Multi-objective optimisation of reliable product-plant network configuration.

Area of Science:

• Operations Research
• Supply Chain Management
• Industrial Engineering

Background:

• Manufacturing reliability is crucial for meeting demand amidst production disruptions.
• Supply network reliability depends on product redundancy but involves balancing costs.
• Existing models lack the complexity to address real-world supply networks.

Purpose of the Study:

• Develop a generic measure for evaluating network reliability.
• Create a multi-objective evolutionary optimization model for cost-reliability trade-offs.
• Apply and compare genetic algorithms for optimizing automotive production networks.

Main Methods:

• Developed a generic reliability measure for product-plant configurations.
• Framed the problem as a multi-objective evolutionary optimization model.
• Utilized NSGA2, SPEA2, and PAES genetic algorithms on an automotive production network.

Main Results:

• Successfully identified cost-reliability trade-off solutions using the developed model.
• NSGA2 algorithm demonstrated superior performance in Pareto front spread.
• Algorithm choice significantly impacts the exploration of the solution space.

Conclusions:

• The proposed model effectively optimizes the cost-reliability balance in complex supply networks.
• Genetic algorithms are valuable tools for supply chain optimization.
• Algorithm selection is critical for achieving optimal results in multi-objective optimization.