Importance measure construction and solving algorithm oriented to the cost-constrained reliability optimization model

Importance measures are used to prioritize the system components for achieving high efficiency and economy of reliability optimization. The existing importance measures pay more attention to evaluating the impact of component performance changes on the objective function (such as system reliability, cost and so on). Still, these importance measures do not consider the influence of constraints in the system reliability optimization models during the ranking process of components, which may leave the component with the largest importance measure almost no space to improve its reliability. For the cost-constrained reliability optimization model (ROM), this paper proposes a cost-constrained ROM-based importance measure (CRIM) by comprehensively considering the objective function and constraints, and several properties of CRIM are discussed. Then, the CRIM-based genetic algorithm (CRIGA) is developed to solve the cost-constrained ROM. The numerical experiment for different scale systems is implemented to evaluate the performance of CRIGA by comparing it with other optimization algorithms. Experimental results show that CRIGA can get better solutions with faster convergence speed and better robustness. Finally, the case on the coal transportation system of the thermal power plant is introduced to illustrate the application of CRIM and CRIGA in reliability optimization.