A practical Markovian model of the availability and reliability of a mass transport service with non-exponential repair times

摘要: The impact of the shape of the repair time distribution on the reliability and availability of a fleet of trams is analyzed. Trams are treated as multi-state systems. Failure times and repair times are modeled as mutually s-independent exponential and non-exponential random variables, respectively. Only one maintenance team, which serves one tram at a time, is assumed to be available. So trams that need repairs are arranged in a queue. This sharing of resources gives rise to a stochastic dependence among the states of different trams. This dependency is directly modeled describing the entire fleet, as a single system, via a homogeneous Markov process extended by the “Device Of Stages’’technique. Thus, the behaviour of a complex semi-regenerative process is emulated using two relatively simple tools. The proposed methodology is presented developing an application to a real case study.