Generating dynamic activity-travel schedules
Year: 2006
Type: conference-paper
Abstract: While making decisions, governments and policy makers are \nsupported by models in order to estimate the impact of their \ndecisions on society as a whole. Transportation models \ncompromise a major example of such decision supporting models. \nAfter all, transportation models are applied to monitor travel \nbehaviour, to evaluate policy decisions, to assess the \nenvironmental impact of traffic, etc. To this end, transportation \nmodels previously consisted of aggregate, trip-based models to \nestimate travel demand. \nDuring the past decade, activity based-transportation models \nentered the area of transportation modelling. Such models assume \nthat travel patterns are the result of activity schedules that \nindividuals execute in their attempt to achieve certain goals. \nTherefore, activity-based models aim at simulating the individual \ndecision-making behaviour concerning several activity-travel \nrelated dimensions simultaneously. For instance one decision \ncompromises the following facets: which activity to perform at \nwhich location, when to start this activity and for how long, which \ntransport mode along which route to use in order to get to the \ndesired location, who will accompany the individual during the course of the activity, etc. The resulting activity-travel schedules \nconstitute the basis of the assignment of the individual routes to \nthe transportation network, and as such estimating aggregate travel \ndemand. (Ettema and Timmermans, 1997; Timmermans, 2000) \nSuch activity-based transportation models offer the opportunity of \npredicting travel demand more accurately as they provide a more \nprofound insight into individual activity-travel schedules. \nFurthermore these models enable estimating more realistically the \nimpact of a policy on transportation related issues, for instance \ntraffic safety, environmental pollution and land use patterns. \nYet, the majority of such models is still quite static. Therefore, \ndynamic activity-based models are introduced quite recently in this \nresearch area. These models attempt at simulating the dynamic \ndecision-making process of individuals interacting with the \nenvironment. For this purpose, agents simultaneously enter the \ntransportation network and determine dynamically their activity \nschedules and routes. This signifies that, after having determined \nan activity schedule for each agent, the system executes the \nschedules of all agents simultaneously, i.e. all agents are processed \nin a given time step before going to the next time step. (Arentze, et \nal., 2005) \nIn the course of each time step, actual travel times are estimated \nand updated. Furthermore the effect of possible unforeseen events \nin the course of the execution of activities – e.g. negative or positive delays – or unexpected travel times, causing a mismatch \nbetween a scheduled and actual end time of an episode, can be \nsimulated. After all, the resulting time-surplus or time-lack \nsituation at the moment of completing an episode can trigger shortterm \nrescheduling. In addition, long-term learning can be \nincorporated into this model as well as an agent – after having \nexecuted the entire schedule – will update his knowledge and \nbeliefs with respect to the different attributes concerning his \nactivity and travel behaviour. (Arentze, et al., 2005) \nBefore being able to account for the above described processes of \nrescheduling and learning, the dynamic activity-travel schedules \nneed to be captured in a model. The goal of this research thus \nconsists of developing such activity-based scheduling algorithm. \nThe present research aims at contributing to this purpose, relying \non modelling techniques that originate from the area of scheduling \nand planning. The selected micro simulation technique will have to \ntake into consideration a range of temporal and spatial constraints, \nfor instance coupling constrains, opening hours of shops and \npublic authorities, availability of transport mode and available time \nwindows in which the activity can take place. To this end some \nprevious researches point the way to the area of operational \nresearch, e.g. the pickup and delivery problem with time windows \n(Recker, 1995), the area of project network scheduling, e.g. \nresource-constrained project scheduling (Icmeli, 1993), or the area \nof artificial intelligence, e.g. (relational) reinforcement learning \n(Arentze and Timmermans, 2003; Charypar and Nagel, 2005; \nDžeroski, et al., 2001) or Bayesian updating (Jha, et al., 1998).
Language: English
Subfield: Transportation
Field: Social Sciences
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W2794209131