| In this thesis, a generalized approach for the analysis of large-scale systems under uncertainty is developed. The key elements of the approach are formal methods and principles to decompose a large-scale system into smaller subsystems, to aggregate the results from analyzing the small subsystems, and to evaluate the uncertainty associated with the analysis.; The method to quantify uncertainty is developed from the three principles of uncertainty, namely the principle of minimum uncertainty, the principle of maximum uncertainty, and the principle of uncertainty invariance. This method is different from the existing methods in that it treats uncertainty in a problem specific and analyst specific context. It recognizes that the amount of uncertainty (or the amount of information desired) is dependent on the specific problem that is being addressed and the specific analyst or analysts who conduct the analysis. A sound procedure to incorporate the context and the analyst attitude is incorporated in the method.; The governing principles to conduct decomposition and aggregation in a consistent and clear manner are developed. The principle of decomposition is based on maximal exclusiveness among the subsystems, and the principle of aggregation is based on maximum consistency in combining information. It is shown that the adherence to these principles minimizes the loss of information and global inconsistency, which are bound to creep into the analysis of a large-scale system.; The application of the proposed approach is illustrated using three specific problems related to the field of transportation: one, the dynamic control of stochastic traffic flow on a city street network; two, the estimation of route choice decisions under route guidance with uncertainties in the advisory information; and three, an examination of the feasibility of achieving regional transportation goals through the implementation of certain Intelligent Transportation System (ITS) alternatives in the region. |
