Research On Travel Time Reliability Of Urban Expressway

With increasing demands for better reliable service, many systems (e.g. electrical power systems, water distributions, communication networks, etc.) have incorporate reliability analysis as an integral part in their planning, design, and operation. However, reliability analysis has received little or not attention in the study in transportation systems in the past. In recent times the stability of road networks has become an increasing important issue, since the value of time has increased considerably and unexpected delay can results in substantial loss to road users. Road networks are now required to have high degree reliability to ensure drivers smooth travel under normal traffic flow fluctuations and to avoid serious unexpected delays for disruptions within the network. Road network reliability has now become an important performance measure for evaluating road networks, especially when considering changes in OD traffic demand and link flow capacity over time. According to the literature about reliability of road network, there exist several concept of reliability of road network in the world, namely connectivity reliability, travel time reliability, capacity reliability, and unblocked reliability. On the basis of studying the characteristic of urban expressway, we choired the travel time reliability as our study target. This paper investigates the transportation network reliability (in terms of travel time reliability) based on the randomicity analysis of demand-supply. Two models are formulated for calculating travel time reliability of urban expressway: the first model, in which the occasional incidents (e.g. traffic accidents, road maintain, poor weather etc.) are not considered, focuses on the fluctuations of origin-destination (OD) demand. A Monte Carlo simulation procedure is developed to estimate the travel time reliability. The application of the proposed model is illustrated by a numerical example; The other model take into account the disturbance arising from randomly incidents, in which the Stochastic User Equilibrium (SUE) model assuming user's route choice behavior under uncertain network condition is incorporated in the travel time reliability model. Due to large variability associated with link capacities resultfrom the interrupting of randomly incidents, this paper adopt the technique of BP neural network to calculate the actual capacity...