A model for allocating passenger traffic in a multiple-airport system is proposed and tested. The model treats the supply side of the system as endogenous, and represents it implicitly using the principle that as traffic using an airport increases, the airport becomes more attractive. This positive feedback effect and the preference of air travelers for airports near their trip terminus are argued to be the key factors determining the distribution of traffic in a multiple-airport system. The model is applied to the San Francisco Bay Area, served by airports in Oakland, San Francisco, and San Jose. A disaggregate logit airport-choice model, in which airport utility is a function of travel time to trip origin, airport traffic in the origin-and-destination market, and airport enplanements, is estimated. The calibrated choice model, combined with estimates of the distribution of trip origins, is used to calculate equilibrium market shares. Comparison of equilibria with observed shares shows high correlation, particularly for larger markets, and for the San Francisco and Oakland airports.

The performance of rubber (from used tires) as an asphalt cement modifier was evaluated by performing experiments on the shear-flow properties and creep-rupture behavior of modified asphalt cements. Results indicate that various asphalt/rubber samples exhibit shear-thickening, Newtonian, and shear-thinning behavior depending on the shear rate. The shear viscosity, linear viscoelastic functions, elasticity, and creep resistance of asphalt cement increased with the addition of rubber. These increases indicate that the addition of rubber will improve the low temperature properties of asphalt cement, and that such modifications are necessary to reduce the tendency of asphaltic-paving materials to crack in cold climates and to increase the life cycle of the road.

A structural deterioration relationship is developed using the U.S. Army Corps of Engineers accelerated traffic tests for plain concrete airfield pavements. The relationship predicts the performance of the rigid pavement in terms of the structural condition index from the pavement's Westergaard edge stresses, flexural strength, and number of load applications. The model is compared to other design criteria and unfailed test sections and gives reasonable agreement. The relationship between the structural condition index and the coverages at initial failure of the pavement is determined, and an alternative fatigue curve is suggested for use in the thickness design of rigid airfield pavements with joints capable of load transfer. The relationships developed only account for the tensile fatigue-loading-induced deterioration of the rigid airfield pavement. Actual in-service pavements will show additional deterioration because of other jointed rigid pavement distresses such as pumping, settlement or faulting, durability D cracking, pop-outs, scaling, or joint seal damage.

No-fines concrete is made from a uniformly graded coarse aggregate and a cement-water paste. Because of its excellent drainage properties, no-fines concrete is found to be beneficial in areas that receive frequent and sometimes excessive rainfall. This paper discusses thickness design of no-fines concrete parking lot pavements. Based on the engineering properties of the no-fines concrete mixtures developed in the laboratory, and various traffic conditions and subgrade characteristics, the thickness requirement of no-fines concrete pavements was determined. Two design procedures, American Association of State Highway and Transportation Officials (AASHTO) and Portland Cement Association (PCA), were used for the thickness evaluation of no-fines concrete parking lots. Test results obtained in the laboratory indicate that, with proper proportioning and densification, no-fines concrete can be successfully utilized as a surface paving material for the construction of parking-lot pavements. The thickness design tables indicate that PCA design procedure is more reasonable for thinner pavements, whereas AASHTO yields a more conservative outcome for thicker pavements. For intermediate traffic categories, both design methods provided similar thickness results.

A new whole-life pavement-performance model (WLPPM) that is capable of making deterministic pavement-damage predictions due to realistic traffic and environmental loading has been developed. A vehicle simulation is used to generate dynamic tire forces that are a function of distance along the pavement. These dynamic tire forces are then combined with the appropriate pavement primary-response influence functions to give primary-response histories at regularly spaced points along the pavement. The primary-response histories are then transformed into an increment of damage at each point along the pavement due to a single vehicle pass, and the surface-roughness profile is updated. This procedure is then repeated until pavement failure occurs. Using the WLPPM, the performance of two classes of pavement are examined operating under climatic conditions typical of the United Kingdom. It is concluded that both dynamic vehicle loads and asphalt-layer stiffness variations can significantly influence long-term flexible-pavement performance.

A mechanistic method for predicting the response of concrete pavements overlying a multilayer soil system is proposed. The method incorporates a finite-element discretization of a concrete slab and a variable foundation support that depends on the nonlinear resilient moduli of the underlying soil layers. The method allows for determination of excess pore-water pressure in saturated granular layers and the corresponding loss of foundation support under the pavement structure. To illustrate the idea, the proposed method was applied to investigate the behavior of a concrete pavement supported by granular bases over a soft clay layer. Moist-compacted and saturated conditions were examined for typical dense-graded and open-graded unbound granular bases. Pavement deflections, stresses, excess pore pressure, and compression settlement associated with pore-pressure dissipation for the different support conditions were evaluated and compared.

Resilient-modulus tests were performed on a laboratory compacted dune sand following the standard test procedure described in the 1982 AASHTO Designation T 274-82 . The samples were prepared by impact- and vibratory-compaction methods. Impact-compacted samples were prepared using standard Proctor energy. Vibratory-compacted samples were prepared using a vibration table generating vibrations of 60 Hz at a peak-to-peak amplitude of 1.73 mm (0.0681 in.). Compaction curves by impact and vibratory compaction are identical at lower water contents, but vibratory compaction produces comparatively higher dry density at increased water content. At the same water content and dry density, a vibratory-compacted specimen exhibits significantly less permanent deformation and about 40% larger resilient modulus than an impact-compacted specimen. The compaction water content has little effect on the resilient properties of dune sand. The resilient modulus increases linearly with dry density for relative compaction of 95–103%. A regression equation and a design chart to estimate the resilient modulus were developed.

As a part of the construction of Interstate 70 through Glenwood Canyon, in Colorado, three new types of barriers for protection against rockfall were developed through a program of prototype testing, numerical simulation of rockfall events, and analysis of barrier response to impact by boulders. New barriers include a suspended pendulum system that attenuates the kinetic energy of falling boulders, a flexibly mounted fence of steel cables and mesh, and a free-standing mechanically stabilized earth wall. The capacity of rockfall barriers, stated as a limiting translational kinetic energy of falling boulders, ranges from 41 kJ (30,000 ft$\dot{s}$lb) for light fences to more than 1,400 kJ (1,000,000 ft$\dot{s}$lb) for mechanically stabilized earth walls. Earth walls can halt falling boulders that weigh as much as 13,700 kg (30,100 lb). The three barriers reported here are useful as options to ditches, berms, rigid fences, and rigid walls of rockfall protection for highways. This paper provides a summary of barriers developed in Colorado, a summary of associated numerical simulation and analytical tools developed specifically for the barriers, as well as a guide to more detailed information on development and testing programs for rockfall barriers.

Corrective maintenance of pavements involves identifying the existing problems and selecting the most effective remedial measure to correct the deficiencies. Regular inspection and minor repairs ensure a satisfactory level of pavement performance, and it prevents the occurrence of future major problems. Due to the unbalanced proportion of too many pavements in need of repairs and too few experienced engineers to carry out the task of condition and treatment evaluation, pavement inspections and necessary repairs are irregular and are not performed as often as needed. More often than not, methods of rehabilitation are more practically available than competent engineers. The dissemination of expert knowledge about pavement inspection and treatment evaluation by means of knowledge-based systems to a wider group of engineers (such as local maintenance personnel) is needed. This paper discusses the maintenance requirements in Ontario, Canada, and presents the development of a prototype object-oriented knowledge-based system for the inspection and repair of pavements.

The objectives of this engineering investigation were to determine the structural responses of typical transit buses under various combinations of seat-belt use and seat type as well as the changes in these responses from one bus design to the next. To this end, finite-element computer models were developed for two 7.6-m transit buses designed in 1989 and 1992. These models were each analyzed under bus deceleration with full and partial seat-belt use, with seats attached to the bus floor only, and with seats attached to the bus sidewall and floor. The study concluded the following with respect to the structural responses of these typical transit buses to bus deceleration: (1) Maximum member stresses should be lower with full versus partial seat-belt use; (2) maximum member stresses should be higher with wall-mounted versus floor-mounted seats; and (3) small to moderate changes in the number, size, and location of structural members from one bus design to the next can have a substantial impact on the maximum member stresses.

It has long been recognized that medians are an effective method of increasing vehicular safety and capacity on urban and suburban arterials. Medians can provide an additional lane for through traffic by removing left turning vehicles from the traffic stream. Medians are also generally considered to be beneficial to pedestrian safety and operations. A number of studies have been conducted on modeling the expected number of vehicle accidents, and, hence, incremental benefits resulting from the installation of different median types. These studies did not, however, determine the expected pedestrian benefits resulting from the installation of medians. This paper presents the results of a study conducted for the Federal Highway Administration. It presents the activities and results of developing accident predictive equations for both vehicle-vehicle and pedestrian-vehicle accidents due to different median types. The model was developed for urban and suburban unlimited access arterials. The results are not applicable to rural environment or to limited access arterials.

Category analysis is a method to find the average response of trips made by a particular category of analysis unit such as a person or household. Conventional analysis is done by household categories whereby the causal household socioeconomic characteristics that are supposed to influence trip generations are cross-classifeid in a two-dimensional plane, and the household trip rates are calculated. A modified approach that would take care of not only the household socioeconomic characteristics but also an individual's travel attribute characteristics has been taken up in this study. The analysis has been extended from the household plane to the individual's travel attribute plane. The anomalies pertaining to sparse or no data due to stratification of household socioeconomic variables and travel attributes such as trip length have been taken into account in this study. A probabilistic approach to model formulation has been attempted. Calibration and validation of the model have also been undertaken.

Scheduling of urban transit network can be formulated as an optimization problem of minimizing the overall transfer time (TT) of transferring passengers and initial waiting time (IWT) of the passengers waiting to board a bus/train at their point of origin. In this paper, a mathematical programming (MP) formulation of the scheduling problem at one transfer station is presented. The MP problem is large and nonlinear in terms of the decision variables, thereby making it difficult for classical programming techniques to solve the problem. We apply genetic algorithms (GAs)—search and optimization methods based on natural genetics and selection—to solve the scheduling problem. The main advantage of using GAs is that the problem can be reformulated in a manner that is computationally more efficient than the original problem. Further, the coding aspect of GAs inherently takes care of most of the constraints associated with the scheduling problem. Results from a number of test problems demonstrate that the GAs are able to find optimal schedules with a reasonable computational resource. The paper concludes by presenting a number of extensions to the present problem and discusses plausible solution techniques using GAs. The success of GAs in this paper suggests their efficacy as a solution tool for similar optimization problems arising in transportation systems.