Abstract

The AMBR Model Comparison Project: Round III — Modeling Category Learning Session Organizers: Kevin A. Gluck (kevin.gluck@williams.af.mil) Air Force Research Laboratory 6030 S. Kent St., Mesa, AZ 85212 USA Richard W. Pew (pew@bbn.com) BBN Technologies 10 Moulton St., Cambridge, MA 02138 USA The goal of the Agent-based Modeling and Behavior Representation (AMBR) Model Comparison Project is to advance the state of the art in cognitive modeling. It is organized as a series of model comparisons, moderated by a team from BBN Technologies. In each comparison, a challenging behavioral phenomenon is chosen for study. Data are collected from humans performing the task. Cognitive models representing different modeling architectures are created, run on the task, and then compared to the collected data. The current effort focuses on models of category learning in a dynamic, dual-task environment. Model comparisons such as this, especially with directly comparable human data are rare. While models of category learning are commonplace, the fact that these are models of integrative performance, not just models of category learning in isolation, makes this set of presentations unique. Experiment Design and Comparison of Human and Model Data David Diller (ddiller@bbn.com) Yvette Tenney (ytenney@bbn.com) BBN Technologies This experiment involved a classic concept learning task embedded in an air traffic control situation. Subjects had to learn to make correct decisions to accept or reject altitude change requests, based on three bi-variate properties of the aircraft (percent fuel remaining, aircraft size, and turbulence level). A novel feature of the experiment was the addition of multi-tasking to this concept learning paradigm. In addition to the altitude change requests (the concept learning task), the participant had to hand-off a number of aircraft to adjoining controllers (secondary task). The design consisted of 9 conditions, defined by 3 category structures and 3 workload levels. The three category structures, borrowed from Shepard, Hovland, and Jenkins (1961), were: single attribute relevant (Type I), a single-attribute rule plus exceptions (Type III), and no rule (Type VI). The three workload levels consisted of 0, 12, or 16 required handoffs, in addition to the 16 altitude requests. It was expected that both category structure and workload level would affect performance. There were 8 scenarios, or trials, lasting ten minutes each. One hour of training on the mechanics of the tasks preceded the trials. Ninety humans and four different human performance models described in subsequent abstracts were run through the scenarios. The interface, consisting of a radar screen with moving aircraft and action buttons, was designed to accommodate both humans and models. Humans were randomly assigned to one condition (ten per condition). The models were run one or more times in each condition. All of the modelers were given the human learning data as soon as they were collected, and while the models were still under development. It was expected, therefore, that they would fit the data fairly well. However, a transfer test (for which the modelers were not given the human data in advance) provides an opportunity to test the generalizability of the models predictions. Results for both humans and models will be presented on the effects of category structure and workload over trials. Human data and model data are available for the following measures: learning curves (probability of error) on the concept learning task, performance errors on the secondary task (missed and incorrect actions), reaction time on both the concept learning and secondary task, self rated workload ratings (collected from models too!), and self- reports on rule discovery and other strategies on the concept task (humans only). This presentation will set the stage for the modelers to describe the mechanisms and assumptions that allow their models to replicate the results. An EPIC-Soar Model of Concurrent Performance on a Category Learning and a Simplified ATC Task Ron S. Chong (rchong@gmu.edu) George Mason University Robert E. Wray (wray@soartech.com) Soar Technology, Inc. During the first phase of the AMBR project, we developed a model of a simplified en-route air traffic control task. That model was built using the EPIC-Soar architecture, an integration of the perceptual and motor systems of the EPIC architecture with Soar, a learning cognitive architecture. The task to be modeled for the current phase of AMBR is the combination of the same ATC task with a new concept acquisition task. Our approach to building the new model has been to reuse, in a modular fashion, previous Soar models for the subtasks. The ATC model is essentially the same as that of the previous AMBR phases. To produce the learning behavior, we have incorporated an existing process model of concept learning called SCA (symbolic concept acquisition). SCA was developed in Soar and has been

Abstract

Free Parking or Free Markets DONALD SHOUP It is no doubt ironic that the motorcar, superstar of the capitalist system, expects to live rent-free. W OLFGANG Z UCKERMAN C ITIES SHOULD CHARGE THE RIGHT PRICES FOR CURB PARKING because the wrong prices produce such bad results. Where curb parking is underpriced and overcrowded, a surprising share of cars on congested streets can be searching for a place to park. Sixteen studies conducted between 1927 and 2001 found that, on average, 30 percent of the cars in congested downtown traffic were cruising for parking. More recently, when researchers interviewed drivers stopped at traffic signals in New York City in 2006 and 2007, they found that 28 percent of the drivers on a street in Manhattan and 45 percent on a street in Brooklyn were cruising for curb parking. In another study in 2008, the average time it took to find a curb space in a 15 block area of the Upper West Side of Manhattan was 3.1 minutes and the average cruising distance was 0.37 miles. For each individual driver, 3.1 minutes is not a long time, and 0.37 miles is not a long distance, but because there are so many drivers, the cumulative consequences are staggering. In a year, cruising for underpriced parking on these 15 blocks alone creates about 366,000 excess vehicle miles of travel (equal to 14 trips around the earth) and 325 tons of CO 2 . Donald Shoup is Professor of Urban Planning in the Luskin School of Public A ffairs at the University of California, L os Angeles (shoup@ucla.edu). A C C E S S

Abstract

Ninety-five percent of automobile commuters in the United States park free at work. To deal with the traffic congestion and air pollution caused by parking subsidies, California law now requires many employers to offer employees the option to cash out their parking subsidies. Similar Federal legislation has been proposed. This nationwide survey found that employers in the United States off employees 84.8 million free parking spaces. Employers own 65.3 million of these free parking spaces, and rent the other 19.5 million. Employers of fewer than twenty employees provide more than half of all employer-paid parking spaces.

Abstract

The political feasibility of using prices to mitigate congestion depends on who receives the toll revenue. We argue that congestion pricing on freeways will have the greatest chance of political success if the revenue is distributed to cities, and particularly to cities through which the freeways pass. In contrast to a number of previous proposals, we argue that cities are stronger claimants for the revenue than either individual drivers or regional authorities. We draw on theory from behavioral economics and political science to explain our proposal, and illustrate it with data from several metropolitan areas. In Los Angeles, where potential congestion toll revenues are estimated to be almost $5 billion a year, distributing toll revenues to cities with freeways could be politically effective and highly progressive.

Abstract

Traffic congestion is a disruptive fact of urban life. It inflicts delays and frustrations in virtually all major cities in the world, new or old, rich or poor. Although many countermeasures have been tried, it is hard to name a city in which there is much satisfaction with the existing state of affairs. About the only positive aspect of congestion is that it reflects the pulse of life, a demand for travel and trade that typically accompanies economic activity.

Abstract

Little is known about demand for EVs in the mass market. In this paper, we exploit a natural experiment that provides variation in large EV subsidies targeted at low- and middle-income households in California. Using transaction-level data, we estimate two important policy parameters using triple differences: the subsidy elasticity of demand for EVs and the rate of subsidy pass-through. Estimates show that demand for EVs amongst low- and middle-income households is price-elastic and pass-through is complete. We use these estimates to calculate the expected subsidy bill required for California to reach its goal of 1.5 million EVs by 2025.

Abstract

In this thesis we examine how fuel price variation affects the optimal mix of services in intercity transportation. Towards this end, we make two main contributions. The first is the development of an analytic total logistics cost model of intercity transportation, which is sensitive to fuel price and incorporates multiple classes of vehicles serving passengers with differentiated values of time. The second is an empirical investigation of the cost relationship between fuel price and operating cost for intercity transportation vehicles. The analytic total logistics cost models are combined with the empirical models to gain insights into the impact of fuel price on optimal service mixes in representative corridors. We consider a scheduled intercity transportation corridor on which different classes of intercity transportation vehicles serve passengers with differentiated values of time. In determining optimal service mix, we consider a central planner choosing the vehicles and service frequencies that provide the minimum total logistics cost for an intercity transportation corridor. The total logistics cost is the sum of the two main intercity transportation cost components: vehicle operator cost and passenger cost. In considering operating and passenger costs together, we balance cost efficiency and level of service of alternative vehicles with different cost structures and service attributes. In developing the total logistics cost model, we seek both analytic insights and numerical examples. To keep the model analytically tractable while at the same time incorporating multiple objectives, including fuel cost, operating cost, schedule delay, and line-haul time, we incorporate the continuum approximation method from logistics. In employing the continuum approximation, discrete variables are considered continuous, leading to analytic functions from which we can evaluate qualitatively the relationships among fuel price, service level, and comparative vehicle cost. An investigation of the analytic model suggests that, while a fuel price increase would increase costs for any corridor, the rate of cost increase for a corridor served by a mix of vehicle technologies diminishes more rapidly with fuel price. We also find that an increase in fuel price causes vehicles to become more differentiated with respect to the value of time of the passengers they serve. In other words, under high fuel prices the total logistics cost can be minimized by effectively segregating passengers on different types of vehicles according to their values of time. We complement the analytic findings with an empirical investigation of the cost relationship between fuel price and operating cost for different classes of intercity transportation vehicles. We perform this analysis for a subset of intercity transportation vehicles for which data is readily available: jet and turboprop aircraft. In developing a translog operating cost model for jet aircraft, we estimate a flexible functional form that provides a detailed representation of the empirical relationship between fuel cost and operating cost, allowing for substitution, scale, aircraft age, and other effects – including interactions – to be captured. The function reveals that as fuel price increases, airlines will take steps to use fuel more efficiently by leveraging other inputs; however, the potential for this supplier input substitution for fuel is rather modest. This finding reinforces the formulation of the analytic total logistics cost model, in which the only actions available to a central planner to reduce costs are changing technologies and service frequencies. It also proves that empirical models with simpler functional forms are able to accurately predict operating costs, despite the lack of variable interactions. Using linear empirical operating cost models, we estimate operating cost and total logistics costs for intercity transportation corridors served by single vehicle fleets of three different aircraft classes. We find that a specific turboprop aircraft model, with a relatively low fuel consumption rate, provides intercity transportation service with the minimum operating cost compared with a jet with smaller seating capacity over all fuel prices considered and medium-capacity jets for some fuel prices. However, this is no longer the case when total logistics cost is considered, due to the lower quality of passenger service turboprops provide. At a given intercity transportation corridor distance, the fuel price for which the total logistics cost per passenger is equal across turboprops and low-capacity jets is in the fuel price range experienced from 2004 and expected through 2020. For this fuel price range, the total logistics cost per passenger for the medium-capacity jet is generally lower than the turboprop and always lower the lowcapacity jet. This suggests that a mix of services between intercity transportation vehicles could minimize cost for this range of fuel price. To investigate the possibility of mixing services to reduce costs further, we combine the analytic total logistics cost model with the empirical models. In addition to a jet and turboprop aircraft model, we build a high speed rail cost model and consider high speed rail as an additional intercity transportation technology. We find the minimum cost vehicle combination to be sensitive to fuel price in a small transition zone within which the cost ordering of vehicle combinations changes significantly, whereas outside this zone the orderings are stable. As the transition area is in the range of fuel prices forecasted between the years 2010-2035, the results indicate fuel price changes between 2010 and 2035 may dramatically alter the most cost-effective ways to provide intercity passenger transport. We find that high speed rail is a part of a mixed vehicle service that can reduce total logistics cost, suggesting that an integrated air and rail strategy could be an effective tool to manage costs and fuel consumption for an intercity transportation system.

Abstract

Through an extended numerical example, this paper develops a diagrammatic analysis of steady-state parking and traffic congestion in an isotropic downtown and provides systematic policy analysis. Unlike our previous work, the model incorporates curbside parking, garage parking, and price-sensitive travel demand in a unified setting. We examine the deadweight loss associated with underpriced curbside parking, as well as first- and second-best curbside parking capacities. We also explore the transient dynamics and stability of various downtown traffic equilibria. Keywords: cruising for parking, curbside parking, parking garages, parking policy, traffic congestion JEL Classification: D04; L91; R41; R48