This paper is intended to develop a Bayesian decision model for the repair of deteriorating system. A non-homogeneous Poisson process with a power law failure intensity function is used to describe the behavior of the deteriorating repairable system. The decision on whether to have minimal repair or imperfect repair should be made on the occurrence of a failure. However, it is difficult to make a reasonable decision due to many uncertainties intrinsic in repair actions. In this paper, prior distributions are used in order to analyze the uncertainties embedded in the decision alternatives. Especially, a prior distribution for imperfect repair with probabilistic reduction in the failure intensity is proposed. In addition, mathematical expressions to calculate the expected prior loss of each repair alternative are proposed.

This paper is intended to develop a Bayesian decision model for the repair of deteriorating system. A non-homogeneous Poisson process with a power law failure intensity function is used to describe the behavior of the deteriorating repairable system. The decision on whether to have minimal repair or imperfect repair should be made on the occurrence of a failure. However, it is difficult to make a reasonable decision due to many uncertainties intrinsic in repair actions. In this paper, prior distributions are used in order to analyze the uncertainties embedded in the decision alternatives. Especially, a prior distribution for imperfect repair with probabilistic reduction in the failure intensity is proposed. In addition, mathematical expressions to calculate the expected prior loss of each repair alternative are proposed.

Intrusion Detection System (IDS) is intended to detect anomalous usages or attacks on system and network. Even though many IDS models have been developed, they still may have problems, such as involvement of components contaminated by anomalous process, in detecting anomalous behavior patterns. To distinguish anomalous patterns from normal patterns, we propose a Bayesian approach which includes the presentation of the relations among normal pattern components.