Discrete traffic network design model under supply and demand uncertainty

Regarding the uncertainty of traffic supply, demand and decision principles in traffic design, a design model for uncertain discrete traffic network based on uncertainty theory is proposed. Firstly, supply and demand elements are assumed to be stochastic variables. The semi-standard error and value at risk are used to measure uncertainty and a multi-objective bi-level optimized model considering decision risk attitude is built. Then, SPEA2 (strength pareto evolutionary algorithm 2) based on the conception of Pareto is applied to solve the multi-objective bi-level optimized model. The numerical results of Nguyen-Dupuis network show that the risk value can reflect decision maker's attitude towards risk in measuring traffic design. In various targets there are alternative relations in multi-object design. The increase of the construction funds can effectively improve the network performance. However, such improvement is not obvious when the investment is beyond a certain limit.