<p>Additive manufacturing (AM), particularly laser beam powder bed fusion (PBF-LB), enables the fabrication of complex metal parts. However, these parts often necessitate support structures to prevent distortion and support overhanging features during printing. While crucial, support structures can be bulky and require significant post-processing for removal, impacting cost and efficiency. This study presents a generative optimization approach for designing support structures that reduce material usage, simplify removal, and minimize distortion. The generative optimization model developed in this research integrates process simulation, voxel-based mesh creation, and mathematical modeling to create support structures that balance material efficiency and structural integrity. Six distinct support structure types (block, Hcell, contour, rod, tree, and generative) were evaluated at three different levels: non-optimized supports, optimized using a generative approach, and Simufact optimized. The generative optimization model minimizes support volume and mass while ensuring structural integrity under yield strength limits. The proposed optimization strategy achieved a substantial reduction of 18% in support material usage compared to non-optimized designs. Additionally, the generative optimization approach demonstrated superior performance in reducing the induced distortions, with a maximum reduction of 47% compared to other designs. Ultimately, successful 3D printing of an exhaust manifold using the optimized generative supports established a firm correlation between predicted and observed distortions, validating the accuracy of the developed approach. This research underlines the potential of the generative support optimization approach for designing effective support structures in powder-based AM processes, leading to enhanced efficiency, material savings, and dimensional accuracy.</p>