Availability Optimization for Repairable Parallel-Series System by Applying Tabu-GA Combination Method
This study looks at the system availability optimization problem under different resource and design configuration constraints by applying Tabu-GA combination method. From the point of view of logistics engineering, availability optimization applied in the initial system development period, plays a key role to affect system reliability, system maintenance planning, logistics requirements, and related costs during system planned life cycle. Generally, system availability can be improved by increasing component reliability level, component repair rate or the number of components in each subsystem. This proposed model is to combine Tabu method with Genetic Algorithm to solve system availability optimization problem. Through our method applied in the initial phase of system design and development, we can find the optimal allocation of component redundancy number, reliability and maintainability levels to minimize the total system cost under different configuration constraints such as weight, volume, and system availability requirements. The proposed numerical examples are demonstrated based on different system constraint requirements and parameter values. Through a simulation study of 30 times of calculations by applying this proposed combination method, stable results are clearly showed. We also perform the sensitivity analysis based on the cost parameters associated with reliability level and maintenance rate to provide very helpful information for system design and development process. Finally, the performance of this proposed Tabu-GA combination method is compared with SA-GA combination method and it turns out to be very good results in many aspects.
Availability Tabu Search Simulated Annealing Genetic Algorithm Parallel-Series System
Gia-Shie Liu
Department of Information Management Lunghwa University of Science and Technology Taoyuan County, Taiwan
国际会议
IEEE 10th International Conference on Industrial Informatics(第十届IEEE工业信息学国际学术会议 INDIN2012)
北京
英文
803-808
2012-07-25（万方平台首次上网日期，不代表论文的发表时间）