Optimization Framework for Road Network Directed by Unblocked Reliability for Given Network Topology and Inelastic Demand with Stochastic User Equilibrium

To provide more reliable service for drivers, it is necessary to establish an optimization framework in which a reliability index is adopted as a performance index to evaluate road networks. First, a modified four- level model of unblocked reliability (link, path, Origin-Destination pair and entire road network) is proposed, in which the unblocked reliabilities of the Origin-Destination pair and the entire road network are formulated by the law of total probability when the unblocked reliability of path is a conditional probability. Then, a bi-level program is established based on the new model of unblocked reliability since planners design the road network and drivers respond the change in the road network. The objective function of the upper-level program is the maximized balance between the unblocked reliability of the entire road network and the road network expansion ratio which implies cost of improving the road network. The program equivalent of Stochastic User Equilibrium is adopted as the lower-level model so as to achieve consistency in route choices and to model the congestion effect in the road network. Next, a set of link capacity expansions is determined as a planning scheme by solving the bi-level program. This planning scheme not only improves the link of lower reliability but also takes into account the performance of the Origin-Destination pair and the entire road network. The proposed optimization framework is capable of improving the road network to its highest possible reliability level with a minimum scale of road network expansion.