Study On Dynamic Reliability Of Day-To-Day Road Network

As the economy grows at a high speed and people enjoys a higher standard of living, the performance stability of urban road network system is required in a deeper sense by the rapid rhythm of urban life. The road network system is needed to be able to quickly respond to and deal with urgent incidents, and is rather needed to be able to provide stable service for daily repeated trips of travelers. So, through the study of dynamic reliability of day-to-day road network system, the laws of both dynamics and stochasticities of demand-supply interactions of road network system can be disclosed, afterwards, the corresponding planning schemes and management actions can be forged to improve the dynamic stability of day-to-day road network system.Although the classic reliability analysis can disclose the stochasticities of day-to-day road network system, it is incapable to describe the dynamics appropriately. Since the participation of travelers as users and the occurrence of stochastic accidents, urban road network system has both stochasticities and dynamics affected by the interactions between demand side and supply side. No doubt, conclusions from imperfect reliability analysis must have limited worth to serve travelers and decision makers, therefore the congestion problems of urban city are still not handled essentially.For the purpose, based on the classic road network reliability analysis, the static reliability analysis is extended to dynamic version by introducing the dynamic analysis, and the dynamic reliability of day-to-day road network is studied thoroughly in the dissertation, and the main contents are described as follows.The sense of integrating dynamic analysis into the classic framework of road network reliability analysis is stated. The reviews of state-of-art methods both on road network reliability and day-to-day network dynamic assignment are carried out, the existing issues of current researches are also analyzed. The related issues about time dimension, basic structure, stochasticity and dynamics of day-to-day road network are discussed, and the feasibility of using time-dependent probability distribution for describing the stochastic dynamic variation both of demand side and supply side of road network is also argued. The theories and models for travelers’ route choice behavior in stochastic dynamic road network system are analyzed.The essence of individual traveler’s route choice is analyzed in framework of probability theory, a joint probability distribution of individual traveler’s route choice decision random vector is established. Assuming the homogeneity of travelers in the same user class, a joint probability distribution of route flow random vector is deduced. Then, the existence and uniqueness of a stationary probability distribution of route flows obtained from a stochastic dynamic assignment model are proved. By considering the influence of information upon perceptions of travelers, a perception-updating-based stochastic dynamic assignment model with information provision is proposed. In numerical studies, the state evolutions of road network with and without information provision are compared; it is found that total travel time of road network with information provision is higher than that of the one without information provision. This means that even though traffic information provision might increase the convenience of trip making and information acquiring for travelers, it somewhat satisfies the efficiency of network system.Traffic incidents are categorized due to their duration lengths, the similarity and dissimilarity of characteristics between hours-lasting incidents and days-lasting incidents are compared. Numerical simulations of variations of road network travel time are implemented through stochastic dynamic assignment model. Under scenes with different incident durations, locations and severities, the travel time variations are also analyzed respectively. A travel time dynamic reliability is defined, and a feasible appraisal procedure is designed by applying the stochastic dynamic assignment model with stochastic supply of road network.The method of stochastic failure sequential analysis from reliability engineering is introduced, a dynamic reliability indicator for measuring the ability of road network system providing uninterrupted reliable services based on the first failure time of stochastic sequence. A performance index function is established, and an assessment method is designed by utilizing a binary Markov chain and stochastic dynamic assignment model, the test and analysis for proposed indicator and assessment method are carried out in numerical examples.