Toward smart and adaptive apparel manufacturing: pattern-driven genetic algorithm optimization for modular garment assembly

Abstract

Purpose This study proposes a pattern-driven genetic algorithm (GA)–based machine allocation framework for modular garment assembly systems to improve production efficiency, workload balance, and adaptability. Design/methodology/approach Seam-level information extracted from computer-aided design garment patterns is aggregated into workstation-level workloads using machine-specific sewing speeds. Given a fixed modular assembly structure, a GA optimizes the allocation of limited sewing-machine resources across workstations. The optimized layouts are evaluated through deterministic event-based production simulation under different production scales and additional-order insertion scenarios. Findings Simulation results show that GA-optimized machine allocation consistently reduces total production time and improves workload balance compared with non-optimized machine-allocation layouts, with performance gains increasing as production volume grows. In rescheduling scenarios involving additional-order insertion, GA-based re-optimization further shortens production time without physical line reconfiguration, demonstrating flexible production adjustment in response to urgent order changes. Originality/value This study introduces a novel integration of pattern-based workload modeling, GA-driven machine allocation, and simulation-based evaluation for modular garment assembly. By directly linking CAD-derived sewing information to optimization and production-performance evaluation, the proposed framework extends apparel manufacturing optimization beyond conventional line-balancing approaches.