comparison of metaheuristic algorithms for weight optimization of a semi-submersible vawt substructure with hexagonal pontoons

In response to rising global energy demand and the urgent need to reduce greenhouse gas emissions, offshore wind turbines (owts) have emerged as promising renewable energy solutions. among deep-water support structures, semi-submersible platforms offer superior motion stability and design flexibility, but their high structural weight significantly affects construction and installation costs. this study compares five metaheuristic algorithms—genetic algorithm (ga), ant colony optimization for continuous domains (acor), artificial bee colony (abc), firefly algorithm (fa), and particle swarm optimization (pso)—for weight optimization of a four-column semi-submersible substructure supporting a vertical axis wind turbine (vawt) with hexagonal pontoons. the algorithms were first validated with a reference platform optimized using the generalized reduced gradient (grg) method. they were then applied to minimize the vawt substructure weight by optimizing pontoon and column geometry, spacing, and draft under hydrostatic stability, motion, airgap, and feasibility constraints. each algorithm was executed five times, and kolmogorov–smirnov tests confirmed normality of optimized weight and number of function evaluations (nfe). analysis of variance (anova) indicated statistically significant differences among algorithms, and the technique for order of preference by similarity to ideal solution (topsis) was used for multi-criteria decision-making, considering average weight, nfe, accuracy, variance, and stability. results indicate that acor achieved the highest rank, achieving ~37.6% (3690 tons) weight reduction. the findings demonstrate acor's effectiveness as a decision-support tool for conceptual design of semi-submersible substructure of owts. however, it is expected that hydrodynamic loading, aero-structural coupling to be also considered for further detailed design.