Scholarlink

<p>This study proposes a simulation-optimization framework for daily production planning in a garment sewing line under a quality- delivery trade-off. Prior scheduling studies often optimize throughput while treating quality as fixed; this study models quality as a decision-dependent outcome within the scheduling process. A logistic regression model estimates daily defect probability from operational decisions, including machine cleaning frequency, overtime hours, and the ratio of experienced operators. These probabilities are embedded in an evaluation module that jointly captures defect rate and backlog formation under capacity constraints (setup/cleaning time, speed limits, and a discrete overtime ladder). A Genetic Algorithm minimizes defect rate with a backlog cap, while NSGA-II generates Pareto-optimal schedules balancing defect reduction and backlog control. A case study using 51 production days from Sewing Line 4 of a Vietnamese garment firm (June- July) shows practical applicability. GA achieves a 1.6616% average defect rate, and a selected NSGA-II trade-off solution yields 1.7% defect rate with 5.42% backlog ratio. The resulting Pareto front offers transparent and actionable decision support for production managers facing competing quality and delivery priorities.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Multi-objective optimization for reducing defects and balancing production timeliness: a case study of sewing line in the garment industry

  • 期刊论文

  • 发表日期：

    2026年5月15日

查看全文
  • Le Duc Dao,
  • Nguyen Thi Diep Tran,
  • Le Duc Hanh

摘要

This study proposes a simulation-optimization framework for daily production planning in a garment sewing line under a quality- delivery trade-off. Prior scheduling studies often optimize throughput while treating quality as fixed; this study models quality as a decision-dependent outcome within the scheduling process. A logistic regression model estimates daily defect probability from operational decisions, including machine cleaning frequency, overtime hours, and the ratio of experienced operators. These probabilities are embedded in an evaluation module that jointly captures defect rate and backlog formation under capacity constraints (setup/cleaning time, speed limits, and a discrete overtime ladder). A Genetic Algorithm minimizes defect rate with a backlog cap, while NSGA-II generates Pareto-optimal schedules balancing defect reduction and backlog control. A case study using 51 production days from Sewing Line 4 of a Vietnamese garment firm (June- July) shows practical applicability. GA achieves a 1.6616% average defect rate, and a selected NSGA-II trade-off solution yields 1.7% defect rate with 5.42% backlog ratio. The resulting Pareto front offers transparent and actionable decision support for production managers facing competing quality and delivery priorities.