Multi-objective optimization design for accelerated degradation test based on game theory.

Area of Science:

• Engineering
• Reliability Engineering
• Systems Engineering

Background:

• Accelerated degradation testing (ADT) is crucial for rapid reliability assessment of mechanical systems.
• Existing ADT optimization methods often focus on single objectives, leading to suboptimal results.
• A need exists for methods that balance multiple ADT objectives.

Purpose of the Study:

• To develop a multi-objective optimization design method for ADT based on game theory.
• To optimize prediction precision of lifetime distribution, estimation precision of model parameters, and robustness of parameter deviation.
• To ensure the total experimental cost does not exceed a predetermined budget.

Main Methods:

• Formulated a multi-objective optimization model for ADT.
• Transformed the model into a cooperative game problem using game theory.
• Applied fuzzy clustering to determine strategy spaces for each player.
• Developed a collusive cooperation model considering mutual benefits.
• Solved the model to achieve a compromised optimal test plan.

Main Results:

• The proposed method yields a test plan that balances multiple optimization objectives.
• Demonstrated the necessity and effectiveness of the game theory-based approach.
• Validated the method using examples of motorized spindles.

Conclusions:

• The multi-objective optimization design based on game theory provides a superior approach for ADT.
• This method effectively balances competing objectives like prediction accuracy, parameter estimation, and robustness.
• The approach offers a practical solution for optimizing ADT plans within cost constraints.