuristic-machine learning frameworks achieve superior performance compared to single-algorithm approaches, with improvements in convergence speed ranging from 30% to 99% in benchmark problems. Multi-objective optimization using Pareto-optimal solutions proves particularly effective for managing trade-offs in structural design. The review identifies critical research gaps, including limited work on reliability-constrained design, insufficient integration of manufacturing constraints, and the need for adaptive parameter tuning strategies. Future directions should focus on developing robust hybrid frameworks that combine evolutionary computation with deep learning techniques, and establishing standardized benchmarking protocols for algorithm comparison in structural engineering applications