Dynamic transport mode estimation subject to joint decisions and spatio-temporal variations

Abstract

This thesis studies the dynamic route and transport mode optimization for participating in a jointly decided activity subject to spatio-temporal variations. Joint activity participants have to travel from their current locations to a common location which is the location of the joint activity. Apart from the recurrent joint activities (such as work, school etc.), there are several joint leisure activities where activity participants have to decide about the location and the starting time of the joint leisure activity along with the transport mode(s) that each one of them has to use for commuting from his/her current location to the joint leisure activity location.The objective of this thesis is the development of a comprehensive system for the optimization of joint leisure trips that are not related to work by optimizing (a) the location and the starting time of the joint leisure activity (b) the public transport operations (c) the transport mode(s) selection for each activity participant in order to arrive there as fast as possible while satisfying his/her personal trip preferences.Activities not related to work can be responsible for more than 60\% of trips at an urban environment. Non-working travel patterns differ from the more stable, recurrent travel patterns of work-related activities such as trips from/to work, school etc. and the three main differences of those activities are: (1) The location of a leisure activity can differ on a daily basis (it is not static like the location of the working or studying place); (2) The starting time of a leisure activity has greater elasticity and can differ on a daily basis (it is not stable such as the starting time of work, school etc.); (3) The alternative journey options to and from a leisure activity location are not well-known to the users (users are more aware of their journey alternatives when it comes to transfers from/to work-related activities since those activities are re-current) . Given that a significant number of transfers is related to leisure activities, the optimization of the (1) location selection, (2) starting activity time, (3) transport mode selection and (4) route selection are of paramount importance for both the commuters' total travel cost and the transport network performance. In addition, the prediction of non-recurrent activities in time and space can be an important step forward for the tactical and dynamic planning of transport networks since the volume and the non-recurrent nature of such activities lead to significant travel demand variations compared to the more stable, work-related activities. Due to the above, this thesis focuses on: (i) Understanding the State-of-the-Art (SoA) work on utilizing user-generated data for increasing the efficiency level of joint leisure activities and proposing actions towards this direction; (ii) Capturing users' willingness to travel certain distances for participating in different types of activities; (iii) Optimizing the selection of locations and starting times of joint leisure activities (iv) Re-scheduling the starting times of public transportation trips in order to adjust to the joint leisure activity demand without deteriorating the Quality of Service (QoS) for other passengers; (v) Optimizing the journey/path selection of users' who are willing to travel from one point of the network to another for participating at one activity and, possibly, utilize multiple modes while also satisfying their preferences.