Availability optimization of a redundant dependent system using genetic algorithm
Abstract
Parallel redundancy is a fundamental concept widely used to maximize system reliability/availability. However, any enhancement in the system reliability requires resources which consequently increases the costs. Hence, it is important to optimize the system configuration at the design phase. It is demonstrated that modeling the dependency between system components can be an effective option to improve system reliability in an economic way. But, the resolution of the optimization problem by analytic methods may be difficult and expensive in terms of computational time especially for large scale and complex systems. This paper proposes a rapid and effective approach to optimize the cost of a reparable system with dependent components under the availability constraint. The optimization model is solved using the software tool LINGO and the genetic algorithm (GA). A comparison between LINGO and the GA results is established. The results show the superiority of the GA in term of speed with a small difference in the exactitude of the obtained solutions.