Reliability Modeling And Optimization Of Linear Sliding Window System

System reliability research mainly aims to transform the complex,versatile and random systems in real life production into some specific mathematical models that can be used in scientific research,and then use mathematics,operations research,computer science and other related disciplines to solve it.The model under the hypothetical framework carries out research on system reliability improvement,and the research results are fed back to the actual applied science,which is the branch direction of management science and engineering discipline.This paper mainly studies the linear sliding window system model in the field of system reliability.Linear sliding window system is a common model in the field of reliability.It is widely used in power systems,non-random clustering,quality control and industrial inspection program systems.It has good application value.Conventional simple linear sliding window systems only consider a single evaluation criterion and the state of the components within the system also tend to determine event or binary distribution.However,the actual industrial production system is often accompanied by many uncertain factors.Therefore,such a simple linear sliding window system has great limitations in practical applications.Hence,we innovatively proposes a new two-failure criteria sliding window system——“SWS-mg/kn”.In this system,there exist n multi-state elements and each element has different performance states ranging from complete failure state to the perfect functioning.The system fails if either the total number of failed groups of r consecutive elements reaches k or the gap between any two adjacent failed groups of r consecutive elements is less than m normal working groups.In short,the main contributions of this paper are as follows: 1.This paper studies the reliability of a polymorphic system;2.The model in this paper allows multiple failed working groups,and the fault tolerance rate is higher;3.The model of this paper equips with more than one reliability evaluation standard so that the universality and expansion are stronger.To evaluate the system reliability,we extend the universal moment generating function technique and suggest a corresponding algorithm.The reliability importance is analyzed and both analytic and numerical examples are given to illustrate the model and algorithm.Besides,the optimization model of the multicriteria decision model is constructed from the perspective of the Allocation Strategy and the optimal model is solved by genetic algorithm.Finally,the paper also proposes a corresponding calculation algorithm,which calculates the importance of component reliability in the system,and lays a foundation for further optimization.