Development of Criterion Models for Use in
Simulation Analysis
Gerald W. Evans (University of Louisville) and
Nicholas P. Anderson (Decisive Analytics Corporation)
Abstract:
Criterion models are important components in analyses
of simulation output. These models represent the basis by which decision
makers can rank alternatives in a simulation study. This work gives an
overview of criterion models, especially as related to their uses with
simulation models. In particular, examples of criterion models used in
simulation studies are given. In addition, methodologies for choosing a set of
attributes for a criterion model are described, along with a discussion of how
these attributes should be “combined” in order to form a criterion model.
Finding Probably Best Systems Quickly Via
Simulation
Takayuki Osogami (IBM Tokyo Research Laboratory)
Abstract:
The performance of computer systems can in theory be
optimized by selecting the system configuration having the best simulated
performance. Unfortunately, it is often computationally intractable to
estimate the performance of all system configurations accurately via
simulation. We propose an algorithm for finding the best system configuration
without estimating the performance of each system configuration accurately, so
that the total simulation time is minimized. We prove that with high
probability our algorithm finds the best system configuration. In addition,
our algorithm has an advantage in that it only infrequently changes the system
configurations to be simulated. Numerical experiments show that our algorithm
improves upon existing algorithms with respect to the total simulation time
and the frequency of changing the system configurations.
Epidemics Alert & Response Framework and
Technology based on Spreading Dynamics Simulation
Chunhua Tian, Xun
Zhang, Wei Ding, and Rongzeng Cao (IBM China Research Lab)
Abstract:
There have been many research works on epidemic
spreading using athematical or simulation methods. However, to make it
practical as a decision support tool for the policy maker, it needs to be
combined with business operation of epidemic control. Thus, an alert and
response framework covering outbreak detection, policy analysis, and policy
execution is proposed in this work. Key technologies for such a framework at
the modeling, analysis, and data collection aspects are discussed. Sensitivity
analysis is proposed for model verification and confidence analysis. Critical
point identification is another important topic for control policy making by
network topology analysis. The importance and feasibility of these
technologies are illustrated by the experiment on Beijing SARS (Severe Acute
Respiratory Syndrome) data in 2003.
Volume Fusion: A Unified Frame Work for Multi-Scale
Dynamic and Heterogeneous Scenes
Duoduo Liao (George Washington
University) and Shiaofen Fang (Indiana University/Purdue University
Indianapolis)
Abstract:
This work presents a novel unified solution -- a volume
fusion framework for 3D modeling, interaction, and visualization applications
with complex volumetric scenes composed of any type of conventional and
unconventional object representations. The process of volume fusion includes
scene expression and scene evaluation. A volume scene evaluation based on
slice sweeping approach is employed to provide a general computational
framework that extracts volumetric information of the scene expression and
then fuse them by specified fusion functions. The key issues of volume fusion
are scene tree design and fusion function design. Through the unified
framework, different tree structures and fusion functions may result in
different outputs (i.e., applications) after scene evaluation. Two case
studies, volume modeling and volumetric collision detection, using volume
fusion framework will be discussed in detail, and the experimental results
demonstrate the effectiveness, flexibility and generality of our volume fusion
algorithms for 3D interactive volume applications.
Modeling of Individual Decision Making in Mass
Customization Markets
Zbigniew J Pasek (University of Windsor)
Abstract:
The creation of market simulation models is
increasingly widespread as it enables better enterprise planning and
optimization of product delivery. In case of mass customization approach,
which aims to simultaneously target scores of individual customers,
understanding customer behavior under growing market stratification conditions
is critical. This work describes development of a market simulator in which
individual customers are represented as software agents, behavior of which is
based on a set of predefined rules. Parameters in those rules can be either
fixed or randomly follow some probabilistic distributions. Customers/agents
operate in a market which offers them a line of products. These products have
a set of customizable features, presumably addressing the needs of customer
population. Our work describes in detail the decision process of an agent,
market fragmentation design, and presents some sample results which can be
used to determine desired variation offering of a product.
Quantitative Studies on the Organizational Structure
of SMEs: A Case Study of Pharmaceutical Industry in China
Linlin
Cui (Tsinghua University), Jin Dong (IBM China Research Lab ) and Yueting Chai
(Tsinghua University)
Abstract:
Organizational reformation is inevitable, during the
pursuing of larger scales of Small and Medium Enterprises (SMEs). In this
work, a case study of the pharmaceutical industry in China is presented to
show how process-based simulation can be applied to the quantitative studies
on the organizational structure reformation. Through modeling core business
processes and limited resources in the sales and marketing division of the
enterprise, a simulation experiment is performed to analyze how the employees'
working hours distribute in daily work. It is found that salesmen only spend
40% working time on the effort of marketing expansion and sales increase. It
takes more than half the working time for them to track the progress of
logistics affairs. So, it may be concluded that the company in the case should
improve the specialization of the organizational design, and especially should
consider outsourcing the professional logistics service.
Information Fusion and Component Development
for Space Launch Risk Analysis
Serge N. Sala-Diakanda and Luis
Rabelo (University of Central Florida)
Abstract:
We discussed previously the need for a risk analysis
tool which could account simultaneously for the variety of hazards produced by
the explosion of a space launch vehicle. We particularly argued that by
analyzing separately the risks posed by the pressure of the explosion blast,
the dispersion of toxic gases and the fragments cloud, current practices fail
to account for the interdependencies between those hazards, potentially
miscalculating the overall risk on the public and the surrounding
infrastructure. In this work we set two objectives to expand on our claim. The
first is to propose a methodology based on Distribution Envelope Determination
(DEnv) to address the above need. The second on the other hand focuses on
describing the ongoing software development activities aimed at implementing
this methodology as an ArcGIS platform component.
An Object Oriented Library for Sawmill
Simulation
Francisco Javier Ramis (Universidad del Bio Bio, Chile),
Ivan Santelices and Felipe Baesler (Universidad del Bio Bio) and Josef
Sepulveda and Reinaldo Moraga (University of Central Florida)
Abstract:
This work shows an object oriented library that was
designed to facilitate the development of simulation studies of sawmills. A
group technology approach was used to minimize the number of objects in the
library, which includes objects that have the functionality of the equipments
found in these facilities, so that the analyst only needs to provide the
layout, locate the transports and machines, and insert the parameters of the
process. For large sawmills having state of the art production equipment, with
optimization capacities in each machine, two applications of the library are
shown. The study demonstrates the advantage of using an object orientation,
and the usefulness of simulation is a tool for aiding in both the design and
in the improvement of the operations of these facilities.
Risk Analysis Using Monte Carlo Simulation and
Bayesian Networks
Cláudio Flores, Fernando Makiyama, Silvia Nassar,
and Paulo Jose Freitas (Federal University of Santa Catarina, Brazil) and
Carlos Magno Jacinto (PETROBRAS S. A.)
Abstract:
The management of a global activity that has individual
tasks as its components is very difficult, because an unexpected interruption
in any individual task can create extra costs or even disrupt the whole
activity. To resolve this problem, this work presents the development of a
decision-support tool using Bayesian networks (BN). Our research illustrates
how to model the relationship between the total time of a process and the time
of the individual tasks selected as relevant. We use a Monte Carlo simulation
to construct dynamic scenarios on the BN which allow us to track and manage a
global activity. The BN is useful because the activities have random
characteristics and the information about individual tasks can be propagated
throughout the global activity scenario and associated with costs. This offers
the administrator a tool for proactive task management and risk reduction.
Interactive Simulation and Probability
Management
Sam Savage (Stanford University)
Abstract:
A new generation of software performs Monte Carlo
simulation nearly instantaneously on arbitrary spreadsheet models. In effect,
it does for probability distributions what the spreadsheet did for numbers. In
addition, the concept of a Stochastic Information System is introduced, in
which distributions are stored as realizations of a stochastic process. A
single variable is stored as a Stochastic Information Packet (SIP), a
multivariate distribution as a Stochastic Library Unit, Relationships
Preserved (SLURP). This leads to the Fundamental Equality of SLURP algebra:
P(F(x,y)) = F(P(x,y)) where x and y are random variables, P(x,y) is the SLURP
representing their joint distribution, and F is a function of x and y. Live
computer demonstrations will include distribution arithmetic, the Flaw of
Averages, a retirement simulation developed for a New York trust company, and
interactive portfolio model used at a major petroleum firm.
Parallel Simulations of Manufacturing Processing
using SimPy, a Python-based Discrete Event Simulation Tool
Victor M
Castillo (Lawrence Livermore National Laboratory)
Abstract:
SimPy is a Python-based, interpreted simulation tool
that offers the power and convenience of Python. It is able to launch
processes and sub-processes using generators, which act autonomously and may
interact using interrupts. SimPy offers other advantages over competing
commercial codes in that it allows for modular development, use of a version
control system such as CVS, can be made self-documenting with PyDoc, and is
completely extensible. The convenience of an interpreted language, however, is
offset for large models by slower than desired run times. This disadvantage
can be compensated for by parallelizing the system using PyMPI, from the
Lawrence Livermore National Laboratory. This work was performed under the
auspices of the U.S. Department of Energy by University of California,
Lawrence Livermore National Laboratory under Contract W-7405-Eng-48.
SOA-Standard-Conform Modeling as a
High-Level-Standard for Discrete Simulation
Thomas Wiedemann
(University of Applied Science Dresden) and Peter Eichhorst (SOCON Inc.)
Abstract:
Simulation models, created with leading simulation
tools can not be exchanged between the systems, because there is no universal
standard for discrete simulation. This results in very high investments and
maintenance costs for simulation studies and some additional problems with
flexibility in large simulations. A new approach seems now possible by using
Service Oriented Architectures (SOA). The integration of SOA-interfaces in
simulation systems connects them to the IT-infrastructure of ERP-systems,
business process modeling systems and web-services. This option allows a dual
use of the generated models for simulation and real control software. The SOA
standards BPMN and BPEL are also capable for simulation modelling and are
accepted by all major IT companies. The poster presents examples, how BPMN
models can be converted to single simulation programs, which eliminates the
critical performance issues of SOA-webservice interfaces. In conclusion, the
SOA technology can open a new chapter in discrete simulation modelling.
Use of a Computer Algebra Software in the
Functional Validation of Continuous Simulations
Yvon Resplandy
(MBDA)
Abstract:
Due to the numeric integration algorithms with multiple
iterations and the coupling effects between variables, verification and
validation of simulations of continuous systems is a complex task. A method
involving a Computer Algebra Software (CAS) is presented that allows preparing
general parameterized test cases that cannot be achieved by hand. It applies
to flat systems that are causal systems, whose outputs and states can be
expressed through derivation and algebraic operations by a set of given
variables (mechanical systems for example). These variables are set as
temporal expressions with no or few symbolic expansion through derivation such
as trigonometric or polynomial functions. Formal expressions of the inputs and
outputs are calculated by means of a CAS and coded into the environment test.
In addition to the verification and validation, this test is useful to assess
the precision level of the integration method with respect to the variability
of the inputs.
A Visual Aproach To Enhance Discrete-Event Simulation
Model Interoperability
Tai-Chi Wu (I-Shou University) and Allen
Greenwood (Mississippi State University)
Abstract:
With the continued advance of discrete-event simulation
technology, various software packages have evolved that greatly facilitate the
development and analysis of discrete-event simulation models. However, each of
these applications uses a different approach and set of terminology for
representing the behavior of the systems being studied. As a result of this
disparity in implementations, models developed in different applications
cannot interact with other models. Also, the disparity makes it more difficult
for humans to understand the models. In this work, a visual modeling approach
is proposed that allows the modeler, and those not expert in simulation, to
create conceptual models that are independent of the implementation
application. The approach uses eight common simulation elements. This approach
facilitates the model interactions at the model formulation stage and at the
application stage.
SETAC 2 Web: A Web Based Tactical Training
Simulator for the Chilean Army
Luis F. Robledo (Chilean Army,
Computerized Tactical Training Center(CEOTAC)), Jose Robles (Universidad del
Desarrollo/ GBO Ltda.) and Juan C. Perez (GBO Ltda.)
Abstract:
The SETAC2WEB is a web based tactical training
simulator designed for the Chilean Army to train commanders and staff officers
in Decision Making Process from a Division/Brigade level to Battalion level.
This system allows recreation of all phases of war, in real time, and without
physical distance restrictions. SETAC2WEB is being built with state of the art
technology developed between the Chilean Army and the Civilian Industry. This
system has a web based and service oriented architecture with an enhanced
interoperability and integration capability, MIL-STD-2525-B NATO's military
coding system, authentication with digital certificate, and rich graphical
user interfaces.
PLSE-Based Generic Simulation Training Platform for
Typical Weapon Equipments
Ying Liu (Optical and electric
engineering Department of Mechanical Engineering College)
Abstract:
With the development of simulation technologies,
virtual training for the military has become more and more important.
Combining system and software engineering theory, based on the PLSE (Product
Line Software Engineering) idea and method, we analyze and design simulation
training characteristics for typical field equipment. We set up the
domain-oriented system architecture and implement the domain framework. The
research involves: putting forward the conceptual simulation platform based on
PLSE, applying the domain engineering method, analyzing commonalities in the
training for typical equipment, and implementing a general architecture. Using
relevant technologies, we develop reusable core assets systematically and
strategically, and build the object-oriented development platform and the
platform-based developing models.
In Silico Analogues of Epithelial Cell Growth and
Morphogenesis
Sean H.J. Kim and Mark R. Grant (University of
California, Berkeley and San Francisco) and Sunwoo Park and C. Anthony Hunt
(University of California, San Francisco)
Abstract:
We present simulation models of in vitro cell culture
systems and their application in systematic exploration of the simulated cell
growth mechanisms and their impairment in epithelial morphogenesis. Our
referent systems are in vitro cell cultures commonly used for studies of
morphogenetic processes that underlie the development of the embryo and its
organ systems. We have constructed discrete event, agent-based models with
agent rules that capture posited processes of cell proliferation and death
involved in the formation of multicellular epithelial structures such as cysts
and tubules. We report on their application in exploring the phenotypic
consequences of disabling the simulated processes, and show that varying
degrees of the disablement cause different growth patterns to emerge,
including a cancer-like phenotype.
Barriers To and Driving Forces for the
Implementation of Manufacturing Simulation in the Swedish Foundry
Industry
Patrik Thollander (IKP) and Petter Solding (Swecast)
Abstract:
The deregulation of the European electricity market has
led to increasing electricity prices in Sweden, and prices are expected to
raise even further, posing a major threat to the Swedish foundries. This study
present results from a study of barriers to and driving forces for the use of
manufacturing simulation and optimization methods at nine Swedish foundries.
The method used is in-depth interviews and a questionnaire. The largest
barriers were found to be: lack of information regarding what opportunities
manufacturing simulation may gain, lack of time or other priorities, and lack
of staff awareness. The largest driving forces were found to be: lowered
production costs, faster production periods, and possible simulation of the
daily production plan.
Conspace Simulation Application
Tarek M
Zayed (Concordia University)
Abstract:
Simulation becomes essential in the modern world,
especially for construction industry. Simulation principles are used to
provide efficient solutions for different construction operations. The use of
simulation packages becomes popular in modern construction technologies. Using
existing packages involves training on specific applications. In addition,
space constraints for different processes are not entirely presented, and when
presented, the required computing resources are intense, especially for 3D
simulation applications. Therefore, adding space constraints to simulation
packages becomes crucial in order to accommodate the dimensions required by
various resources as well as the available space in the construction
operation. Consequently, Construction Space Simulation Application (CSSA) is
designed to provide a solution for this space problem. The CSSA is coded using
Microsoft Visual Basic 6.0 and Access. It is validated, which shows robust
results.
Comparative Analysis of Process Interaction and
Activity Scanning Approaches in Simulation of Construction
Operations
Rita Awwad, Amir Behzadan, and Omer Tsimhoni (University
of Michigan)
Abstract:
This work illustrates the modeling and implementation
differences between Process Interaction (PI) and Activity Scanning (AS)
approaches in simulation of construction operations. AS has been widely used
in simulation of construction activities while PI is a common approach in
simulating manufacturing and industrial operations. A rock quarry site has
been chosen as a sample construction operation which inherits some degrees of
uncertainty primarily in limited resource levels. The two simulation tools
used are ProModel (for process interaction) and STROBOSCOPE (for activity
scanning). Relevant details of implementation are described on each software
in an attempt to show that both programs can sufficiently be used to model the
task. We compare the results and discuss the differences.
Simulation of the Pig Iron Transportation
System in Companhia Siderurgica Tubarão - Arcelor Group
Alain de
Norman et d'Audenhove (Belge Engenharia e Sistemas)
Abstract:
The objective of this paper is to present the
development of a simulation model involving materials handling in a large
steel company from Arcelor Group. The project had a return of more than 2
million USD and points to a kind of application that seems to be a typical
case where static analysis may cause great errors comparing to a dynamic one
(because the results were also observed in another Brazilian steel company).
The project focused in the processes involving pig iron transportation between
the blast furnaces and its discharge in the steelmaking plant. It allowed the
evaluation of several operational parameters such as for the 7.5 Mt/year
production rhythm, a great increase from the actual production of 5 Mt/year.
“Bottlenecks” in the flow were identified through the results and the
scenarios gave the support to the decision of increasing the number of
torpedo-cars and locomotives in the system.
Supply-Chain Simulation and Analysis of Petroleum
Refinery Systems: A Reusable Template with Incremental
Approach
Luiz Augusto Franzese, Marcelo Moretti Fioroni, and Daniel
Pablo Paz (Paragon Tecnologia), Rui Carlos Botter (Universidade de São Paulo
(USP)) and Carlos Alberto Gratti, Alberto Omar Martinez, and Carlos Maria
Bacigalupo (Petrobras Energía S.A)
Abstract:
Simulation technology has been successfully applied in
Supply Chain studies, mostly to assist onetime decisions, and rarely in
recurrent situations. Several authors have chosen traditional approaches like
developing simulators or developing templates for better modeling, facing
pitfalls like lack or excess of detail, lack of flexibility of tools, how to
continue to evolve models in a dynamic environment etc. This work addresses
application of a systems approach to better describe and model complex
processes. We also demonstrate how a combined approach of templates and
simulators, described as incremental modeling, was used to create a refinery
simulation toolset, and present results of an application at Petrobras
Energia.