Valuation and optimization for performance based logistics using continuous time Bayesian networks

When awarding contracts in the private sector, there are a number of logistical concerns that agencies such as the Department of Defense (DoD) must address. In an effort to maximize the operational effectiveness of the resources provided by these contracts, the DoD and other government agencies have altered their approach to contracting through the adoption of a performance based logistics (PBL) strategy. PBL contracts allow the client to purchase specific levels of performance, rather than providing the contractor with the details of the desired solution in advance. For both parties, the difficulty in developing and adhering to a PBL contract lies in the quantification of performance, which is typically done using one or more easily evaluated objectives. In this work, we address the problem of evaluating PBL performance objectives through the use of continuous time Bayesian networks (CTBNs). The CTBN framework allows for the representation of complex performance objectives, which can be evaluated quickly using a mathematically sound approach. Additionally, the method introduced here can be used in conjunction with an optimization algorithm to aid in the process of selecting a design alternative that will best meet the needs of the contract, and the goals of the contracting agency. Finally, the CTBN models used to evaluate PBL objectives can also be used to predict likely system behavior, making this approach extremely useful for PHM as well.