Scholarlink

<p>This research investigates Multi-Objective Solid Transshipment Problems (MOSTP) within Ethiopian manufacturing supply chains because their logistics expenses exceed 30% of production costs. The study represents transportation expenses through Triangular Intuitionistic Fuzzy Numbers (TIFNs), which enable experts to assess uncertainty by measuring both their affiliation and non-affiliation probabilities. The research collected parameter data from eight to ten logistics managers in each industry through controlled measurement methods. To advance beyond standard approaches, the research presents a new computational system which uses fuzzy heuristics to mathematically merge buffer-stock network transformations with TIFN parameters before defuzzification. The system enables complete environmental modeling because it prevents premature loss of uncertainty information, thereby enabling Python’s SciPy optimization library with HiGHS solver to solve the resulting precise multi-objective linear program. The epsilon-constraint method produces Pareto frontiers through its execution. The Python-based implementation achieves solution times under one second for both case studies and scales efficiently to large synthetic networks with dimensions of 20 by 50 by 5 nodes. For the textile supply chain, the optimal cost is 52,800 ETB with a fuzzy range of 38,400 to 68,400 ETB, and the optimal time is 720&#xa0;h with a fuzzy range of 480 to 940&#xa0;h. Pareto analysis reveals increasing marginal costs for time reduction, from 120 ETB per hour initially to 288 ETB per hour for the fastest deliveries. Comparative analysis demonstrates up to 74% cost improvement over existing methods; this gain is achieved because the proposed framework utilizes an exact mathematical solver on a preserved network structure, completely avoiding the local-optima traps typical of existing fuzzy genetic algorithms or heuristics. TIFN outputs enable risk-aware planning with quantified contingency estimates of approximately 30% of expected costs, thereby replacing arbitrary buffer factors. The convex Pareto frontier reveals diminishing returns on time-reduction investments, which guides infrastructure prioritization. The open-source Python implementation eliminates software cost barriers, thereby supporting capacity building in Ethiopian institutions. The methodology successfully demonstrates practical fuzzy optimization in developing economy contexts where uncertainty is high and historical data are limited.</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

An integrated python-based methodology for multi-objective solid transshipment problems with triangular intuitionistic fuzzy numbers: applications to Ethiopian manufacturing supply chains

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月21日

查看全文
  • Walle Tilahun,
  • Terefe Getachew Beyene,
  • Melisew Adamu Brlie,
  • Abebaw Bizuayehu Derseh,
  • Emebet Getaneh,
  • Tegbew Addis,
  • S. K. Manikandan,
  • Mulualem Molla,
  • Ashebir Awoke

摘要

This research investigates Multi-Objective Solid Transshipment Problems (MOSTP) within Ethiopian manufacturing supply chains because their logistics expenses exceed 30% of production costs. The study represents transportation expenses through Triangular Intuitionistic Fuzzy Numbers (TIFNs), which enable experts to assess uncertainty by measuring both their affiliation and non-affiliation probabilities. The research collected parameter data from eight to ten logistics managers in each industry through controlled measurement methods. To advance beyond standard approaches, the research presents a new computational system which uses fuzzy heuristics to mathematically merge buffer-stock network transformations with TIFN parameters before defuzzification. The system enables complete environmental modeling because it prevents premature loss of uncertainty information, thereby enabling Python’s SciPy optimization library with HiGHS solver to solve the resulting precise multi-objective linear program. The epsilon-constraint method produces Pareto frontiers through its execution. The Python-based implementation achieves solution times under one second for both case studies and scales efficiently to large synthetic networks with dimensions of 20 by 50 by 5 nodes. For the textile supply chain, the optimal cost is 52,800 ETB with a fuzzy range of 38,400 to 68,400 ETB, and the optimal time is 720 h with a fuzzy range of 480 to 940 h. Pareto analysis reveals increasing marginal costs for time reduction, from 120 ETB per hour initially to 288 ETB per hour for the fastest deliveries. Comparative analysis demonstrates up to 74% cost improvement over existing methods; this gain is achieved because the proposed framework utilizes an exact mathematical solver on a preserved network structure, completely avoiding the local-optima traps typical of existing fuzzy genetic algorithms or heuristics. TIFN outputs enable risk-aware planning with quantified contingency estimates of approximately 30% of expected costs, thereby replacing arbitrary buffer factors. The convex Pareto frontier reveals diminishing returns on time-reduction investments, which guides infrastructure prioritization. The open-source Python implementation eliminates software cost barriers, thereby supporting capacity building in Ethiopian institutions. The methodology successfully demonstrates practical fuzzy optimization in developing economy contexts where uncertainty is high and historical data are limited.