A unifying framework for demand-responsive public bus systems

When demand for transportation is low or highly variable, traditional bus services lose efficiency and tend to frustrate passengers, and a large number of demand-responsive systems have been proposed to fix that. What was missing was a comprehensive overview: the literature was fragmented, with each paper introducing its own terminology and assumptions, making it hard to see how the systems relate to one another.

This review fills that gap with a unifying framework that classifies demand-responsive public bus systems along three degrees of responsiveness — dynamic online, dynamic offline and static — and further distinguishes many-to-one (feeder) from many-to-many systems. For every system it records the optimization problem that was modelled, whether realistic data was used, and the size of the instances tested, with summary tables over the whole corpus.

The screen started from more than 300 records and retained 151 studies. The classification makes the gaps in the field visible: only 42 concerned many-to-one systems, and of those only 9 were semi-flexible against 33 fully flexible. Semi-flexible many-to-one systems in the dynamic online setting — the most operationally demanding and most realistic combination — were covered by exactly one prior paper. That gap is what motivated the direction of my own PhD.

Published jointly with Pieter Vansteenwegen, Lissa Melis, Dilay Aktaş Dejaegere, Paulo Vieira and Kenneth Sörensen.