Reducing Unreliability of Complex Systems by Numerical Optimization: A Hybrid Approach

This paper investigates the effectiveness of numerical optimization techniques in enhancing the reliability of complex systems. The main objective of the study is to reduce the unreliability of complex systems by narrowing the range of component unreliability, thereby reducing the probability of system failure and achieving stability and robustness in the reliability of complex systems. The technique for this task involves formulating the problem as a multi-objective numerical optimization problem under specified constraints using Particle Swarm Optimization and Simulated Annealing algorithms, as well as a hybrid approach combining the two algorithms. The multi-objective optimization techniques were applied to a more complex system modeled with 24 nodes and 38 components. The results demonstrate the effectiveness of the optimization algorithms in reducing the unreliability of the complex system and optimizing its reliability, indicating the potential of both algorithms in optimizing complex systems. Statistical analysis demonstrated the superiority of the Simulated Annealing algorithm over the Particle Swarm Optimization algorithm, while the hybrid approach outperformed both. This work opens up broad prospects for achieving substantial improvements in the design of complex systems that require high accuracy and performance efficiency.