National
Institute of Statistical Sciences
19 T. W. Alexander
Drive
P.O. Box 14006
Research Triangle Park, NC 27709-4006
Tel: 919.685.9300
FAX: 919.685.9310
admin@niss.org
PROGRESS REPORT (DMS-9313013)
Measurement, Modeling and Prediction for Infrastructural Systems
Alan F. Karr, Eric I. Pas, Jerome Sacks, Principal Investigators
May 16, 1997
Contents
Travel Demand
Intelligent Transportations Systems
Materials Science and Deterioration of Concrete
References
This work aims at understanding the relationships between sociodemographics and activity-travel behavior, to support the development of activity-travel patterns for synthetic populations.
A structural equation model system has been developed Lu & Pas (1997) to predict travel behavior (number of trips, number of tours, travel time and car mode share) from sociodemographics, activity participation (both in-home and out-of-home). This work reveals important relationships and shows that a model system that excludes activity participation leads to incorrect inferences regarding the influence of sociodemographics on travel behavior. This work is now being extended to incorporate measurement models of the latent concepts of lifestyle, life cycle and household role.
Another effort -- to identify the important sociodemographic influences on activity-travel behavior -- has required developing a multivariate version of the CART methodology. Initial results appear promising, and work is continuing.
This work is led by Pas, P. Speckman (Statistics, Missouri), and also involves D. Sun (Statistics, Missouri), X. Lu (graduate student, Civil Engineering, Duke) and P. Koskenoja (postdoc, NISS).
Two complementary approaches have been and continue to be pursued.
One, first articulated in Kitamura (1995), models the daily activity-travel pattern of an individual as a Markovian sequence of activities, each described by type, duration and location. A model system to generate synthetic travel patterns has been developed and estimated [Kitamura, Chen & Pendyala (1997)]. Currently, a computer code is being developed to form a microsimulator that can be used as a tool for forecasting and policy analysis, as well as to generate travel patterns for synthetic households [Pas, et al. (1997)].
A second approach, outlined in Pas (1996a), Pas (1996b), aims to synthesize daily activity-travel patterns for all members of each household. This approach assumes that a daily activity-travel pattern has a basic skeletal structure (comprising, for example, the number of activities and tours), that imposes constraints on more detailed aspects of the pattern. This approach synthesizes the activity-travel patterns of the members of a synthetic household by sampling the skeletal structure from observed activity-travel patterns (from a travel behavior survey) and then simulating the details of the activity-travel patterns of household members by models based on observed probability distributions. In the initial work, models of location choice (for activities) are being used in creating the synthesized activity-travel patterns.
This work, led by Pas, involves R. Kitamura (Civil Engineering, Kyoto, and RDC, Inc.), Koskenoja, Lu, Speckman, Sun and K. Vaughn (former postdoc, NISS, now with the Federal Transit Administration).
Models of location choice for activities are a necessary component of generators of activity-travel patterns for synthetic populations. A NISS-developed nonparametric technique for estimation of a continuous version of the gravity model [Speckman (1997)], initially applied to (home,work) location pairs in Portland, OR, has been extended to model shopping locations, conditional on home and work locations. Sample output, showing an estimated distribution of shopping locations, appears below.
A hierarchical Bayesian model has been developed for work location conditional on home location. This basic framework is to be extended during the coming year to incorporate sociodemographic and other covariates, as well as trip chaining.
Key participants are Pas, Speckman, Sun and Lu, with collaboration from R. Wolpert and K. Ickstadt (both, Statistics and Decision Sciences, Duke).
This component of the project addresses econometric models of several aspects of activity-travel patterns, in order to identify important relationships and to provide policy analysis tools; these models can also be used in activity pattern generation. Emphases during the past year included:
This work is being led by C. Bhat (Civil Engineering, Massachusetts) and F. Koppelman (Civil Engineering, Northwestern), and also involves graduate students V. Pulugurta (Civil Engineering, Massachusetts) and C. Wen (Civil Engineering, Northwestern).
During the coming year, a number of extensions of the above efforts are planned:
Research on ITS addresses estimation of travel times in large, urban street networks, estimation of travel times on freeways and network flow models. Additional lines of investigation have been initiated.
A central activity was to complete analysis of data from a field experiment conducted using ADVANCE (Once planned to be a multi-year, 3000-vehicle deployment, ADVANCE was "downsized" to a very limited deployment.) probe vehicles and data collection mechanisms during the summer of 1995. These data, focused on one principal ("study") link on westbound Dundee Road in suburban Chicago, include (1) probe travel times, link-by-link (And also congested times and distances -- time spent at speed less than 2 meters per second and distance traveled at speed less than 10 meters per second.); (2) video tapes of the signal status at the downstream intersection; and (3) detector data, aggregated at five-minute intervals, for all lanes approaching the downstream intersection.
Models were developed that predict travel on the study link as a function of the relativized entry time to the link (Entry time relative to the signal cycle, with zero corresponding to arrival at the signal just as it turns green.) as depicted below. Results are reported in Graves, Thakuriah & Karr (1997).
These models reveal unambiguously the central role of signal status within cycle as predictor of travel times. In reality, though, such data are rarely available. Probe vehicles, however, for which travel time is known, may be used to impute signal status. One method, in effect, inverts the estimated function shown above to obtain signal status from probe travel times.
This method is effective, however, only over short time intervals, because the length of the signal cycle and individual phases change in response to traffic conditions. An alternative method Thakuriah (1996a), based on dynamic clustering of probes, is startlingly effective, as shown below.
Motivated by and using data from the ADVANCE project, Sen has led a continuing effort on other aspects of travel time prediction, focused on
Principal sites for this activity are NISS headquarters and the University of Illinois at Chicago (UIC). Karr and A. Sen (Urban Planning and Urban Transportation Center, UIC) have led it; other participants are T. Graves (postdoc, NISS), P. Thakuriah (postdoc, NISS, and now Urban Planning, UIC), and X. Zhu (graduate student, Urban Planning, UIC).
Using a very rich set of single loop detector data collected by the PATH (Partners in Advanced Highways) Project (Berkeley, CA), methods have been developed for accurate estimation of travel times on freeways. The PATH data comprise count and occupancy, at a time resolution of one second, separately for each lane, for some 20 locations on Interstate Highway 880 in Oakland, CA. Successive locations are approximately one-third mile apart.
Unlike other procedures, the new methods estimate the entire travel time distribution, rather than a single summary measure such as the mean. Critical steps were to limit the support of the estimated distribution and dynamically to use the "naive" speed estimate (count divided by the product of occupancy and a nominal vehicle length) to center the estimated distribution.
Data on velocity from (a small number of) double loop detectors and from probe vehicles were used to assess the accuracy of travel time estimates based on single loops.
The results are reported in Bickel, et al. (1996).
Continuing efforts address travel times over multiple links.
This component of the project is housed at the University of California, Berkeley (UCB). Participants are P. Bickel (Statistics, UCB), J. Rice (Statistics, UCB), Y. Ritov (Statistics, Hebrew National), J. Jiang, K. Petty (graduate student, Civil Engineering, UCB) and F. Schoenberg (graduate student, Statistics, UCB), with the cooperation of P. Varaiya (Civil Engineering, UCB, and Director, PATH).
Bayesian methods have been applied to to infer origin-destination data from link flows [West & Tebaldi (1996)]. These methods originated in the context of communication networks, and -- despite daunting computational requirements -- show strong promise.
Based on initial results [Daley (1995)], Daley has completed studies of link-to-link dependence in the Dundee Road network technical reports are in preparation.
These activities have been centered at NISS and Duke University, and have involved D. Daley (Statistics, Australian National), M. West (Statistics and Decision Sciences, Duke) and C. Tebaldi (graduate student, Statistics and Decision Sciences, Duke).
An important issue for management of freeways is recognizing incipient congestion. Methods have been developed to predict the onset of freeway breakdown, a readily observed but not easily characterized phenomenon that occurs when, in heavy traffic and without an external cause (such as an accident or lane closure), flows and speeds drop rapidly and dramatically, across all lanes. From the standpoint of congestion management, the central need is to recognize breakdown before it occurs (in order to implement control measures such as changeable speed limits or access restrictions).
Over the summer of 1996, a unique data set was assembled by NISS (with cooperation
of transportation engineers from the Washington State Department of Transportation):
in real time and over the Internet, counts and occupancies were obtained from
more than 30 single loop detector stations on a 10-mile stretch of I-5 in northern
and north suburban Seattle, WA. Data were collected for weekday morning and
afternoon peak periods, amounting to ten hours per day, over two months.
(
Real-time video data for I-5 are available over the World Wide Web.)
Visualizations of the data, exemplified below, revealed a clear structure for breakdowns: a breakdown propagates upstream (as it is fed by demand) and clears downstream (as demand slackens). (The direction of traffic flow is downward.)
Thus, if breakdown as defined as "speed less than some threshold (e.g., 30 mph)," in space-time, it has a characteristic triangular structure. Using this definition (with speed estimated "naively" from count and occupancy), CART-based methods have been constructed that predict breakdown with (from the perspective of traffic engineering) startling accuracy: using 10-minute blocks of data, more than 60 per cent of breakdowns are predicted, with a false alarm rate of less than 5%.
A typical classification tree is shown below. The results are reported in Click, et al. (1996); additional manuscripts are in preparation.
Karr and N. Rouphail (Civil Engineering, North Carolina State University) have led this effort, which also involved T. Graves (postdoc, NISS), V. Thakuriah (postdoc, NISS) and S. Click (graduate student, Civil Engineering, North Carolina State University).
Issues similar to those for freeway breakdown arise in a safety context as well, treating accidents rather than flow as the response points to need for recognizing incipient unsafe conditions. Although mildly suggestive that accidents occur under high flow--low speed (i.e., congested) conditions, results such as those shown below, where the plot is of (count,speed) conditions for approximately 70 accidents together with an equal number of conditions selected randomly from the entire (two-month) data set, point mainly to the need to additional data, especially covariates (e.g., weather and road geometry).
This question was studied by Karr and Graves collaboratively with Rouphail and R. Hughes (Highway Safety Research Center, University of North Carolina at Chapel Hill).
Emphases of this component of the project were:
D. Boyce (Civil Engineering and Urban Transportation Center, UIC) leads this effort; other participants are B. Janson (Civil Engineering, University of Colorado at Denver), P. Mirchandani (Industrial Engineering, University of Arizona), M. Tatineni (postdoc, UIC) and X. Tian (graduate student, UIC).
Other ITS issues addressed include:
Three new research directions have been established. They not only expand the the project in timely ways, but also leverage other activities at NISS.
Evaluation of Algorithms for Dynamic Signalization. Dynamic control of traffic signals, especially using new sensor technologies (e.g., infrared or video) not subject to the same unreliabilities as street-embedded loop detectors is a key issue in ITS.
Importance notwithstanding, methods do not exist for prospective evaluation of control algorithms and data sources. Many pieces (network flow models such as CORSIM, e.g.) are available, but lack the hooks to allow statistically based evaluation. Also lacking are means to simulate some forms of data, techniques to combine empirical data and numerical experiments and "benchmark" control algorithms to which alternatives can be compared.
Negotiations are under way with the Chicago Department of Transportation (CDOT) for data collection and potential experimentation with benchmark algorithms.
Key participants are Karr, Rouphail, Sacks, Sen, S. Stidham (Operations Research, University of North Carolina at Chapel Hill), Thakuriah and T. Ozturkkul (graduate student, Operations Research, University of North Carolina at Chapel Hill).
Estimation of Emissions from Traffic Data. Planned for the Fall of 1997 is a pilot experiment to explore the feasibility of predicting emissions (of carbon monoxide and hydrocarbons) by vehicles from aggregated traffic data (counts and speeds). If the experiment succeeds, the resultant technology will be of significant value to transportation planners and engineers seeking to measure air quality or to quantify air quality implications of various policies and control measures. The experiment will use roadside infrared sensors to measure emissions of individual vehicles and the Mobilizer video imaging and analysis system to assemble detailed traffic data from several locations in Chapel Hill, NC, to which models will be fit. The crucial issues will then be to assess loss of predictive capability when only restricted data are available. Key issues will be the effects of two uncontrollable (and for most purposes, unobservable) variables: vehicle (car/truck, in simple case) mix (Needed to convert infrared emission measurements to masses emitted.) and the proportion of "cold start" vehicles (For which emissions are dramatically higher than those for "warm" vehicles.). Validation will involve fitting a model with data from one site and testing its predictive capability at another.
Those involved will be C. Frey (Civil Engineering, North Carolina State University), Karr, R. Ranjithan (Civil Engineering, North Carolina State University), Rouphail, Sacks, and a NISS postdoc.
Model Validation. Beginning in the summer of 1997, we will estimate and validate a regional travel forecasting model at the level of detail required by transportation planning organizations to forecast traffic, ridership and vehicle emissions. Specifically, a detailed model of the Chicago region will be implemented, estimated and validated at a scale that is realistic for planning -- 1500 zones, four trip purposes, two modes (automobile and an integrated transit mode), and two time intervals (peak and shoulder) during the morning or evening peak commuting period. Such values go well beyond what has been feasible previously.
Primary data for estimating model parameters describing choice behavior will come from the Chicago Area Transportation Study (CATS) 1990 home interview survey. Validation data include the 1990 Census of Population and Housing, highway and transit flow data collected by CATS and the Regional Transportation Authority (RTA). Criteria for validation we be regional trip length distributions and vehicle miles of travel.
Sensitivity analyses, propagation of error through a bi-level optimization model and computation are primary issues to be confronted.
Because the mechanism of chloride permeability of concrete is not fully understood, designed experiments and statistical models are perhaps the only viable path to find factors that influence permeability so that concrete mixes can be designed to have maximum resistance to chloride ion penetration.
Two pilot studies of the effects on chloride permeability of design ("mix") variables (principally the volume fraction and size distribution of aggregate and the water-cement ratio) and microstructural variables (specimen-level volume fraction and size distribution of aggregate, the separation between aggregates, the paste radius, the area of the aggregate-paste interface and measures of tortuosity) have been completed.
One study Karr, et al. (1997), which employed spherical alumina aggregate, allowed exact reconstruction of three-dimensional of sliced specimens, using NISS-developed software. Even with limited data, tortuosity -- the average length of an ion diffusion path through the specimen, accounting for obstacles posed by the aggregate -- proved an important predictor.
The second pilot study, using "real" aggregate but a gapped size distribution, became a prelude (and, in particular, provided important preliminary information on variability) to one principal activity during the year, a 120-specimen experiment with water-cement ratio, volume fraction of aggregate and aggregate grade (selected from several standard size distributions) as the design variables. Two-inch thick, four-inch specimens to be tested were selected from various locations in deeper (8-inch) cylinders.
A full factorial design was used, utilizing 24 different compositions of concrete. Because variability between batches was potentially as important as the variability between cylinders cast from each batch, multiple batches of some compositions were made; location of the specimen within its parent cylinder is yet another source of variability. The time-consuming nature of the experiments, however, precluded full replication to explore these sources completely. Instead, a novel design was devised, which required either the casting of multiple cylinders from a particular batch or the mixing of multiple batches for each composition. Currently, the design is being compared on a theoretical basis to designs in which both (inter- and intra-batch) variance components are estimated for all compositions, with a manuscript in preparation Ankenman, et al. (1997). (Location-induced variability is removed by specimen-level measurements of microstructural variables.) The results will be transferable to numerous other contexts.
The experimental work employed techniques developed during the pilot studies. In these studies, air voids formed in some specimens, and could only be eliminated by vibration of the concrete before setting. Because this vibration induced the settling of coarse aggregate within cylinders, batch-level volume fractions are adequate predictors of specimen-level volume fractions. Instead, stereological estimates of the volume fraction of (coarse) aggregate have been constructed, which involve intricate image manipulations to convert raw images such as
to segmented versions such as
In addition to their being incorporated in models for permeability, the image-derived data will be used to assess the fit of random set models for simulating the structure of concrete.
Because physical experiments are costly and time-consuming, a "numerical experiments" component of the research is being initiated. The initial focus is on random set models that accurately simulate the micro- and macro-structure of concrete. These models will be used to investigate tortuosity, which is hypothesized to be the scientific link between structure and permeability: coarse aggregate contort the cement phase of the concrete into a complex network of channels of varying widths, which impede the diffusion of chloride ions, thereby decreasing permeability. The goal is a quantitative measure of tortuosity that predicts permeability. Existing models (from NIST, for example) for ion diffusion in concrete will comprise the permeability part of the numerical experiments.
A second major focus is permeability of cracked samples of concrete, in order to understand the feedback between permeability and deterioration. An initial study is reported in Wang, et al. (1996).
A more substantial follow-on is now in the experimental stage. This phase involves cracking slices from cylinders of four different cement-based materials (cement paste, cement with sand, normal strength cement and high strength cement) under controlled displacement conditions, then testing specimens not only for chloride permeability but also water permeability (using a much slower, less precise test). The experimental design varies material, specimen thickness and maximum crack opening displacement, but does not allow for variance component estimation. Efforts are now underway to define a suitable measure of water permeability, and to develop methods to describe crack structure. A further extension of the experiment is planned that investigate the effects of larger cracks, particularly in the case of the more chloride-resistant high strength concrete.
This work has been overseen by Karr and S. Shah, Director, Center for Advanced Cement Based Materials (ACBM), Northwestern University. Other participants are C. Aldea (postdoc, ACBM), B. Ankenman (Industrial Engineering and Management Sciences, NU), T. Igusa (Civil Engineering, NU), H. Lu (graduate student, Industrial Engineering and Management Sciences, NU), J. Picka (postdoc, NISS), K. Wang (postdoc, ACBM) and S. Jaiswal (graduate student, Civil Engineering, NU).
Continuing earlier practices, meetings of personnel associated with specific components of the project are held regularly at NISS headquarters in Research Triangle Park, NC, the University of Illinois at Chicago, Northwestern University and the University of California Berkeley.
Conference presentations featuring activities under this project have been given at the national meeting of INFORMS (May, 1996; Washington); the Activity-Based Travel Forecasting Conference (June, 1996; New Orleans); the Workshop on the Relationship Between GIS and Behavioral Travel Modeling (June, 1996; Santa Barbara); the Joint Statistics Meetings (August, 1996; Chicago, IL); the Conference on Theoretical Foundations of Travel Choice Modeling (August 1996; Stockholm); the Joint Statistical Meetings (August, 1996; Chicago); the Fourth World Congress of the Bernoulli Society (August, 1996; Vienna); the 36th Congress of the European Regional Science Association (August, 1996; Zurich); the national meeting of INFORMS (November, 1996; Atlanta); and the Transportation Research Board (January, 1997; Washington). The last of these, the major national transportation meeting, featured more than a dozen talks on NISS research.
A workshop on travel demand forecasting and activity-travel behavior was held at NISS on May 8-9, 1997.
Ankenman, B. E., Karr, A. F., Lui, H., and Picka, J. D. (1997). Experimental designs for estimating a response surface and variance components. Technical Report, National Institute of Statistical Sciences. (in preparation)
Bhat, C. R. (1996). Incorporating observed and unobserved heterogeneity in urban work mode choice modeling. Submitted to Transp. Sci.
Bhat, C. R. (1997a). Work mode choice and number of nonwork commute stops. Transp. Res.31B 41-54.
Bhat, C. R. (1997b). Accommodating flexible substitution patterns in multi-dimensional choice modeling: formulation and application to travel mode and departure time choice. Submitted to Transp. Res.
Bhat, C. R. (1997c). An analysis of travel mode and departure time choice for urban shopping trips. Submitted to Transp. Res.
Bhat, C. R., and Misra, R. (1997). Discretionary activity time allocation of individuals between in-home and out-of-home and between weekdays and weekends. Submitted to Transportation.
Bhat, C. R., and Pulugurta, V. (1997). A comparison of two alternative behavioral mechanisms for car ownership decisions. Transp. Res. (to appear)
Bickel, P., Jiang, J., Ostland, M., Petty, K., Rice, J., Ritov, J., and Schoenberg, F. (1996). Accurate estimation of travel time from single loop detectors. Transp. Res. Rec. (to appear)
Boyce, D. E., Lee, D.-H., Janson, B. N., and Berka, S. (1997). Dynamic user-optimal route choice modeling of a large-scale traffic network. ASCE J. Transp. Engrg. (to appear)
Boyce, D. E., Lee, D.-L. and Janson, B. N. (1996a). Roadway incident analysis with a dynamic user-optimal route choice model. Submitted to Beckmann, M., et al. , eds., Knowledge and Networks in a Dynamical Economy. Springer-Verlag, Heidelberg.
Boyce, D. E., Lee, D.-H., and Janson, B. N. (1996b). A variational inequality model of an ideal dynamic user-optimal route choice problem. Submitted to Bell, M. G. H., ed., Proceedings of the Euro96 Working Group on Transportation. Elsevier, London.
Boyce, D. E., and Mattsson, L.-G. (1996). Formulation and solution of IMREL as a network equilibrium model. Submitted to Papers Regional Sci.
Boyce, D. E., and Zhang, Y. (1996a). Calibrating a combined model of trip distribution, modal split and assignment. Submitted to Transp. Res. Rec.
Boyce, D. E. and Zhang, Y. (1996b). Parameter estimation for combined travel choice models. Submitted to Lundqvist, L., Mattsson, L.-G., and Kim, T. J., eds., Network Infrastructure and the Urban Environment. Springer-Verlag, Heidelberg.
Click, S. M., Rouphail, N. M., Hughes, R., and Graves, T. L. (1996). Using advanced vehicle monitoring systems to extend system capacity along North Carolina freeways. Final Report, Center for Transportation Engineering Studies, Department of Civil Engineering, North Carolina State University.
Daley, D. J. (1995). Some statistical properties of link travel times. Technical Report, National Institute of Statistical Sciences.
Graves, T. L., Thakuriah, P., and Karr, A. F. (1997). Effect of signals and volume on arterial travel times. Technical Report, National Institute of Statistical Sciences. (in preparation)
Karr, A. F., Shah, S. P., Styer, P., Igusa, T., Jaiswal, S., and Picka, J. (1997). Permeability and microstructure relationships for spherical alumina aggregates in cement. Technical Report, National Institute of Statistical Sciences.
Karr, A. F., Shah, S. P., Ankenman, B. E., Igusa, T., Picka, J., Wang, K., and Jaiswal, S. (1997). Models for chloride permeability of concrete. Technical Report, National Institute of Statistical Sciences. (in preparation)
Kitamura, R. (1995). Generation of synthetic daily activity--travel patterns: outline of the approach. Technical Report, National Institute of Statistical Sciences.
Kitamura, R., Chen, C., and Pendyala, R. M. (1997). Generation of synthetic daily activity--travel patterns. Transp. Res. Rec. (to appear)
Koppelman, F. S., and Wen, C.-H. (1997a). Alternative nested logit models: structure, properties and estimation. Submitted to Transp. Res.
Koppelman, F. S., and Wen, C.-H. (1997). The paired combinatorial logit model: properties, estimation and application. Transp. Res. B (to appear)
Lawton, T. K., and Pas, E. I. (1996). Resource paper: Survey methodologies workshop. In Conference on Household Travel Surveys: New Concepts and Research Needs, Conference Proceedings 10. Transportation Research Board, Washington.
Lee, D.-H., Boyce, D. E., and Janson, B. N. (1996a). Formulation and solution of a dynamic user-optimal route choice model on a large-scale network. Technical Report, National Institute of Statistical Sciences.
Lee, D.-H., Boyce, D. E. and Janson, B. N. (1996b). Solution of a large-scale, dynamic route choice model and an application to roadway incident analysis. Submitted to Transp. Res. C.
Liu, L. N., and Boyce, D. E. (1996). A variational inequality formulation of optimal congestion pricing in a general transportation network. Submitted to Regional Sci. Urban Econ.
Lu, X., and Pas, E. I. (1997). A structural equation model of the relationships among socio-demographics, activity participation and travel behavior. Submitted to Transp. Res. A.
Metaxatos, P., and Sen, A. (1997). Trip chaining models: theoretical framework and empirical analysis. Submitted to Transp. Res. Rec.
Nobile, A. (1995). A hybrid Markov chain for the Bayesian analysis of the multinomial probit model. Submitted to J. Amer. Statist. Assoc.
Nobile, A. (1997). Some advances in simulating contingency tables with given margins. Technical Report, National Institute of Statistical Sciences. (In preparation)
Nobile, A., Bhat, C. R., and Pas, E. I. (1996). A random effects multinomial probit model of car ownership choice. In Gatsonis, C., et al. , eds., Proceedings of the Third Workshop on Bayesian Statistics in Science and Technology: Case Studies. Springer, New York. (To appear)
Noland, R. B., Small, K. A., Koskenoja, P. M., and Chu, X. (1996). Simulating travel reliability. Submitted to Reg. Urban Econ.
Pas, E. I. (1996a). Time and travel demand Modeling: Theory, data collection and models. Proceedings, Conference on Theoretical Foundations of Travel Choice Modeling. (to appear)
Pas, E. I. (1996). Advances in activity-based travel modeling. In Proceedings of the Conference on Activity-Based Travel Forecasting. (to appear)
Pendyala, R. M., Kitamura, R., Chen, C., and Pas, E. I. (1997). Activity-based microsimulation analysis of transportation control measures. Transp. Policy (to appear)
Ran, B., and Boyce, D. E. (1996a). Modeling Dynamic Transportation Networks, 2nd revised edition, Springer-Verlag, Heidelberg.
Ran, B., and Boyce, D. E. (1996b). Modeling dynamic transportation networks with variational inequalities. Submitted to Nijkamp, P., and Emmerink, R., eds., Behavioural and Network Impacts of Driver Information Systems, Springer-Verlag, Heidelberg.
RDC, Inc. (1996). An Exploration of AMOS Survey Time-Use and Travel Data Set.
Sen A., and Thakuriah, P. (1995). Estimation of static travel times in a dynamic route guidance system. In Rodin, E. Y., ed., Network, Control, Communication and Computing Technologies for Intelligent Transportation Systems, a special issue of Mathematical and Computer Modelling 22 83-102.
Sen, A., Soot, S., and Christopher, E. (1995). Household travel survey non-response estimates: the Chicago experience. Transp. Res. Rec. 1443 170-177.
Sen, A., Thakuriah, P., Zhu, X., and Karr, A. F. (1996). Frequency of probe reports: variances of link travel time estimates. ASCE J. Transp. Engrg. (to appear)
Sen, A., Soot, S., Ligas, J., and Tian, X. (1997). Arterial link travel time estimation: probes, detectors and assignment-type models. Submitted to Transp. Res. Rec.
Sen, A., Soot, S., Thakuriah, P., and Condie, H. (1997). Estimation of static travel times in a dynamic route guidance system. J. Math. Computer Modelling (to appear)
Sen, A., and Zhu, P. (1997). Static profiles using piecewise linear functions. Technical Report, National Institute of Statistical Sciences.
Soot, S., Sen, A., Marston, J., and Thakuriah, P. (1995). Multiworker household travel demand. In National Personal Transportation Survey, Demographic Special Reports, 1990 NPTS, Chapter IV.
Speckman, P., Pas, E. I., and Vaughn, K. (1997). A continuous spatial interaction model: application to home-work travel in Portland, OR. Technical Report, National Institute of Statistical Sciences. (in preparation)
Tatineni, M., Boyce, D. E., and Mirchandani, P. (1996a). Solution properties of stochastic route choice models. Technical Report, National Institute of Statistical Sciences.
Tatineni, M., Boyce, D. E., and Mirchandani, P. (1996b). Comparisons of deterministic and stochastic traffic loading models. Transp. Res. Rec. (to appear)
Tatineni, M., Boyce, D. E., and Mirchandani, P. (1996c). Solution properties of route choice models: Some comparisons. Submitted to Transp. Res. B.
Thakuriah, P. (1996a). Link travel time estimation with imputed signal control information. Technical Report, National Institute of Statistical Sciences.
Thakuriah, P. (1996b). Planning with intelligent transportation systems. Submitted to J. Amer. Planning Assoc.
Thakuriah, P., and Sen, A. (1996). Quality of information given by advanced traveler information systems. Transp. Res. C 4 249-266.
Thakuriah, P., Sen, A., and Karr, A. F. (1996a). Analysis of probe-based information on signalized arterials. Technical Report, National Institute of Statistical Sciences.
Thakuriah, P., Sen, A., and Karr, A. F. (1996b). Quality of real-time traffic information using probe vehicles. In Behavioral and Network Impacts of Driver Information Systems. Wiley, New York. (to appear)
Thakuriah, P., Sen, A., Soot, S., and Christopher, E. J. (1996). Non-response and urban travel models. Transp. Res. Rec. 1551 82-87.
Thakuriah, P., Sen, A., and Soot, S. (1997). Effects of recurrent and non-recurrent congestion on arterial streets: implications for traffic management and driver information systems. Submitted for presentation at the 1997 annual meeting of the Intelligent Transportation Society of America.
Vaughn, K., and Pas, E. I. (1996). Report of a workshop on activity-travel pattern modeling. Technical Report, National Institute of Statistical Sciences.
Wang, K., Jansen, D. C., Shah, S. P., and Karr, A. F. (1997). Permeability study of cracked concrete. Cement Concrete Res. 27 381-393.
West, M., and Tebaldi, C. (1996). Bayesian inference on network traffic using link count data. Submitted to J. Amer. Statist. Assoc.
 |
 |
| Help |