Abstract
This paper proposes an integrated scenario tree model that incorporates recurrent congestion conditions and sporadic disasters into a stochastic degradable road network design problem (SDNDP). The traffic pattern of the stochastic degradable network (SDN) under the recurrent congestion condition is evaluated by probit-based stochastic user equilibrium (called SDN-SUE), whereas the system optimum is used to assess the traffic pattern of the SDN under the sporadic disaster (called SDN-SO). The proposed model determines optimal link capacity expansions that minimize the sum of the total network travel time costs of all recurrent congestion conditions plus the total investment cost subject to the desired total network travel time constraint for evacuation purpose and the SDN-SUE and SDN-SO conditions. A solution algorithm is also developed for solving the SDNDP. Numerical examples are given to demonstrate the potential pitfall in considering the network improvement policies separately and to show the benefit from the integrated model.
