|
The 2002 INFORMS National Meeting was held in San Jose,
California, November 17-20, 2002. For more information, contact the
cluster chair:
Natasha Lukiantseva Click here for information about the full INFORMS conference. |
Setting the Table: Supply Chain Applications in the Rail Industry
Rail-centric Supply Chain Services: Selling to our customers and ourselves
Use of optimization in rail-centric supply chain consulting and services
CP/Tronicus experience with rail-provided supply chain services
Intelligent supply chain control
A prototype for a successful implementation of rail-based supply chain
services
RASIG Business Meeting - Sunday 5:45 - 6:45
Trent Sommers, Norfolk Southern Railroad, Network Optimization, Atlanta, GA.
Paper: Industry Changing Gateway Optimization (CN/CXT, BN/CSXT)
Natasha Lukiantseva, CSX Transportation, Operations Research, Jacksonville, FL.
Gateway Optimization between Class One railroads is bringing the benefits of a merger without a merger, allowing railroads to improve their service to the customer on the interline traffic. cThe presentation will give an overview of several projects that were undertaken this and last year.
Scheduled Railroads
Carl Malazdrewich, Canadian National, Operations Research.
Christopher D Neuman, Karen Smilowitz, Northwestern University, Department of IE/MS, Evanston, IL
In the Chicago area, an estimated 25,000 containers are moved daily between customers and railheads. Each movement generates empty movements that do not generate revenue. We develop a mathematical programming formulation to coordinate all movements that yields solutions for small problems but scales poorly. We also introduce a scalable heuristic.
Modeling Truck-Rail Interaction at Intermodal Yards
Bhuwan Bhaskar Agrawal, Peng Duan, Athanasios Ziliaskopoulos, Northwestern University, Evanston, IL
This presentation deals with modeling issues that lie in the interface between modes of intermodal freight operations, as well as interactions among stakeholders. An important interaction is between truck and train, which happens in intermodal yards. The efficiency of yard operations is critical for the whole industry. This not only affects the commercial vehicle operators and the railroads, but also most of the other stakeholders in the freight industry. We discuss an effort to build a yard simulation model that will not only help in understanding the operations in the yard, but also in testing various scenarios and impacts of possible changes/improvements in the yard operations and commercial vehicle operator reactions. Emphasis will be given to data issues, modeling tools, validation and finally issues related to evaluating the impact of changes in the yard operations to improve the efficiency.Rail car distribution under uncertainty using the separable, projective approximation routine (SPAR)
Huseyin Topaloglu, Warren B. Powell, Sanjay Melkote, Selcuk Avci. Department of Operations Research and Financial Engineering, Princeton University.
We have developed a method for allocating rail cars which takes forecasting uncertainties into account. The method, which is an adaptation of the SPAR algorithm for two-stage stochastic resource allocation problems, handles multiple car types, provides integer solutions and easily scales to very large scale problems. Numerical results are reported.
Erick Wikum, Ted Gifford, Jeff Day, Schneider National, Inc., Green Bay, Wisconsin
We consider how to allocate empty repositioning costs among the loads and lanes of a transportation network, a problem that can arise in pricing, profitability analysis, and operations. We examine a variety of approaches based on standard accounting techniques, linear programming duality, and game theory.
Kalman Silverberg, Burlington Northern Santa Fe Railroad.
The Burlington Northern Railroad attempted a redesign of it's service design process in the early 1990s. The team developed approaches to the service design question, which are unique to the railroad industry. The presenter will examine those approaches and their application to the operations research arena.
Charles Martin, CSX Transportation.
The presentation will discuss the next generation dispatching models. It outlines the background of the uses, limitations and future direction. The next generation dispatching models have evolved through the advancements in computer technology. With the improved capabilities of these tools come more complex questions, resulting in more increased demands on the user. In this presentation, we will utilize examples from two new dispatching models, RTC and RAILS 2000, to illustrate the process changes in simulations.
Eric Wilson, Berkeley Simulations.
The speaker will provide his observations about dispatching model features most demanded by the user community and the continuing challenges faced in creating and enhancing dispatching simulators. His model is widely used in the U. S., and is key to the Chicago Terminal improvement project currently underway.
Jian Liu, University of Florida, Gaiinsville, FL
The locomotive scheduling problem is to assign locomotives with sufficient power to trains on a pre-planned train schedule. In this paper, we report the results of a study of the locomotive scheduling problem faced by CSX Transportation. We developed a solution technique that uses ideas from problem decomposition, linear programming, integer programming, and neighborhood search. Comparing our solution with that of in-house software developed by CSX showed that we obtained a savings of over 400 locomotives, which translates into a savings of over one hundred million dollars annually.
Jodelson Sabino, Federal University of Espirito Santo, Vitoria ES, Brazil
This paper explores the application of a multiple ant colony systems meta-heuristic method called COMPETants in railroad yard operations optimization. Some modifications are proposed to make the original algorithm more suitable for use in the railroad yard context. The problem proposed here can be classified as a Full Truckload Capacitated Pickup and Delivery Problem with Time Windows and Capacity Constraints (PTPTWC) and we aim to attend a set of transportation requests with minimal cost.
Richard Freling, Dennis Huisman, Leo G. Kroon, Ramon M. Lentink, Erasmus University Rotterdam, The Netherlands.
In this paper we introduce the problem of shunting passenger train units in a railway station. Shunting occurs whenever train units are temporarily not necessary to operate a given timetable. We discuss several aspects of this problem and focus on two subproblems. We propose mathematical models for these subproblems together with a solution method based on column generation. Furthermore, a new efficient and speedy solution technique for pricing problems in column generation algorithms is introduced. Finally, we present computational results based on real life instances from Netherlands Railways.
Blocking problem is one of the most important problems in railroad scheduling and is a very large-scale multi-commodity flow network design problem containing millions of fixed charge variables. We propose new approaches of solving this problem based on neighborhood search techniques, which are able to solve the problem very efficiently in practice. We expect major savings by the use of these approaches.
Erika Yazid and David Friedman, MultiModal Applied Systems, Princeton, NJ.
The FirstClass project between CSX and MultiModal (called Algorithmic Class Tracking at CSX) involves two major algorithmic technologies: (1) the ability to rapidly sequence blocks, and (2) the ability to generate detailed classification tables. This talk focuses on the block sequencing, and explains the blocking network with its various artificial arcs, the shortest-path technology used, and our experiences with parallelization of the algorithm.
Session Chair: Chip Kraft, Reebie Associated, CKraft@reebie.com .
This tutorial will provide a high-level overview of the state-of-the-practice versus state-of-the-art in railroad-related operations research. Students and other researchers who think they might have an interest in rail OR topics, are encouraged to attend. Sample applications include train dispatching, rail yard operations, service planning and network design, car scheduling, service monitoring and service recovery, revenue management, empty car distribution, locomotive distribution and crew management.
Return to RASIG Home Page
Updated 2 June 2003