Fuzzy modelling and cost optimization of fault-tolerant system with service interruption

Area of Science:

• Industrial Engineering
• Operations Research
• Reliability Engineering

Background:

• Fault-tolerant machining systems (FTMS) are crucial for industries requiring high reliability.
• Existing models may not fully capture complex failure and repair scenarios in FTMS.
• System availability and maintainability are key performance indicators.

Purpose of the Study:

• To develop a non-Markov queueing model for FTMS with server vacations, breakdowns, and imperfect repairs.
• To analytically predict the performance of FTMS using a finite population M/G/1 queueing model.
• To optimize FTMS design by minimizing total cost using meta-heuristic and classical optimization techniques.

Main Methods:

• Development of a non-Markov queueing model incorporating server vacation, breakdown, and imperfect repair processes.
• Application of the supplementary variable technique for analytic solutions of the M/G/1 queueing model.
• Implementation of parametric non-linear programming, Particle Swarm Optimization (PSO), Genetic Algorithm (GA), and quasi-Newton method for performance evaluation and optimization.

Main Results:

• An analytic solution for the performance prediction of FTMS was derived.
• Performance measures were evaluated in both crisp and fuzzy environments.
• Optimal design descriptors were determined by minimizing total cost using various optimization techniques.
• Sensitivity analysis of performance indices with respect to system parameters was conducted.

Conclusions:

• The developed non-Markov model provides a robust framework for analyzing FTMS with complex failure and repair dynamics.
• The optimization techniques effectively identify optimal design parameters for cost-efficient FTMS.
• The study offers valuable insights for enhancing the reliability and maintainability of industrial machining systems.