Abstract

In this study we will focus at comparing the impact of different holding-only control strategies determining which buses are to be held where and for how long. The comparison study is based on two approaches: one deterministic able to optimize over the entire simulation period, but assuming that future stochasticity only depends on the mean values of demand at stops: the second approach is stochastic, based on a hybrid predictive control formulation, which assumes explicitly the stochastic behavior for future demand prediction, but only considering a finite number of steps ahead to perform the online optimization. Different scenarios regarding design frequency and demand levels (capacity being reached and not reached) will be studied to identify under which conditions each strategy outperforms the other.