WSC 2003 Final Abstracts

Tutorial on Business and Market Modeling to Aid Strategic Decision Making: System Dynamics in Perspective and Selecting Appropriate Analysis Approaches
Donna D. Mayo and Knud Erik Wichmann (PA Consulting Group)

Abstract:
System dynamics models have been used to address strategic questions in many hundreds of companies and government agencies around the world over the past 40 years, including a broad range of organizations in the transport sector. However, this technique remains less well known than other approaches among potential client organizations and within the simulation community. This paper provides a pithy tutorial on the system dynamics method and the modeling process, uses transport sector case examples to illustrate how such models have been valuable in practice, and compares key characteristics of system dynamics to discrete event simulation. We close with some guidance on factors to consider when selecting an analysis approach that is appropriate to the problem under study.

Steering Strategic Decisions at London Underground: Evaluating Management Options with System Dynamics
Donna D. Mayo and William J. Dalton (PA Consulting Group) and Martin J. Callaghan (London Underground Ltd.)

Abstract:
System dynamics simulation models provide a powerful and rigorous means of evaluating management options in dynamically complex settings such as a metro subway system. This paper explores how a system dynamics model has been used over several years to address a variety of management challenges at London Underground. Several short case illustrations are described to demonstrate how such models are used in practice to aid strategic decision making, carry out robust business planning, and communicate effectively with key stakeholders.

Simulation – A Key Tool to Accelerate and Add Confidence to Postal Network Configuration
Niels Erik Larsen (PA Consulting Group)

Abstract:
This paper discusses the reasons for why the postal industry should use discrete event simulation to improve the performance of postal networks. Simulation will accelerate and add confidence to decision making. It improves the match between strategy and the applied network structures, product portfolio, technology etc. A configurable and flexible Postal Network Planner (PNP simulation tool) is introduced for “End-to-end” analysis of the postal logistic value chain. It can establish a nation wide general view of the cause-effect relationships in a complex postal system. Several real examples from the postal industry show why simulation is a prerequisite for creation of innovative and improved solutions.

Semiconductor Supply Network Simulation
Gary W. Godding (Intel Corporation), Hessam S. Sarjoughian (Arizona State University) and Karl G. Kempf (Intel Corporation)

Abstract:
More efficient and effective control of supply networks is conservatively worth billions of dollars to the national and world economy. Developing improved control requires simulation of physical flows of materials involved and decision policies governing these flows. This paper describes our initial work on modeling each of these flows as well as simulating their integration through the synchronized interchange of data. We show the level of abstraction that is appropriate, formulate and test a representative model, and describe our findings and conclusions.

Railroad Simulator on Closed Loop
Luis Augusto Gago Franzese and Marcelo Moretti Fioroni (Paragon Tecnologia) and Rui Carlos Botter (University of São Paulo)

Abstract:
This paper presents the procedures and results of a simulation project of a railroad coal transportation system. Locomotives and wagons are modeled at the system level, as opposed to other work that inserts trains at start of the line, and removes them on the other side, just to analyze line utilization and traffic. The case was based on the Estrada de Ferro Vitória-Minas (EFVM), managed by Companhia Vale do Rio Doce, one of the biggest iron coal mining companies in the world. The project helped to find the best train size, the impact of various changes on the physical line, the influence of failures and accidents, and provided a full customized interface, with the ability to do many more experiments.

Modelling a Continuous Process with Discrete Simulation Techniques and its Application to LNG Supply Chains
Niels Stchedroff and Russell C. H. Cheng (University of Southampton)

Abstract:
This paper discusses the problem of modelling an LNG supply chain efficiently. The production, processing, transportation and consumption of LNG (Liquid National Gas) and the associated products are a topic of major interest in the energy industry. While the problem is apparently continuous, analysis suggests that this problem can be modelled using discrete, deterministic techniques. A method involving a modification of the Three Phase discrete technique was used. Analysis of the way in which the effects of an event spread leads to a method by which excessive recalculation can be avoided, yielding a model that is computationally very efficient.

A Prototype Object-Oriented Supply Chain Simulation Framework
Manuel D. Rossetti and Hin-Tat Chan (University of Arkansas)

Abstract:
In this paper, we discuss the design, development and testing of a prototype object-oriented framework for performing supply chain simulations. We define the primary objects required for supply chain simulations and design how each of these objects are related to each other to form a supply chain network. We also present how persistence is handled for instantiating supply chain network simulations from a database. Finally, we present a small example simulation to validate and illustrate the concepts.

Initializing a Distribution Supply Chain Simulation with Live Data
Malay A. Dalal (Union Pacific Railroad), Henry Bell and Mike Denzien (Simulation Dynamics, Inc.) and Michael P. Keller (Insight Network Logistics)

Abstract:
This paper describes VinLogic, a simulation of a supply chain network for the distribution of new automobiles. The model is implemented in Simulation Dynamic’s Supply Chain Builder product. One of the key features of the model is its integration with a database containing the status of all vehicle shipments-VinVision. The information in the database is used to distribute vehicles and resources through the network at model start, and the model can then be used to project forward from the current situation. The model is thus able to have a “warm start,” and does not require the warm-up period necessary in simulation models that start with the system empty and idle.

Simulation Analysis of International-Departure Passenger Flows in an Airport Terminal
Soemon Takakuwa (Nagoya University) and Tomoki Oyama (Recruit Staffing Co., Ltd.)

Abstract:
An entire airport terminal building is simulated to examine passenger flows, especially international departures. First, times needed for passengers to be processed in the terminal building are examined. It is found that the waiting time for check-in accounts for more than 80 percent of the total waiting time of passengers spent in the airport. A special-purpose data-generator is designed and developed to create experimental data for executing a simulation. It is found that the possible number of passengers missing their flights could be drastically reduced by adding supporting staff to and by making use of first- and business-class check-in counters for processing economy- and group-class passengers.

Online Simulation of Pedestrian Flow in Public Buildings
André Hanisch, Juri Tolujew, and Klaus Richter (Fraunhofer Inst. for Factory Operation & Automation) and Thomas Schulze (University of Magdeburg)

Abstract:
Online simulation of pedestrian flow in public buildings is a new tool which can be especially useful for improving the aspects of safety and short-term planning in the phase of organizing and operating large public buildings. These might be places such as a train station, an airport or a shopping center. This paper provides an insight into the different concepts of pedestrian flow simulation. Special emphasis is placed on explaining the mesoscopic approach as applied to the area of traffic simulation. This approach is transferred to the context of analyzing and predicting the pedestrian flow. A first prototypical implementation of a simulation supported control center is briefly presented, also.

Development and Validation of a Flexible, Open Architecture, Transportation Simulation
Michael Hunter (Georgia Institute of Technology) and Randy Machemehl (The University of Texas at Austin)

Abstract:
Simulation has been utilized in the planning and development of almost all sectors of the transportation field. The practicing transportation community primarily relies on simulation packages, as opposed to “ground up” simulation development. Unfortunately, the use of these simulation packages has several disadvantages, most notably the “black box” phenomenon and reduced modeling flexibility. The simulation approach described in this paper lays the foundation for a transportation simulation approach that minimizes the “black box” problem and increases modeling flexibility, while still providing an easy to use package in which highly capable models may be quickly and accurately built. This simulation approach utilizes SIMAN and ARENA. This paper includes a brief discussion of the simulation approach, a comparison of the proposed simulation and CORSIM results for an intersection and an arterial, and a comparison of the proposed simulation control delay to delays collected for a twelve intersection grid north of downtown Chicago.

Total Airport and Airspace Model (TAAM) Parallelization Combining Sequential and Parallel Algorithms for Performance Enhancement
Neera Sood and Frederick Wieland (The MITRE Corporation)

Abstract:
This paper describes how to achieve a desired speedup by careful selection of appropriate algorithms for parallelization. Our target simulation is the Total Airport and Airspace Model (TAAM), a worldwide standard for aviation analysis. TAAM is designed as a sequential program, and we have increased its speed by incorporating multi-threaded algorithms with minimal changes to the underlying simulation architecture. Our method was to identify algorithms that are bottlenecks in the computation and that can be executed concurrently, producing a hybrid sequential and parallel simulation. Our results show a performance gain that varied between 14% and 33%.

Discrete Event Simulation Model for Airline Operations: SIMAIR
Loo Hay Lee, Huei Chuen Huang, Chulung Lee, Ek Peng Chew, Wikrom Jaruphongsa, Yean Yik Yong, Zhe Liang, Chun How Leong, Yen Ping Tan, and Kalyan Namburi (National University of Singapore) and Ellis Johnson and Jerry Banks (Georgia Institute of Technology)

Abstract:
SIMAIR is a C++ based research tool meant for the simulation of airline operations. Various aspects of airline operations like aircraft, crew and passengers might be simulated to keep the simulation as close to reality as possible. It provides a means for devising and evaluating various airline recovery mechanisms to handle disruptions, and can also be used as a tool to evaluate the performance of a given schedule in operations. The performance of a given recovery mechanism or a given schedule can be quantified for research and evaluation purposes.

Simulating Aircraft Delay Absorption
Justin Boesel (The MITRE Corporation)

Abstract:
An airplane's ability to absorb delay while airborne is limited and costly. Because of this, the air traffic control system anticipates and manages excessive demand for scarce shared resources, such as arrival runways or busy airspace, so that the delay necessary for buffering can be spread out over a larger distance, or taken on the ground before departure. It is difficult to model these important dynamics in a standard queue-resource simulation framework, which does not account for limited delay absorption capacity. The modeling methodology presented here captures these dynamics by employing a large number of independent threads of execution to monitor and enforce a large number of relatively simple mathematical relationships. These relationships calculate feasible time windows for each portion of each flight. The model was implemented in the SLX simulation language. The speed and scalability of SLX are essential to the approach, which would otherwise be impractical.

Runway Schedule Determination by Simulation Optimization
Thomas Curtis Holden and Frederick Wieland (The MITRE Corporation)

Abstract:
Many airport runway expansion projects are restricted by space limitations imposed by development in the vicinity of the airport. This often causes planners to choose configurations for new runways that limit the use of these runways in time and/or space. Studies that model airports with new runways that are not yet operational need to develop plausible operational models for these new runways since historical data is not available. We look at a runway schedule problem encountered during the configuration and validation step of an earlier study. We develop a method using simulation optimization to approach the runway schedule problem and compare it to a manual approach developed in the earlier study. We use the Total Airspace and Airport Modeler to model the airport and airspace operations and Fast Simulated Annealing for the optimization.

A Simulation Approach to Manpower Planning
Massoud Bazargan-Lari, Payal Gupta, and Seth Young (Embry-Riddle Aeronautical University)

Abstract:
This paper describes the development of a computerized line maintenance simulation model for strategic manpower planning at Continental Airlines for one of their major maintenance stations at Newark airport. The simulation model provides guidelines to the development of enhanced staffing models and a better understanding of resource requirements on a daily basis. The proposed simulation model could be used as a tool to support the management of the line maintenance department in solving various capacity planning issues related to the manpower requirement and scheduling. The recent capabilities of simulation modeling, namely optimization modeling is adopted in search of enhanced shift schedule of technicians that would improve the efficiency of the existing system.

A Discrete Event Simulation for the Crew Assignment Process in North American Freight Railroads
Rainer Guttkuhn, Todd Dawson, and Udo Trutschel (Circadian Technologies, Inc.) and Jon Walker and Mike Moroz (Canadian National Railroad)

Abstract:
This paper introduces a discrete event simulation for crew assignments and crew movements as a result of train traffic, labor rules, government regulations and optional crew schedules. The software is part of a schedule development system, FRCOS (Freight Rail Optimization System), that was co-developed by CN Rail and Circadian Technologies, Inc. The simulation allows verification of the impact of changes to trainflow, labor rules or government regulations on the overall operational efficiency of how crews are called to work. The system helps to evaluate changes to current crew assignments and can test new crew assignment scenarios such as crew schedules. Potential problems can be detected before the actual implementation, saving unnecessary costs. The software is also used to assess the impact of traffic changes on existing crew schedules in order to implement reactive corrections to these schedules.

Simulation of Freeway Merging and Diverging Behavior
Daiheng Ni (Georgia Institute of Technology) and John D. Leonard, II (Georgia State Road & Tollway Authority)

Abstract:
Simplified theory of kinematic waves was proposed by Newell and uses cumulative arrival and departure counts to describe kinematic waves of freeway traffic. The original paper deals only with traffic on freeway mainline. It is of great interest, at least practically, to investigate whether the simplified theory can be used to simulate freeway traffic merging and diverging behavior. In his paper, Newell assumed that on-ramp traffic always has the priority and can bypass queues, if any. This assumption will be released so that traffic from the mainline and the on-ramp will have to compete for downstream supply. For off-ramps, Newell assumed that all vehicles that want to exit can always be able to do so. Again, this assumption is also released so that queues from either downstream can build up and block upstream traffic.

Modeling Ambulance Service of the Austrian Red Cross
Othmar Koch (Vienna University of Technology) and Heinz Weigl (ESLA Solutions)

Abstract:
We discuss a simulation model used in the analysis of the transport logistics of the Austrian Red Cross rescue organization. The emphasis is on the details of modeling the scheduling of ambulance service in the simulation environment ARENA. A heuristic (near-) optimal strategy is employed to coordinate patients' transports, where some parameters with an intuitive interpretation, which are involved in the decision process, have to be suitably chosen. The validity of the model is apparent from the interpretation of the results in terms of the structure of the organization and coordination of services provided.

Simulation Modelling in Support of Emergency Fire-Fighting in Norfolk
Alan C. Cowdale (Royal Air Force Waddington)

Abstract:
In the fall of 2002 the Fire Brigade Union within the United Kingdom proposed a ballot on strike action in support of a wage demand. Consequently the Ministry of Defence agreed that in the event of strike action, emergency assistance to the local authorities would be provided by deploying military staff to man fire-fighting equipment and to provide command and control functions. The military Detachment Commander for the County of Norfolk region was faced with a number of issues regarding basing of assets and asset management. This paper describes how simple simulation modelling was used to provide insights into the type of issues that the military would face in undertaking this role. Two models were developed; one using SIMUL8 to investigate asset utilisation, and the other using VisualBasic to generate a master event list for use in mission planning and control centre training.

Dynamic Freight Traffic Simulation Providing Real-Time Information
Jinghua Xu and Kathleen L. Hancock (University of Massachusetts) and Frank Southworth (Oak Ridge National Laboratory)

Abstract:
The paper describes a prototype Dynamic Freight Traffic Simulation model called DyFTS, constructed for studying the effects of highly developed information technologies and logistic strategies on freight transportation. DyFTS is designed as a highly adaptable system that can be easily embedded into a more comprehensive transportation simulation model. Various decision-making processes are formulated, such as goods-to-vehicle assignment, departure time choice and pre-trip routing, and en-route vehicle redirection. As part of the modeling system, descriptive real-time information for the network is simulated to study the influence of such information on freight transportation. A knowledge-based learning process is established to refine the perceptions of decision-makers to the transportation network based on past experience. Numerical examples are designed to compare a set of freight movements operating both with and without the aid of real-time information, as such freight operations vary according to different delivery time requirements and using different fleet configurations.

The Modal-Shift Transportation Planning Problem and its Fast Steepest Descent Algorithm
Masami Amano, Takayuki Yoshizumi, and Hiroyuki Okano (IBM Research)

Abstract:
The Modal-Shift Transportation Planning Problem (MSTPP) is the problem that finds a feasible schedule for carriers with the minimum total cost when sets of facilities, delivery orders, and carriers are given. In this paper, we propose a fast steepest descent algorithm to solve the MSTPP. Our solution generates a set of candidate routes for each delivery order as a preprocess. Then, it finds a schedule by iteratively updating selections of the candidate routes in descent directions, while computing a configuration of carrier movements at each iteration by a greedy algorithm. Intensive numerical study using artificial data modeled from the manufacturing industry in Japan is also presented.

A Monte Carlo Simulation Approach to the Capacitated Multi-Location Transshipment Problem
Deniz Özdemir and Enver Yücesan (INSEAD) and Yale T. Herer (Israel Institute of Technology)

Abstract:
We consider a supply chain, which consists of N retailers and one supplier. The retailers may be coordinated through replenishment strategies and lateral transshipments, that is, movement of a product among the locations at the same echelon level. Transshipment quantities may be limited, however, due to the physical constraints of the transportation media or due to the reluctance of retailers to completely pool their stock with other retailers. We introduce a stochastic approximation algorithm to compute the order-up-to quantities using a sample-path-based optimization procedure. Given an order-up-to S policy, we determine an optimal transshipment policy, using an LP/Network flow framework. Such a numerical approach allows us to study systems with arbitrary complexity.

Modeling Ship Arrivals in Ports
Eelco van Asperen, Rommert Dekker, Mark Polman, and Henk de Swaan Arons (Erasmus University Rotterdam)

Abstract:
Ports provide jetty facilities for ships to load and unload their cargo. Since ship delays are costly, terminal operators attempt to minimize their number and duration. Here, simulation has proved to be a very suitable tool. However, in port simulation models, the impact of the arrival process of ships on the model outcomes tends to be underestimated. This article considers three arrival processes: stock-controlled, equidistant per ship type, and Poisson. We assess how their deployment in a port simulation model, based on data from a real case study, affects the efficiency of the loading and unloading process. Poisson, which is the chosen arrival process in many client-oriented simulations, actually performs worst in terms of both ship delays and required storage capacity. Stock-controlled arrivals perform best with regard to ship delays and required storage capacity.

Optimization of a Barge Transportation System for Petroleum Delivery
Nicholas P. Anderson, Gerald W. Evans, and William E. Biles (University of Louisville) and Todd C. Whyte (American Commercial Barge Lines)

Abstract:
This paper describes a simulation model of a liquid fuel supplier operating on the Ohio River. Each day, orders arrive for six different fuel types at six different locations. The goal of this study was to determine the appropriate number of tow boats required to meet the demand for fuels. The system was analyzed using Arena and Optquest.

Iterative Optimization and Simulation of Barge Traffic on an Inland Waterway
Amy Bush, W. E. Biles, and G. W. DePuy (University of Louisville)

Abstract:
This paper describes an iterative technique between optimization and simulation models used to determine solutions to optimization problems and ensure that the solutions are feasible for real world operations(in terms of a simulation model). The technique allows for the development of separate optimization and simulation models with varying levels of detail in each model. The results and parameters of the optimization model are used as input to the simulation model. The performance measures from the simulation output are compared to acceptable levels. These performance measures are then used to modify the optimization model if the simulation results are not acceptable. This iterative approach continues until an acceptable solution is reached. This iterative technique is applied to barge traffic on an inland waterway as an example. Linear programming is used as the optimization technique for the example while a simulation model is developed using Arena software.