Experimental study and modeling of departure time choice behavior in the bottleneck model with staggered work hours

This paper designed an experiment to explore the effects of staggered work hours on the departure time choice behavior of commuters based on a discrete Vickrey's bottleneck model. The experiment included a benchmark scenario with one work start time and a treatment scenario with two work start times. In the treatment scenario, subjects were classified into two groups with different work start times, and such two groups had interactions because of the small gap between the two work start times. Our experimental results indicated that (i) subjects' choice behavior deviated from the pure-strategy equilibrium in both scenarios, (ii) subjects' average travel cost in the benchmark scenario was higher than that of the pure-strategy equilibrium, and (iii) in the treatment scenario, the average travel cost in the group with earlier work start time did not significantly differ from the pure-strategy equilibrium, whereas subjects in the group with later work start time had lower average travel cost than the pure-strategy equilibrium. Finally, we proposed a reinforcement learning model to simulate the departure time choice behavior of the commuters. The simulation results were consistent with the collective behavior of the experimental observations. Our work is expected to shed light on the theoretical study on the departure time choice behavior concerning staggered work hours.

Automated summary

This study designed an experiment to explore the effects of staggered work hours on the departure time choice behavior of commuters. The experimental results indicated that the subjects' choice behavior deviates from the pure-strategy equilibrium and staggered work hours have a significant impact on travel costs.