SUMMARY
The purpose of this study is to apply network analysis to evaluate the service performance of a freeway system, in which the speed on each section is stochastic owing to user behavior, accidents, tunnels, road gradient, and road repairs. A stochastic-flow freeway network (SFFN) is constructed to model the freeway system with stochastic speed. System reliability is defined as the probability of a certain number of vehicles in a freeway system completing their journey smoothly within a restricted time. From the quality management viewpoint, this can serve as an important performance indictor in the planning and management of freeway system traffic flow. Two SFFN models were built to describe the relationships among traffic flow, system time and stochastic speed. After obtaining the probability distribution of speed by collecting historical data, two algorithms with respect to different models were proposed to generate all minimal speed vectors (MSVs). The system reliability can be subsequently computed in terms of MSVs. The Taiwan freeway system was considered as a case study to demonstrate the effectiveness of the proposed models and algorithms.