NEWS>>news archive>> ITS Researchers at INFORMS 2005
A specialty airline for college athletes, solving the "bullwhip
effect" in supply chains, infrastructure maintenance optimization,
flight scheduling methods and understanding causes of flight delays were
among the topics addressed by ITS faculty, researchers, and students
at the 2005 annual meeting of INFORMS, the top Operations Research conference
in the world. It was held this year in San Francisco, Nov. 13-16.
They included delegations from the Institute of Transportation
Studies' National Center of Excellence for Aviation Operations
Research (NEXTOR)—a major contributor to INFORMS—as well
as ITS Faculty and researchers from other departments, including the
Department of Civil and Environmental Engineering, as well as the Industrial
and Engineering Operations Research Department. Below is a list of research
papers presented by UC Berkeley researchers, with links to their work
and contacts. They are ordered alphabetically, based on the Berkeley
faculty author.
********
En-Route Air Traffic Modeling and Strategic Flow Management using Mixed Integer Linear Programming: An Eulerian model of air traffic flow is developed for the National Airspace System and validated against real data at the level of one Center. The model relies on an integer linear dynamical system framework. This is used to assess real-time controllability of user-defined metrics (sector counts, inbound flows). Empirical evidence of fast running time using Mixed Integer Linear Programming is provided, which makes the method applicable to online strategic traffic flow management for one Center
Charles-Antoine Robelin, Graduate Student Researcher, UC Berkeley, Alexandre Bayen, Assistant Professor of Civil and Environmental Engineering, UC Berkeley, Dengfeng Sun, Graduate Student Researcher, UC Berkeley, Guoyuan Wu, Graduate Student Researcher, UC Berkeley
Presented in a session titled, "Measuring the Impact of National Air Transportation System." Entire Session (PDF) (56K)
The Bullwhip Effect in Decentralized Supply Chains: This paper examines decentralized supply chains with general inventory replenishment policies. It presents: (i) demand-independent tests to determine if the bullwhip effect arises; (ii) exact formulae for the variance of the stage-n order stream under ergodic demand; and (iii) the variance-reduction benefits of advance demand information.
Yanfeng Ouyang, Assistant Professor, University of Illinois at Urbana-Champaign, Carlos Daganzo, Professor of Civil and Environmental Engineering, UC Berkeley
Presented in a session titled, "Performance Analysis of Supply Chains Systems II" Entire Session (PDF) (60K)
Delay Normalization: An Econometric Analysis of US Airline Flight Delays: We normalize airline flight delay by estimating an econometric model of average daily delay. With this model we are able to quantify major causal factors of delays, such as flow, weather, and seasonal effects. The model is applied to analyzing the flight delays of US National Airspace System.
Chieh-Yu Hsiao,UC Berkeley Graduate Student, Mark Hansen, NEXTOR Co-Director and Associate Professor of Civil and Environmental Engineering, UC Berkeley
Presented in a session titled, "Measuring Air Transportation System Performance." Entire Session (PDF) (60K)
Scenario-based Air Traffic Flow Management: From Theory to Practice: Recent developments in solving the single airport ground holding problem use static or dynamic optimization to manage the uncertainty of how airport capacities will evolve. Scenario trees of airport arrival capacity profile provide the basis for formulating multistage recourse problems. We present methodologies for generating scenario trees from empirical data and examine the performance of scenario-based models in a real-world setting.
P. Barry Liu, UC Berkeley Graduate Student, Mark Hansen, NEXTOR Co-Director and Associate Professor of Civil and Environmental Engineering, UC Berkeley, Avijit Mukherjee, Research Associate (formerly at Berkeley and now at University of Maryland).
Presented in session titled, "OR Applications for Off-Schedule Operations (OSO) in the Airline Industry." Entire Session (PDF) (59K)
Dynamic and Collaborative Planning of Ground Delay Programs: We first present a scenario-based optimization model for assigning ground delays to flights in event of a Ground Delay Program. We then present a methodology for allowing airlines to perform flight substitutions and cancellations when slot allocations are scenario-specific. The substitutions are subject to certain non-anticipativity constraints capturing uncertainty in airport capacity. Finally, we present an optimization model that functions analogously to the CDM Compression algorithm.
Avijit Mukherjee, Research Associate (formerly at Berkeley and now at University of Maryland) and Mark Hansen, NEXTOR Co-Director and Associate Professor of Civil and Environmental Engineering, UC Berkeley
Presented in session titled, "Air Traffic Management." Entire Session (PDF) (59K)
Designing a Real-Time Intermodal System for Airline Use: In the national airspace system, most of the delay is caused by adverse weather. To address the problem of a hub airport facing capacity reduction due to adverse weather, we propose a real-time intermodal system by embedding surface transportation modes into solutions for schedule recovery problems. Passenger-related information communication and market issues involved in gaining consumer acceptance of the system are also discussed.
Yu Zhang, UC Berkeley Graduate Student, and Mark Hansen, NEXTOR Co-Director and Associate Professor of Civil and Environmental Engineering, UC Berkeley
Presented in session titled, "Air Traffic Management." Entire Session (PDF) (59K)
Econometric Analysis of Aircraft Size and Service Type: the Case of La Guardia Airport (same session and page as above): A given level of air traffic can be served with more flights on small aircraft or fewer flights on large ones. There are many factors, such as slot controls, that influence this tradeoff. This research compares LGA markets with other markets to investigate if there are systematic differences in the service type and the size of aircraft used, or in the influence of factors such as stage length and market density on aircraft size.
Chieh-Yu Hsiao, UC Berkeley Graduate Student, and Mark Hansen, NEXTOR Co-Director and Associate Professor of Civil and Environmental Engineering, UC Berkeley
Presented in session titled, "From Ground to Air: Capacity Planning" chair, Mark Hansen. Entire Session (PDF) (59K)
Finding Alternative Routes between Airports under Weather Uncertainty: The stochastic dynamic ground-holding and rerouting model has been developed to enhance strategic planning tools, Ground Delay Programs for ground holding, and National Playbook for rerouting. The model combines the ground-holding and rerouting and is formulated as a heuristic shortest path problem. In order to yield the better storm predictability from the weather forecast, the probabilistic model is used to generate an input for the stochastic dynamic ground-holding and rerouting model
Wanjira Jirajaruporn, UC Berkeley Graduate Student, and Mark Hansen, NEXTOR Co-Director and Associate Professor of Civil and Environmental Engineering, UC Berkeley.
Presented in cluster titled, "Recent Advances in Aviation." Entire Session (PDF) (59K)
The Feasibility of a Private Airline for a College Sports League: In this research we address the feasibility of a dedicated airline for a sports league. Given a game schedule, fleet and operating constraints, a routing and scheduling is determined. A generalized model is developed and results are compared to those of a case-based simulation tool. We comment on strategies for creating schedules to minimize transportation cost, and the economic feasibility of semi-private sports league operations.
Anne Goodchild, and Gautam Gupta, UC Berkeley Graduate Students, and Mark Hansen, NEXTOR Co-Director and Associate Professor of Civil and Environmental Engineering, UC Berkeley
Presented in session titled, " Scheduling Modeling in Transportation,"chair, Anne Goodchild, NEXTOR Ph.D. student. Entire Session (PDF) (59K)
Scheduling and Due-date Quotation in a Make-To-Order Supply Chain: We consider a manufacturer with a single supplier that has to quote due dates to arriving customers in a make-to-order production environment. The manufacturer is penalized for long lead times, and for missing due dates. We consider several variations of this problem, and design effective due-date quotation and scheduling rules for centralized and decentralized versions of the model.
Onur Kaya, UC Berkeley Graduate Student, Philip Kaminsky, Associate Professor of Industrial and Operations Research Department, UC Berkeley
Presented in a session titled, "Scheduling and Due Date Quotation," chair Philip Kaminsky. Entire session (PDF) (60K)
Transportation Contracts in Stochastic Production/distribution Systems: Third-party logistics contracting has now become a widely utilized practice across many industries. Motivated by this observation, we develop models to analyze the optimal operation of stochastic production/distribution systems with transportation contracts. We characterize the optimal integrated production/distribution policy structures and provide numerical illustrations.
Engin Alper, UC Berkeley Graduate Student, Hyun-soo Ahn, Ross School of Business, University of Michigan, Philip Kaminsky, Associate Professor of Industrial and Operations Research Department, UC Berkeley
Presented in a session titled, "Inventory and Transportation in Supply Chains," chair, Philip Kaminsky. Entire session (PDF) (60K)
Infrastructure Maintenance Decisions Optimization using a History-dependent Markov Decision Process: A reliability-based model of deterioration is developed and used in a facility level optimization model of bridge maintenance decisions, using a Markov chain whose state includes part of the history of deterioration and maintenance. This formulation allows the use of standard optimization techniques, while using realistic, history-dependent deterioration models. The framework to formulate a continuous deterioration model as a Markov model with limited loss of information is presented.
Charles-Antoine Robelin,UC Berkeley Graduate Student, Samer Madanat, Professor of Civil and Environmental Engineering, Director of the Institute of Transportation Studies, UC Berkeley
Presented in a session titled, "Transportation Infrastructure Management II," chair, Jasenka Rakas,UC Berkeley Researcher. Entire session (PDF) (60K)
Robust Maintenance and Rehabilitation Policies for a System of Infrastructure Facilities: This paper introduces the use of robust optimization to deal with epistemic uncertainty in network-level infrastructure management. It is found that epistemic uncertainty is significant in pavement deterioration modeling. A robust approach may reduce maintenance expenditures. The methods and results can be used to identify the cost of modeling uncertainty.
Kenneth Kuhn,UC Berkeley, Samer Madanat, Professor of Civil and Environmental Engineering, Director of the Institute of Transportation Studies, UC Berkeley
Presented in a session titled, "Transportation Infrastructure Management I." Entire session (PDF) (60K)
Optimal Sequencing of Un-staffed Facility Inspection Sites in the National Airspace System: The evaluation of the Facility Condition Index (FCI) for 9,000 structural towers and 5,000 un-staffed facilities in the National Airspace System requires proper sampling methodology. This research focuses on developing an appropriate sequence of facility inspection sites with the objective to minimize sample variance. The proposed sequencing technique is supported by data analysis, including comparisons to highlight the relative benefits of such a technique.
Juaida Norrell, Federal Aviation Administration, Gautam Gupta, University of California Berkeley, Jasenka Rakas, UC Berkeley NEXTOR Researcher.
Presented in a session titled, "Transportation Infrastructure Management II," chair, Jasenka Rakas, UC Berkeley NEXTOR Researcher. Entire Session (PDF) (60K).
Balancing Service Components of the National Airspace System: Quality, Value, and Cost: This research explores questions of balance among three components of service of the National Airspace System (NAS): quality, value and cost. We also explore how value and willingness to pay, either consciously or unconsciously, ultimately drives quality of NAS service. Service Level Agreements are discussed identifying how risk tolerance, as a new quality measure, influences cost. Deterministic and stochastic models are presented to illustrate the decision process.
Jady Handal, Federal Aviation Administration, Jasenka Rakas, UC Berkeley NEXTOR Researcher
Presented in a session titled, "Transportation Infrastructure Management II," chair, Jasenka Rakas, UC Berkeley NEXTOR Researcher. Entire Session (PDF) (60K).
Effect of Controller-pilot Data Link Communications on En-route Sector Routing Efficiency: We analyze effects of controller-pilot data link communications (CPDLC) on sector routing efficiency and hypothesize that reduced voice channel occupancy results in more timely vectoring of aircraft onto their original routes in cases where they previously have been vectored off-route to resolve conflicts. We test this hypothesis by studying the relationship between voice channel occupancy and the excess distance caused by conflict resolutions.
Jasenka Rakas, UC Berkeley NEXTOR Researcher, Michael Bennett and David Knorr, Federal Aviation Administration
Presented in a session titled, "Measuring the Impact of National Air Transportation System Changes." Entire Session (PDF) (60K)
INFORMS annual meeting home page
