Solution Methods for Dynamic Offline Relaxation-Based O-D Estimation Problem

摘要: The relaxation-based method was recently proposed to estimate origin-destination (O-D) travel demands from limited traffic measurements. This does not enforce the user equilibrium condition so that difficulties associated with traditional bi-level programming are avoided. paper a column generation procedure for solving dynamic off-line O-D estimation problem which is formulated as variational inequality (VI). VI encapsulates flow dynamics through network loading model. It shown space-time expanded can be constructed generate paths needed formulation optimally. Algorithms two key sub-problems: namely restricted and time-dependent minimum cost path problem, presented their implementation issues addressed. Numerical results based on synthetic data also presented.