A Demand Estimator Based on a Nested Logit Model
提出一种新的非线性需求估计器，适用于地面和航空交通，其估计需求分布遵循嵌套Logit模型，并开发了精确求解算法，在包含66,767个起讫对的美国实际网络中验证了收敛性和最优性条件。
The importance of historical travel demand has been well recognized by transportation researchers and practitioners. This paper presents a new nonlinear demand estimator that can be applied in both ground and air transportation. The estimator is formulated such that the distribution of the estimated demand follows a nested logit model. To solve the demand estimator, we develop an exact solution algorithm, which maximizes its dual problem sequentially along unit directions and keeps some of the first-order optimality conditions satisfied for the estimator. We investigate the convergence of the algorithm. We prove that all of the accumulation points produced by the solution algorithm satisfy some optimality conditions. A large example (a real U.S. network), which contains 66,767 origin-destination pairs, is presented.