Bi-level optimization of shared manufacturing service composition based on improved NSGA-II

Area of Science:

• Manufacturing Systems Engineering
• Operations Research
• Computer Science

Background:

• Shared manufacturing environments face challenges in comprehensively representing service composition indexes.
• Existing models may not adequately capture crucial factors like service reliability and confidence.

Purpose of the Study:

• To develop a robust service composition evaluation system for shared manufacturing.
• To propose a bi-level programming optimization model for shared manufacturing service composition.

Main Methods:

• Decomposition of service reliability, confidence, and task fit indexes.
• Establishment of a bi-level programming model with Quality of Service (QoS) as the upper objective and reliability, confidence, and task fit as the lower objectives.
• Application of Criteria Importance Though Intercrieria Correlation (CRITIC) for weight determination and an improved Fast Elitist Non-Dominated Sorting Genetic Algorithm (Improved NSGA-II) for multi-objective optimization.

Main Results:

• The proposed model effectively optimizes shared manufacturing service composition.
• The Improved NSGA-II demonstrated a 23.33% increase in convergence speed compared to traditional NSGA-II.
• The Improved NSGA-II achieved a 69.99% gain in operational efficiency over traditional NSGA-II.

Conclusions:

• The developed service composition optimization model is viable and effective for shared manufacturing.
• The Improved NSGA-II offers significant performance enhancements for solving multi-objective optimization problems in this domain.
• The study provides a more comprehensive approach to evaluating and selecting services in shared manufacturing settings.