Path Choice Model Research Articles

Day-to-day dynamic models for Intelligent Transportation Systems design and appraisal

Transportation Supply Design (TSD) with demand assignment provides a powerful framework to support project appraisal, since modifications of existing and/or introduction of new transport facilities and/or services may greatly affect traveller behaviour, concerning path choice at least. TSD also includes the Design of Intelligent Transportation Systems (ITSs), such as Advanced Traveller Information Systems (ATISs) or Advanced Driver Assistance Systems (ADASs). Solution approaches available for all the above problems are based on user equilibrium (UE) assignment, with either Wardrop or (less frequently) probabilistic path choice models. Still, optimization of transportation supply under equilibrium assumption may not guarantee that an effective solution is obtained; indeed the system may not evolve towards the equilibrium state, if this state is not stable. Thus results of project appraisals based on equilibrium assignment only may be misleading. On the other hand, day-to-day dynamic models provide a more general approach to demand assignment, including as special cases equilibrium state.This paper aims at supporting this conclusion, by describing an extension of a deterministic process model to include the evolution over time of the total user surplus. According to this approach a project appraisal should be based both on the value of user surplus and its stability over time. A simple but effective application shows that the proposed approach can be applied to model the effect of ITS. Results for a small network show that an accurate design of ITS based on the effects on total user surplus requires a day-to-day dynamic analysis.

Cost Functions for Strategies in Schedule-Based Transit Networks with Limited Vehicle Capacities

Passengers in congested transit networks need to understand capacity constraints of vehicles when network congestion prevents boarding a vehicle. A strategy concept was developed for schedule-based networks. The passenger is assumed to know the schedule, to be aware of the capacity constraints, and to realize the possibility of being unable to board a vehicle. Therefore, the passenger must plan for several what-if scenarios at each decision point of the trip. The concept of strategies in networks with capacity constraints was extended in a study. The strategy costs were defined in terms of a random variable. Properties of this random variable, mean value and variance, for example, were used to define new types of cost functions. This approach accounts for passengers' aversion to travel time variability and risk. In the literature, these ideas were originally developed for path choice models in private traffic but had not yet been applied to public transport. In the current study, the developed mean excess approach was adjusted so that it could be applied to public transport.

Large-scale application of MILATRAS: case study of the Toronto transit network

This paper documents the efforts to operationalize the conceptual framework of MIcrosimulation Learning-based Approach to TRansit Assignment (MILATRAS) and its component models of departure time and path choices. It presents a large-scale real-world application, namely the multi-modal transit network of Toronto which is operated by the Toronto Transit Commission (TTC). This large-scale network is represented by over 500 branches with more than 10,000 stops. About 332,000 passenger-agents are modelled to represent the demand for the TTC in the AM peak period. A learning-based departure time and path choice model was adopted using the concept of mental models for the modelling of the transit assignment problem. The choice model parameters were calibrated such that the entropy of the simulated route loads was optimized with reference to the observed route loads, and validated with individual choices. A Parallel Genetic Algorithm engine was used for the parameter calibration process. The modelled route loads, based on the calibrated parameters, greatly approximate the distribution underlying the observed loads. 75% of the exact sequence of transfer point choices were correctly predicted by the off-stop/on-stop choice mechanism. The model predictability of the exact sequence of route transfers was about 60%. In this application, transit passengers were assumed to plan their transit trip based on their experience with the transportation network; with no prior (or perfect) knowledge of service performance.

Mind the map! The impact of transit maps on path choice in public transit

This paper investigates the impact of schematic transit maps on passengers’ travel decisions. It does two things: First, it proposes an analysis framework that defines four types of information delivered from a transit map: distortion, restoration, codification, and cognition. It then considers the potential impact of this information on three types of travel decisions: location, mode, and path choices. 1 Path refers to a unique sequence of entry, transfer, and exit stations/stops in the public transit network. The author differentiates between path and route choices because the latter could refer to a situation among different service routes that follow the same physical path, which is not the purpose of this analysis. 1 Second, it conducts an empirical analysis to explore the impact of the famous London tube map on passengers’ path choice in the London Underground (LUL). Using data collected by LUL from 1998 to 2005, the paper develops a path choice model and compares the influence between the distorted tube map (map distance) and reality (travel time) on passengers’ path choice behavior. Results show that the elasticity of the map distance is twice that of the travel time, which suggests that passengers often trust the tube map more than their own travel experience on deciding the “best” travel path. This is true even for the most experienced passengers using the system. The codification of transfer connections on the tube map, either as a simple dot or as an extended link, could affect passengers’ transfer decisions. The implications to transit operation and planning, such as trip assignments, overcrowding mitigation, and the deployment of Advanced Transit Information System (ATIS), are also discussed.

Adaptive route choices in risky traffic networks: A prospect theory approach

This paper deals with route choice models capturing travelers’ strategic behavior when adapting to revealed traffic conditions en route in a stochastic network. The strategic adaptive behavior is conceptualized as a routing policy, defined as a decision rule that maps from all possible revealed traffic conditions to the choices of next link out of decision nodes, given information access assumptions. In this paper, we use a specialized example where a variable message sign provides information about congestion status on outgoing links. We view the problem as choice under risk and present a routing policy choice model based on the cumulative prospect theory (CPT), where utility functions are nonlinear in probabilities and thus flexible attitudes toward risk can be captured. In order to illustrate the differences between routing policy and non-adaptive path choice models as well as differences between models based on expected utility (EU) theory and CPT, we estimate models based on synthetic data and compare them in terms of prediction results. There are large differences in path share predictions and the results demonstrate the flexibility of the CPT model to represent varying degrees of risk aversion and risk seeking depending on the outcome probabilities.

Adaptive Route Choice Models in Stochastic Time-Dependent Networks

Adaptive route choice models are studied that explicitly capture travelers’ route choice adjustments according to information on realized network conditions in stochastic time-dependent networks. Two types of adaptive route choice models are explored: an adaptive path model in which a sequence of path choice models are applied at intermediate decision nodes and a routing policy choice model in which the alternatives correspond to routing policies rather than paths at the origin. A routing policy in this study is a decision rule that maps from all possible pairs (e.g., node, time) to the next links out of the node. Existing route choice models that can be estimated on disaggregate revealed preferences assume a deterministic network setting from the traveler's perspective and cannot capture the traveler's proactive adaptive behavior under uncertain traffic conditions. The literature includes a number of algorithmic studies of optimal routing policy problems, but the estimation of a routing policy choice model is a new research area. The specifications of estimating the two adaptive route choice models are established and the feasibility of estimation from path observations is demonstrated on an illustrative network. Prediction results from three models–nonadaptive path model, adaptive path model, and routing policy model–are compared. The routing policy model is shown to better capture the option value of diversion than the adaptive path model. The difference between the two adaptive models and the nonadaptive model is larger in terms of expected travel time if the network is more stochastic, indicating that the benefit of adaptivity is more significant in a more unpredictable network.

Path Choice Model Based on Fuzzy Priority Relation

AbstractFuzzy phenomena emerge frequently in the process of travel decision-makings. Based on fuzzy priority relation between paths, a new path choice model was provided under assumption of fuzzy user equilibrium. Firstly, after the path with minimal perceived impedance was obtained with mean area method, a choice set of paths was produced by dominance check. Based on the fuzzy priority relation array, priority values of the choice paths were computed, and the choice possibilities of all paths can be obtained. The model can be regarded as a generalization of classical Logit model. On condition that perceived impedance of path is difficult to be described with fuzzy variable, the choice possibility of the path can be obtained by comparing priority relation between a pair of paths with the model.

Walking infrastructure design assessment by continuous space dynamic assignment modeling

AbstractFor the planning and design of walking infrastructure, characterized by the fact that the pedestrians can choose their paths freely in two‐dimensional space, applicability of traditional discrete network models is limited. This contribution puts forward an approach for user‐optimal dynamic assignment in continuous time and space for analyzing for instance walking infrastructure in a two‐dimensional space. Contrary to network‐based approaches, the theory allows the traffic units to choose from an infinite non‐countable set of paths through the considered space. The approach first determines the continuous paths using a path choice model. Then, origin‐destination flows are assigned and traffic conditions are calculated. The approach to determine a user‐optimal assignment is heuristic and consists of a sequence of all‐or‐nothing assignments. An application example is presented, showing dynamic user equilibrium traffic flows through a realistic transfer station. The example is aimed at illustrating the dynamic aspects of the modeling approach, such as anticipation on expected flow conditions, and predicted behavior upon catching or missing a connection.

Dynamic user-optimal assignment in continuous time and space

To support planning, design, and management of pedestrian infrastructure, dynamic assignment models are useful tools. However, current models are network-based and presume that travelers can choose between a finite number of discrete route alternatives. For walking facilities, where pedestrians can choose their paths freely in two-dimensional space, applicability of these traditional network models is limited. This article puts forward an approach for user-optimal dynamic assignment in continuous time and space. Contrary to network-based approaches, the theory allows the traffic units to choose from an infinite non-countable set of paths through the considered space. The approach consists of three interrelated steps, that is: determining the continuous paths using a path choice model, assigning the origin–destination flows, and calculating the resulting traffic conditions. The approach to determine a user-optimal assignment is heuristic and consists of a sequence of all-or-nothing assignments in continuous time-space. The article presents the mathematical problem formulation, solution approaches, and application examples.

A within-day dynamic traffic assignment model for urban road networks

In this paper a new formulation of within-day dynamic traffic assignment is presented, where a dynamic user equilibrium is expressed as a fixed-point problem in terms of arc flow temporal profiles. Specifically, it is shown that, by extending to the dynamic case the concept of network loading map, one need not introduce the continuous network loading problem in order to ensure the temporal consistency of the supply model. On this basis it is possible to devise efficient assignment algorithms, based on piece-wise linear or piece-wise constant approximation of temporal profiles over predefined time intervals covering the period of analysis, whose complexity is equal to the one resulting in the static case multiplied by the number of time intervals. With specific reference to a Logit path choice model, an implicit path enumeration network loading procedure is obtained as an extension of Dial's algorithm; then, the fixed-point problem is solved through the Bather's method.

Assessment of the Transfer Penalty for Transit Trips Geographic Information System-Based Disaggregate Modeling Approach

Transit riders negatively perceive transfers because of their inconvenience, often referred to as a transfer penalty. Understanding what affects the transfer penalty can have significant implications for a transit authority and also lead to potential improvements in ridership forecasting models. A new method was developed to assess the transfer penalty on the basis of onboard survey data, a partial path choice model, and geographic information system techniques. This approach was applied to the Massachusetts Bay Transportation Authority subway system in downtown Boston. The new method improves the estimates of the transfer penalty, reduces the complexity of data processing, and improves the overall understanding of the perception of transfers.

A joint mode-transit service choice model incorporating the effect of regional transport service timetables

Transit systems offer substantially different services in urban and inter-urban areas; this usually leads to different assumptions about user behaviour in path choice modelling. These differences have not yet been recognised in mode choice modelling. Consistent with the above, in this paper we propose a joint mode-transit service choice model where the elementary alternatives are specific services of the available transit modes together with the private car. Among the attributes of transit services we propose inclusion of the difference between the desired arrival time of the traveller and the arrival time of the specific service (early/late arrival penalty), instead of service/line frequency. Different utility specifications were tested and their results compared for different correlation structures (both nested and cross-nested). In general a high significance was observed in the case of the early/late arrival penalty parameters, indicating the importance of such factors in mode choice.

THE SCHEDULE-BASED DYNAMIC MODELLING FOR PUBLIC TRANSPORT NETWORKS: A NEW APPROACH IN PATH CHOICE AND ASSIGNMENT

This paper presents a general overview of the schedule-based approach in dynamic path choice and assignment models for transit nehvorks. The first part deseribes different specification of path choice models in relation to different user and seMce characteristics. The latter deals with a classification of assignment models.

URBAN FREIGHT MOVEMENT: A QUANTITY ATTRACTION MODEL

The paper describes a multi-step model that considers different decision makers, quantity attraction and distribution models, an acquisition model, models for the choice of service and vehicle type, and path choice model. To obtain preliminary results, some surveys were conducted in a medium size city. The paper specifies and calibrates an attraction model for perishables and household products in terms of quantity and for durables in terms of good trips.

Transit Path-Choice Models That Use Revealed Preference and Stated Preference Data

A study was conducted of the trip-making behavior of transit users in a complex transit network with input of travel information in an increasingly complex and information-rich transit environment. A modeling framework to assess the preferences of commuters toward a range of factors and alternatives was produced. Path-choice models on transit services of Singapore were developed using mixed revealed preference (RP) and stated preference (SP) data. RP models were estimated using data obtained in the 1997 Household Interview Survey. An SP survey based on 44 scenarios was conducted, which yielded 307 responses at 12 different locations in Singapore. Models based on pure RP data, pure SP data, and mixed RP and SP data were estimated. For transit users, the decision of which path to take relies on the trade-offs among attributes of journey time and other factors. The significant attributes in the models indicate that the transit path choice is a multi-criteria decision-making process. This suggests that the usual way of transit path assignment in many planning models based on single criterion, such as minimum time, may not be appropriate in choice situations. Path-choice models developed and calibrated provide important fundamental knowledge of the way transit services are to be utilized by travelers, for use in a wide range of applications, such as multimodal transit service planning models, transit demand modeling, and integration of the transit path-choice models in intelligent transportation systems applications.

A Doubly Dynamic Schedule-based Assignment Model for Transit Networks

In order to improve the performances of transit modeling and in particular to assess the effects of the introduction of ITS technologies, which imply substantial changes in user behavior, in the last 10 years transit path choice and assignment models that use a schedule-based approach, instead of the traditional frequency-based one, have been developed. This paper presents a schedule-based path choice model for high-frequency transit networks, which allows us to consider the evolution in time of transit services, both within-day and day-to-day, as well as the day-to-day learning process of attributes by which users choose. On the basis of this path choice model a dynamic process assignment model, both within-day and day-to-day (from which the term “doubly dynamic” derives), was developed and tested on a realistically sized network to verify its applicability for operations planning.

A DOUBLY DYNAMIC ASSIGNMENT MODEL FOR REGULAR TRANSIT SYSTEMS

This paper, with reference to high frequency transit systems with waiting time user information and with regular arrival of services at stops, proposes a dynamic stochastic path choice model that explicitly considers within-day variation and day to day evolution of the system. The path choice model is used together with a diachronic network model in a transit doubly dynamic assignment model, which is analyzed with a deterministic approach, as a non linear dynamic system.

Stochastic equilibrium traffic assignment with value-of-time distributed among users

In this paper, existing fixed-point models for stochastic equilibrium traffic assignment to transportation networks are extended to deal with the value of time distributed among users. This problem is relevant when transporation demand management measures affecting travelling monetary costs are to be evaluated. Analysis of existence and uniqueness of solutions is carried out, leading to mild assumptions about link cost-flow functions, path choice model, and value-of-time distribution. Some quite simple solution algorithms are also specified and their convergence analysed. Results of some applications are also reported.

Stochastic Equilibrium Traffic Assignment with Value‐of‐time Distributed Among Users

In this paper, existing fixed‐point models for stochastic equilibrium traffic assignment to transportation networks are extended to deal with the value of time distributed among users. This problem is relevant when transporation demand management measures affecting travelling monetary costs are to be evaluated. Analysis of existence and uniqueness of solutions is carried out, leading to mild assumptions about link cost‐flow functions, path choice model, and value‐of‐time distribution. Some quite simple solution algorithms are also specified and their convergence analysed. Results of some applications are also reported.

Integrated Models for Simulating the Italian Passenger Transport System

Abstract In this paper, a system of models simulating intercity passenger travel demand in Italy is presented. The models are part of a wider project for a Decision Support System at the national level (SIMPT) covering both passenger and freight transport systems.The core of the SIMPT is composed by integrated demand, supply and assignment models. Behavioural demand models, calibrated on a large RP survey, tour frequency, destination, mode and path choices for different purposes, days and periods of the year. Supply models consist in infrastructural (roads, railways) and a service (railway services, extraprovincial buses, airlines) networks. The road assignment model is a pseudo-within-day-dynamic SUE with explicit path enumeration and takes into account the calibrated path choice models for different time slices, while assignment to transit services is a SNL line-based modal.