Transportation services have a great importance which are make economic- social activities facilitate in the community. What is important for users of the urban traortation system is the quality of its services and what is widely understood by people in the questionnaires, they have complained that there is a busy and unpredictable travel time. Therefore, one of the main goals of policymakers in urban traort system is to reduce or eliminate the traffic congestion. Achieving this goal requires the adoption of appropriate policies by traort professionals. Though, in order to reduce traffic congestion, studies have been carried out on the selection of urban streets for widening, but the basis of these studies has been based on the traditional analysis of street networks. In this type of analysis, it is assumed that the Origin-Destination demand and street capacity are constant over time, while the demand for travel in a particular hour of the day may vary and the street capacity for various reasons such as traffic accidents, parking lots Incorrect, bad driving and ... may change. In another way, it cannot be said that a street in a particular hour of the day is definitely busy or quiet and should be talked about in a way about the crowded or silent on that street. With this in mind, we should study the performance of urban traort networks in terms of congestion in terms of reliability. One of the objectives of this study is to determine the reliability of the operation of urban traort networks in relation to recurrent congestion. For this purpose, in this thesis, a model has been presented to determine the probability of crowding the streets of a city network. This model has the ability to determine the probability of the congestion using the log-normal distribution function in showing the changes in travel time in an arc by using free travel time information and average travel time in that arc. Then, a new model was developed to calculate the reliability of the network with the probability of a bunch of arcs, and to illustrate how it was used, the model was implemented on a Sioux Falls traortation network and the reliability of the network was calculated. . Another goal of this study was to improve the reliability of the network's performance in connection with recurrent congestion using the street widening policy. In this thesis, a project selection model was used in traortation networks in this connection. The model is a high-level two-level model, the optimal set of projects to be selected from among a set of candidate projects, and with a limited budget allocation, is chosen and at the low level, a problem of equilibrium traffic allocation can be solved. To make since the problem of selecting a project in traort networks is NP-hard from the point of view of solution of NP-hard issues, a quantum evolutionary heuristic algorithm is used to solve the model of the selection of the widening project and its implementation results are presented on the Sioux Falls test network. The results of implementing the entire model on the Sioux Falls network indicate that the method is applicable to large networks Keywords: Project selection in urban traortation network, Frank-Wolfe algorithm, street widennig, Quantum Evolutionary Algorithm