Complex fracture networks, comprising hydraulic and preexisting natural fractures, play a pivotal role in characterizing flow dynamics and enhancing hydrocarbon production in low-permeability, low-porosity unconventional reservoirs. These networks function as the primary high-conductivity flow channels, as fluid flow through the matrix is minimal. Accurately representing their spatial distribution necessitates discrete fracture network (DFN) simulation methods, which treat individual fractures as distinct computational elements rather than averaging their properties into matrix grid blocks. This study introduces a comprehensive workflow for DFN generation, meshing, simulation, and result visualization, designed to improve the representation of complex fracture geometries and their impact on reservoir performance. A key innovation is integrating advanced refinement techniques with optimization-based meshing algorithms, creating high-quality unstructured perpendicular bisector grids that accurately capture fracture geometry and connectivity. The workflow is validated through two applications. The first uses an outcrop-based DFN model to investigate the effects of natural fracture properties on production. Sensitivity analyses reveal a positive correlation between natural fracture conductivity and oil production, with fracture permeability having a greater impact than width. The second employs synthetic DFN models to evaluate hydraulic fracturing strategies. Results demonstrate that non-uniform fracture designs consistently outperform uniform designs at the same stimulation cost, emphasizing the importance of tailored stimulation patterns to reduce undepleted regions and maximize stimulated reservoir volume. These findings highlight key factors controlling hydrocarbon recovery in shale plays, the limitations of uniform stimulation strategies in current practices, and advocate for customized approaches to optimize unconventional reservoir development.