ExtendSim Research Grants Fulfilled

The ExtendSim Research Grant program subsidizes a portion of a student's PostDoc, PhD, or Masters research by donating a full version of ExtendSim for use by the student during the term of the project. Here are some of the research projects awarded an ExtendSim Research Grant that have already been completed. To learn about current projects in the ExtendSim Research Grant program, please see Research Grants in Progress

Biofuel Supply Chain

University of Texas"Simulation-Based Approach for the Optimization of a Biofuel Supply Chain"
Hernan Chavez Paura Garcia
University of Texas at San Antonio
PhD in Supply Chain Optimization • February 23, 2017


Project presented at

IISEThe 2017 IISE Annual Conference.


Abstract

The billion-ton study lead by the Oak Ridge National Laboratory indicates that the U.S. can sustainably produce over a billion ton of biomass, annually. However, the delivery of the biomass required to meet the required goals is particularly challenging. This is mainly because of the physical properties of biomass. This research work focuses on the use of agricultural residues to produce second-generation biofuels. Second generation biomass exhibits more quality variability (e.g., higher ash and moisture contents) than first generation. The purpose of this study is to quantify the cost of imperfect feedstock quality in a biomass-to-biorefinery supply chain (SC) and to develop a discrete event simulation coupled with an optimization algorithm for designing a biofuel SC's. This work presents a novel optimization approach based on an extended Integrated Biomass Supply and Logistics (IBSAL) simulation model for estimating the collection, storage, and transportation costs. The presented extension of the IBSAL considers the cost incurred for having imperfect feedstock quality and finds the optimal SC design. The applicability of this methodology is illustrated by using a case study in Ontario, Canada. A converging set of non-dominated solutions is obtained from computational experiments. Sensitivity analysis is performed to evaluate the impact of different scenarios on overall costs. Preliminary results are presented.


IBSAL-SimMOpt Approach

The approach presented in this work is a two-phase model that uses an extension of the IBSAL model in the initial phase, and searches a near-optimal set of solutions in the second phase by using an optimization procedure based on the SimMOpt model. The IBSAL model is a time-dependent discrete event simulation (DES) model with activity-based costing. The model in the proposed approach estimates the cost of imperfect feedstock quality and evaluates its effect on the performance of the SC.

The SimMOpt model is a simulation-based multi-objective optimization approach based on stochastic Simulated Annealing (SA). 

Chavez model

The solutions for near-optimal quality-related costs are found by using the extension of the IBSAL implemented in ExtendSim and the SimMOpt-based procedure included in written in MS VBA®™ language. This figure is a diagram of the proposed approach.


Case Study

The characteristics of the system described in this research include a geographical implementation (i.e., Southern and Western Ontario, Canada), crop availability (ac), corn yield (bu/ac), sustainable production (dry tonne/ac), among others. The geographical implementation corresponds to 20 farms located in the following counties: Lambton, Chatam-Kent, Middlesex, and Huron.

The initial moisture content is modeled by ~U(0.6, 0.8). The moisture content after the natural drying process is assumed to follow ~U(0.15, 0.3). The moisture content requirement for thermochemical processes is 20%. If the content remains above 20%, the stover goes through a mechanical drying process. The cost of the natural air drying system is given by (0.014 * (Initial Moisture Content – Final Moisture Content) * 100) + 0.05 per bushel and the in-bin, stirred system is given by (0.033 * (Initial Moisture Content) * 100) + 0.048 per bushel. Similarly, the initial ash content follows ~U(0.08,0.12). The ash content after the screening process is assumed to follow ~U(0.1, Initial Ash Content). The screening cost is given by 135(Initial Ash Content – Final Ash Content) per dry tonne. The cost of disposing ash is given by 28.86(Final Ash Content) per dry tonne. The binary decision variables in this model represent the decisions of performing the field drying and the screening activities at each farm.


Results and Conclusions

The SA was tuned through designed computational experiments. The SA schedules have a relevant effect on the set of these preliminary solutions. Figure 3 shows the Pareto front including four schedules. The non-dominated solution that balances out the conformance and nonconformance costs (i.e., (0.4/0.6) and (0.6/0.4) weights) shows a conformance cost (the cost incurred to prevent biomass poor quality) of $49,899.47 and a non-conformance cost (cost incurred to fixed biomass poor quality) of $16,496.83 for all 20 farms. This non-dominated solution was found when using the schedule that computes the largest number of initial solutions (i.e., Schedule 4 → 50 initial solutions). This preliminary results highlighted the trade-off between conformance and conformance activities implemented into a biofuel SC. The impact of quality-related activities in the SC topology, design and planning decisions will be studied next.


Acknowledgements

This material is based upon work supported by the U.S. Department of Energy, Office of Energy Efficiency and Renewable Energy, Bioenergy Technologies Office (4000142556) and U.S. Department of Agriculture/National Institute of Food and Agriculture (2015-38422-24064). The fellowship from the Mexican Council for Science and Technology (CONACYT) is gratefully acknowledged. The support provided by Imagine That!®™ by donating of a full version of ExtendSim®™ through the ExtendSim Research Grant is gratefully acknowledged. The research work on the IBSAL model by Sokhansanj, Turhollow, Ebadian and Webb was relevant for the development of the proposed approach.


pdf buttondownload paper

Download ExtendSim modeldownload ExtendSim model


Previous Publications and Refereed Conference Proceedings

Aboytes, M., Chavez, H., Krishnaiyer, K., Stankus, S. & Taherkhorsandi, M. "Improving Radio Frequency Identification Accuracy in a Warehouse Setting". Abstract accepted in the 2016 Engineering Lean & Six Sigma Conference, San Antonio, TX, September 14 - 16, 2016.

line200

Chávez, H. & Castillo-Villar, K. K. "Stochastic Multi-Objective Simulated Annealing for the Optimization of Machining Parameters". Abstract accepted in the 2016 Industrial and Systems Engineering Research Conference (ISERC), Anaheim, California, May 21 - 24, 2016.

2016 ISERC Best Track Paper: Manufacturing and Engineering Design.

line200

Chávez, H., Castillo-Villar, K. K., Herrera, L., & Bustos, A. "Simulation-based multi-objective model for supply chains with disruptions in transportation". Robotics and Computer-Integrated Manufacturing.. Ms. Ref. No.: RCIM-D15-00130. 2016.

line200

Stankus, S., Chávez, H., Castillo-Villar, K. K., & Feng, Y. "A Simulation-based Optimization Approach to Modeling a Fast-track Emergency Department". Abstract accepted in the 2015 Industrial and Systems Engineering Research Conference (ISERC), Nashville, Tennessee, May 30 - June 2, 2015.

line200

Chavez, H. & Castillo-Villar K.K. "A Preliminary Simulated Annealing for Resilience Supply Chains". Paper on proceedings of the IEEE Symposium Series on Computational Intelligence 2014, Orlando, Florida, USA, December 12, 2014.

line200

Chavez, Hernan, Castillo-Villar, K. K., Herrera, Luis & Bustos, A. "Simulation-based Optimization Model for Supply Chains with Disruptions in Transportation". Paper submitted on the Flexible Automation and Intelligent Manufacturing (FAIM) Conference 2014, San Antonio, TX, May 20-23. 2014.


Scientific Posters

Chavez, H., Webb, E., Castillo-Villar, K.K., Ebadian, M., & Sokhansanj, S. "Modeling Cost of Quality in a Discrete Event Biomass Supply Chain". IBSS (Southern Partnership for Integrated Biomass Supply Systems) Annual Meeting. July 27th, 2016.

line200

Chavez, H., Webb, E., Castillo-Villar, K.K., Ebadian, M., & Sokhansanj, S. "Modeling Cost of Quality in a Discrete Event Biomass Supply Chain". ORISE (Oak Ridge Institute for Science and Education) Summer Graduate, Post Graduate, Employee Participant, and Faculty Poster Session. August 9th, 2016.


Technical Reports

Castillo-Villar, K. K., Rogers, Dwain & Chavez, Hernan. "AFV's Fleet Replacement Optimization (Alternative Transportation Initiatives)". (City Public Services (CPS) Energy (San Antonio Water System (SAWS)) & Austin Public Transit (Capital Metro):). Summer 2015. Granting Agency: CPS through Texas Sustainable Energy Research Institute. 2015.

line200

Castillo-Villar, K.K., & Chávez, H. "Simulation-based Optimization Model for Supply Chains with Disruptions in Transportation", Mexico Center Educational Research Fellowship - International Study Fund, 2013-2014, Funded by Mexico Center, UTSA. 2014.

line200

Castillo-Villar, K.K., & Chávez, H. "Reliability Project in Toyota Manufacturing from Reactive to Proactive Maintenance, San Antonio, TX". The University of Texas at San Antonio, Technical Report, June 2014, 68 and 21 pages. Granting Agency: Toyota Motor Manufacturing Texas. 2014.


Biomass to Commodity Chemicals

University of Tulsa"Simulation-Based Optimization of Biomass Utilization to Energy and Commodity Chemicals"
Ismail Fahmi
University of Tulsa
PhD in Chemical Engineering • August 2013


Project presented at

fahmi aiche12AIChE Conference
Pittsburgh, PA
November 1, 2012

Go to ExtendSim presentationOral presentation
Go to ExtendSim presentationPoster presentation


Project published in

IChemEChemical Engineering Research and Design
Special Issue: Computer Aided Process Engineering (CAPE) Tools for a Sustainable World
Volume 91, Issue 8 • August 2013

Go to ExtendSim paperArticle


Abstract

Incorporating non-traditional feedstocks, e.g., biomass, to chemical process industry (CPI) will require investments in research & development (R&D) and capacity expansions. The impact of these investments on the evolution of biomass to commodity chemicals (BTCC) system should be studied to ensure a cost-effective transition with acceptable risk levels. The BTCC system includes both exogenous, e.g., product demands (decision-independent) and endogenous, e.g., the change in technology cost with investment levels (decision-dependent) uncertainties.

This paper presents a prototype simulation-based optimization (SIMOPT) approach to study the BTCC system evolution under exogenous and endogenous uncertainties, and provides a preliminary analysis of the impact of using three different sampling methods, i.e., Monte Carlo, Latin Hypercube, and Halton sequence, to generate the simulation runs on the computational cost of the SIMOPT approach. We realized that the simulation-based optimization framework that we developed has two major computational costs:

the cost associated with the required number of samples to obtain the model with sufficient statistical significance.

the cost associated with solving the optimization problem.

The results of a simplified case study suggest that annual demand increases is the dominant factor for the total cost of the BTCC system. The results also suggest that using Halton sequence as the sampling method yields the smallest number of samples, i.e., the least computational cost, to achieve a statistically significant solution. Other conclusions that we can take from this work are:

BTCC system evolution is highly dependent on the costs of the raw material.

Investing in the non-renewable processing technologies is attractive in long term.

The developed SIMOPT framework can be used to include the endogenous and exogenous uncertainties of the BTCC system evolution.

The developed optimization module is universal, meaning that it is independent of technology types.

The developed optimization module can also incorporate the relationships between maturity stage and the ability to incorporate to overall production.

Halton series is shown to be the most appropriate sampling method.

Bilinear relaxation with linearly segmented tight relaxation coupled with nonlinear terms relaxation with linear upper and under estimators is shown to be able to obtain a good initialization for the optimization module.


pdf buttondownload paper


Other Publications by this Researcher

Fahmi, Ismail and Selen Cremaschi. "A Prototype Simulation-based Optimization Approach to Model Feedstock Development for Chemical Industry". Proceedings of the 22nd European Symposium on Computer Aided Process Engineering, 2012.

line200

Fahmi, Ismail and Selen Cremaschi. "Stage-gate Representation of Feedstock Development for Chemical Process Industry". Foundations of Computer-Aided Process Operations, 2012.


Call Centers -- Routing Rules

UCSC"Knowledge Management in Call Centers: How Routing Rules Influence Expertise in the Presence of On-the-Job Learning"
Geoffrey Ryder
University of California at Santa Cruz
PhD in Operations Research • March 16, 2011


Abstract

In this paper, the effect that routing rules have on agent learning is researched. A nonlinear optimization framework for two kinds of expertise objectives are developed: one that seeks equal distribution of experience across the workforce (effectively cross-training) and one that aims to develop specialized expertise by prioritizing the routing of specific customer inquiries to specific agents. Analytical models of call center operations are inadequate to handle this task, so instead we turn to discrete-event simulation, and evaluate the effect of routing policies on agent expertise with a custom simulator developed in the ExtendSim modeling environment. Simulation results describe an efficient frontier in routing policies that depends on the underlying expertise objective function.


pdf buttondownload paper

Download ExtendSim modeldownload ExtendSim model


Other Publications by this Researcher

Ryder, G. "Managing Changing Service Capacity Based on Agent Performance Data". INFORMS Annual Meeting, Service Industry III Session, Nov. 7, 2007.

line200

Ryder, G. "How Learning and Forgetting Affect the Optimal Work Policy." INFORMS Annual Meeting, Management of Complex Service Systems Session, Nov. 4, 2007.

line200

Ryder, G. and Ross, K. "Optimal Service Rules in the Presence of Learning and Forgetting". Sixteenth Annual Frontiers in Service Conference, San Francisco. October 4, 2007.

line200

Ryder, G., Ross, K., and Musacchio, J. "Optimal service policies under learning effects". International Journal of Services and Operations Management, Issue 6, Vol. 4, 2008.

line200

Ryder, G. and Ross, K. "Optimal service policies in the presence of learning and forgetting". Applied Probability Track, INFORMS Annual Conference, Pittsburg, PA, 2006.

line200

Ryder, G. "A probability collectives approach to weighted clustering algorithms for ad hoc networks". IASTED CCN Conference, Marina Del Rey, CA, October 2005.


Chemical -- Processing Multi-Products

Carnegie Mellon University"An Efficient Method for Optimal Design of Large-Scale Integrated Chemical Production Sites with Endogenous Uncertainty"
Sebastian Terrazas-Morenoa, Ignacio E. Grossmanna, John M. Wassick, Scott J. Bury, Naoko Akiya
Carnegie Mellon University
PhD in Chemical Engineering • March 2012


Project published in

Computers & Chemical Engineering
Volume 37 • February 2012


Abstract

Integrated sites are tightly interconnected networks of large-scale chemical processes. Given the large-scale network structure of these sites, disruptions in any of its nodes, or individual chemical processes, can propagate and disrupt the operation of the whole network. Random process failures that reduce or shut down production capacity are among the most common disruptions. The impact of such disruptive events can be mitigated by adding parallel units and/or intermediate storage. In this paper, the design of large-scale, integrated sites considering random process failures in addressed. In a previous work (Terrazas-Moreno et al., 2010), a novel mixed-integer linear programming (MILP) model was proposed to maximize the average production capacity of an integrated site while minimizing the required capital investment. The present work deals with the solution of large-scale problem instances for which a strategy is proposed that consists of two elements. On one hand, we use Benders decomposition to overcome the combinatorial complexity of the MILP model. On the other hand, we exploit discrete-rate simulation tools to obtain a relevant reduced sample of failure scenarios or states. We first illustrate this strategy in a small example. Next, we address an industrial case study where we use a detailed simulation model to assess the quality of the design obtained from the MILP model.


pdf buttondownload paper 



Original paper

"A Mixed-Integer Linear Programming Model for Optimizing the Scheduling and Assignment of Tank Farm operations"
Sebastian Terrazas-Morenoa, Ignacio E. Grossmanna, John M. Wassick
Carnegie Mellon University


Abstract

This paper presents a novel mixed-integer linear programming (MILP) formulation for the Tank Farm Operation Problem (TFOP), which involves simultaneous scheduling of continuous multi-product processing lines and the assignment of dedicated storage tanks to finished products. The objective of the problem is to minimize blocking of the finished lines by obtaining an optimal schedule and an optimal allocation of storage resources. The novelty of this work is the integration of a tank assignment problem with a scheduling problem where a dedicated storage tank has to be chosen from a tank farm given the volumes, sequencing, and timing of production of a series of products. The scheduling part of the model is based on the Multi-operation Sequencing (MOS) model by Mouret et al., (2011). The formulation is tested in three examples of different size and complexity.


pdf buttondownload original paper 


Construction -- Production Variability

Pontificia Universidad"Work-In-Process Buffer Design Methodology for Scheduling Repetitive Building Projects"
Vicente González, Luis Fernando Alarcón, and Pedro Gazmuri
Pontificia Universidad de Católica de Chile
PhD in Construction • July 2008


Abstract

gonzalez modelVariability in production is one of the largest factors that negatively impacts construction project performance. A common construction practice to protect production systems from variability is the use of buffers (Bf). Construction practitioners and researchers have proposed buffering approaches for different production situations, but these approaches have faced practical limitations in their application. 

In Multiobjective Design of Work-In-Process Buffer for Scheduling Repetitive Building Projects, a multiobjective analytic model (MAM) is proposed to develop a graphical solution for the design of Work-In-Process (WIP) Bf in order to overcome these practical limitations to Bf application, being demonstrated through the scheduling of repetitive building projects. Multiobjective analytic modeling is based on Simulation–Optimization (SO) modeling and Pareto Fronts concepts. Simulation–Optimization framework uses Evolutionary Strategies (ES) as the optimization search approach, which allows for the design of optimum WIP Bf sizes by optimizing different project objectives (e.g., project cost, time and productivity). The framework is tested and validated on two repetitive building projects. The SO framework is then generalized through Pareto Front concepts, allowing for the development of the MAM as nomographs for practical use. The application advantages of the MAM are shown through a project scheduling example. Results demonstrate project performance improvements and a more efficient and practical design of WIP Bf. Additionally, production strategies based on WIP Bf and lean production principles in construction are discussed.


pdf buttondownload paper

Download ExtendSim modeldownload ExtendSim models


Other Publications by this Researcher

González, V., Alarcón, L.F. and Gazmuri, P. "Design of Work In Process Buffers in Repetitive Projects: A Case Study". 14th International Conference for Lean Construction, Santiago, Chile, July 2006.

line200
González, V. and Alarcón, L.F. "Design and Management of WIP Buffers in Repetitive Projects" (White Paper), 2005.

line200

González, V., Rischmoller, L. and Alarcón, L.F. "Management of Buffers in Repetitive Projects: Using Production Management Theory and IT Tools". PhD Summer School, 12th International Conference for Lean Construction, Helsinore, Denmark, August 2004.

line200

González, V., Rischmoller, L. and Alarcón, L.F. "Design of Buffers in Repetitive Projects: Using Production Management Theory and IT Tools". 4th International Postgraduate Research Conference, University of Salford, Manchester, U.K., April 2004.

line200

González, V. and Alarcón, L.F. "Buffer de Programación: Schedule Buffers: A Complementary Strategy to Reduce the Variability in the Processes of Construction", Revista Ingeniería de Construcción Pontificia Universidad Católica de Chile, Volume 18 . Nº 2, pp 109 - 119, Mayo - Agosto 2003.


Cross-Docking -- Part I

University of Pennsylvania"CDAP Simulation Report"
Zongze Chen
University of Pennsylvania
Masters in Electrical & Systems Engineering • May 13, 2010


Abstract

In this project we collaborated with National Retail Systems (NRS) who provided us with realistic cross-dock situations and data to study and to evaluate our modeling and optimization result. The NRS's North Bergen facility receives goods from multiple vendors, sorts and loads them onto outbound trailer trucks for a number of retail stores. Using ExtendSim, I developed a simulation  to model of two cross-docks in operation of their New Jersey facility. There were several main objectives to be achieved:

Develop the model of cross-dock process. For simplicity, we started the development with a 4×4 cross-dock. Later, this would be expanded to realistic dimensions.

Use the data generated from the GQ3AP algorithm we developed previously to simulate the process and analyze the total cost under the situation and discuss the optimization.

Improve our optimization models to take into account the impact of truck arrival and departure times. Determine how one can improve cross-docking operations and what costs could be reduced through improved operational control.

The final report describes how simulation helps ensure success of cross-docking systems by determining optimal routing costs. Modeling methods and issues are also discussed as they apply to cross-docking. This report includes discussion of the actual processes employed by NRS, description of our models, simulation results and comparisons, and our conclusions.


Approach

The initial phase of the project involved data collection and acquisition including OD freight volumes and scheduling times. In the next phase in cooperation with NRS, we developed a discrete cross-dock simulation model and found the best optimization methods. Finally, we incorporated what we  learned into a set of suggested optimization procedures.


pdf buttondownload paper

Download ExtendSim modeldownload ExtendSim model


Other Reports by this Researcher

Course Reports in OPIM 910 course (introduction to optimization theory):
Personalized Diet Plans Optimization
University of Pennsylvania Evacuation Plan Design
Classroom Scheduling

Course Reports in STAT 541 course (Statistics Methods):
Boston Real Estate Analysis

Course Reports in STAT 520 course (Applied Econometrics):
Salary Weight Analysis by Gender

Research Assistant on Project at Management Department of Wharton:
Entrepreneurship and Strategy


Cross-Docking -- Part II

University of Pennsylvania"Anticipating Labor and Processing Needs of Cross-Dock Operations"
Frederick Abiprabowo
University of Pennsylvania
Operations Research • December 2012


Project presented and awarded at

Abiprabowo awardDuring the INFORMS Annual Conference 2013, the design team of Frederick Abiprabowo, Napat Harinsuit, Samuel Lim, and Willis Zhang, all from the University of Pennsylvania, were awarded the Undergraduate Operations Research Prize. This competition is held each year to honor a student or group of students who conducted a significant applied project in operations research or management science, and/or original and important theoretical or applied research in operations research or management science, while enrolled as an undergraduate student.

Their paper "Designing a Simulation Tool for Commercial Cross-Docking Application" uses ExtendSim to dynamically replicate the operations of a large cross docking facility. The prototype model serves as a tool that enables cross-dock operators to evaluate assignment strategies in a risk-free, costless environment.


Abstract

In the supply chain literature, there are generally three kinds of studies regarding cross-docking:

fundamentals of cross-dock

distribution planning

operations inside the facility

Studies of cross-dock fundamentals take a high-level perspective and discuss the issues of cross-docking in relations to a company’s distribution process, management, etc. Cross-docking can also be viewed as a method in the management of a supply chain. Distribution planning problems pertain to scheduling of trucks, vehicle routing, and network navigation. Finally, there are many research papers regarding the operations inside cross-docks.

Zongze Chen’s CDAP simulation report dealt with the operations inside the cross-docking facility and discussed the formulation of goods movement inside a cross-dock as a GQ3AP. Then, a simulation model was designed after one of the cross-docking facilities of a retail company in New Jersey. This model was then tested to compare the cost of goods movement in two different scenarios: 

GQ3AP-optimized door assignment

manual door assignment

Based on the results of this test, the scenario with optimized door assignment did yield a lower cost than the one without.

This report continues where the previous report left off and addressed its proposed future plans for the model.


pdf buttondownload paper

Download ExtendSim modeldownload ExtendSim model


Education -- Optimizing Performance of Educational Processes

University of Split"Higher Education Management in Relation to Process Organization Theory"
Maja Cukusic
University of Split
PhD in Business Informatics • March 2011


Project published as

"Online self-assessment and students' success in higher education institutions"
Maja Ćukušić, Željko Garača, Mario Jadrić
Computers & Education - An International Journal
March 2014 • ISSN 0360-1315


Abstract

Cukusic modelThis paper validates effects of online self-assessment tests as a formative assessment strategy in one of the first year undergraduate courses. Achieved students' results such as test scores and pass rates are compared for three different generations for the same course but also judged against the exam results of other courses taught in the same semester. The analysis points out that there is a statistically significant difference between the groups for half-semester tests and exam pass rates after online self-assessment tests were introduced. Positive effects on students' success are approximated for the overall institution using a simulation model. Results point out that a small increase in pass rates could significantly impact the overall success i.e. decrease of dropout rates.

The research model that preceded the simulation explored the correlation between ICT support of educational processes and their outcomes. In order to test the impact of ICT support a sub-process was chosen: extensive self-evaluation quizzes delivered via e-learning system were designed, implemented and monitored within a first-year university course. The results were controlled and measured with regards to students’ outcomes achieved during the previous and current academic year in several courses (horizontal and vertical control of the results). Given the correlations between variables that characterize support of the educational process and outcomes on tests and exams, ICT support of the educational process has a positive effect as expressed in terms of relevant performance indicators.

A simulation model was developed which allows extrapolation of the impact on key performance indicators (i.e. drop-out rate and study completion time) for the whole institution enabling analysis of potential opportunities. The model adheres to study regulations of the Faculty of Economics (University of Split) and simulates outcomes for a generation of undergraduate students. Simulated results were compared with the actual data from the information system to verify the correctness of the model.

Not all course environments allow implementation of self-evaluation quizzes that result in slightly better exam pass-rate (roughly about 3%). Consequently, the simulation experiment investigates the process change for only half of the courses and only for the largest group of students (60%). As a result, the percentage of students who drop out from their studies could be significantly lower, 36% compared to 45.67% in real-life. For the entire system, this relatively small per-course improvement in exam results has a strong overall effect.


pdf buttondownload paper (in English)

pdf buttondownload paper (in Croatian)

Download ExtendSim modeldownload ExtendSim model


Other Publications by this Researcher

For a complete list of publications by Maja Ćukušić, please go to http://bib.irb.hr/lista-radova?autor=300571.

Healthcare -- Emergency Department Crowding

Wharton School"The Financial Consequences of Lost Demand and Reducing Boarding in Hospital Emergency Departments"
Bob Batt
University of Pennsylvania • Wharton School
PhD in Operations Management • April 2011


Project published in and presented at

Annals of Emergency MedicineResults of Mr. Batt's project was published in the Annals of Emergency Medicine in October 2011. The American College of Emergency Physicians thought the findings were so significant that they issued a press release about the paper. The main find was that reducing emergency department boarding by one hour could generate approximately $2.7M per year if dynamic admitting policies are used to control elective patient arrivals.  POMSThis is a novel finding in that no previous work has put together the revenue gains from the ED of reducing boarding with the potential revenue reductions from reducing elective patients. Mr. Batt presented this paper at the Production & Operations Management Society conference Conference in Reno, Nevada on April 29, 2011.


Abstract

Bob Batt modelThis project explores the operational ramifications of crowding in hospital emergency departments. A common indicator of crowding is patients “boarding” in the emergency department while awaiting transfer to an inpatient bed in the hospital. Boarding is a controversial topic in the medical community because it has been suggested that it is a way to tacitly prioritize high-dollar elective patients over lower-value emergency patients. However, the financial impact of boarding is not obvious since boarding creates congestion in the emergency department leading to higher levels of lost demand from patients leaving without treatment and ambulances being diverted. We use discrete event simulation to model a hospital under various boarding regimes and patient prioritization schemes. We find that reducing boarding can be not only operationally efficient but also financially beneficial for the hospital.This project explores the operational ramifications of crowding in hospital emergency departments. A common indicator of crowding is patients “boarding” in the emergency department while awaiting transfer to an inpatient bed in the hospital. Boarding is a controversial topic in the medical community because it has been suggested that it is a way to tacitly prioritize high-dollar elective patients over lower-value emergency patients. However, the financial impact of boarding is not obvious since boarding creates congestion in the emergency department leading to higher levels of lost demand from patients leaving without treatment and ambulances being diverted. We use discrete event simulation to model a hospital under various boarding regimes and patient prioritization schemes. We find that reducing boarding can be not only operationally efficient but also financially beneficial for the hospital.


nextpageotherview paper

Download ExtendSim modeldownload ExtendSim model of ED boarding

Download ExtendSim modeldownload ExtendSim model of whole hospital


Healthcare -- Nanomedicine

USC"Nanomedicine"
Janet Cheung
University of Southern California
Non-profit research project • January 2015


Project details

As part of a completely non-profit research project organized by our advisor, we aimed to answer the question of how effective the methodology of using nanobots can be for curing cancer as compared to the existing cancer drug because nanobots could integrate the diagnosis and treatment of cancer in a cohesive, potentially non-invasive unit through precise, targeted operations on the cellular level.

The key problems of conventional technology are the methods of drug delivery and the concentration of the drug cocktail required to destroy the cancerous cells. So nanobots would potentially allow the drugs to be directed to exact location where cancerous cells have been observed. Thus, only the malignant tissues are affected, and healthy tissues are not. As a result, nanomedicine does have the potential to revolutionize the way medicine is practiced around the world, but it is clear that biocompatibility on the nanoscale is one of many major challenges that must be overcome.

Cheung modelThe model created focuses on drug delivery via the bloodstream (which is the most common and preferred method of delivery). It has the skeletons for active and passive targeting nanobots. The nanobots simulate flowing through the bloodstream until it is attracted to the tumor. Additionally, the model accounts for various possible drug delivery failures (i.e. early deployment of the drug, power failure, and time constraints). For now, the model determines when the failures occur using various probability distributions as placeholders for more legitimate values. Upon failure, the representation of the nanobot exits the model. The number of each type of failure can be viewed in a bar graph. In the event of a successful nanobot to tumor attachment, the representative nanobot in the model incurs latency, which accounts for attaching to the tumor and actual drug deployment. The number of successful deliveries to the number of failed deliveries can be seen in another bar graph available.The model focuses on drug delivery via the bloodstream (which is the most common and preferred method of delivery). It has the skeletons for active and passive targeting nanobots. The nanobots simulate flowing through the bloodstream until it is attracted to the tumor. Additionally, the model accounts for various possible drug delivery failures (i.e. early deployment of the drug, power failure, and time constraints). For now, the model determines when the failures occur using various probability distributions as placeholders for more legitimate values. Upon failure, the representation of the nanobot exits the model. The number of each type of failure can be viewed in a bar graph. In the event of a successful nanobot to tumor attachment, the representative nanobot in the model incurs latency, which accounts for attaching to the tumor and actual drug deployment. The number of successful deliveries to the number of failed deliveries can be seen in another bar graph available.

The model also has toxicity and biocompatibility factored into it. The body has an assumed toxicity capacity, which is the level that it can contain the toxins and still be healthy. Each drug has its own toxicity level, which is inversely proportional to its biocompatibility. The nanobots release the toxins into the bloodstream per failed drug delivery, which gradually increases the current toxicity level in the body. As with the probability distributions for the drug delivery failures, I have implemented placeholder values. These placeholder values can all be easily replaced with more accurate values with further research.


Further research

In June of 2016, Michael Schlesinger of USC was awarded an ExtendSim Research Grant to take Ms. Cheung's research one step further. Mr. Schlesinger is investigating how to design and fabricate a robot with biopsy, imaging, and drug delivery capabilities. Details can be found on the Research Grants in Progress page.


Healthcare -- Patient Flow

Purdue University"Essays on Efficiency in Service Oeprations: Applications in Health Care"
John Norris
Purdue University
PhD in Management, Quantitative Methods • December 2007


Abstract

Partnering with Indiana University Medical Group (IUMG), the author focused on outpatient care:

Address the issue of missed appointments.

Analyze variability in patient flow.

Analyze performance of phone system.

...to analyze and improve patient flow at an outpatient clinic of the Indiana University Medical Group. Queuing concepts were used to uncover sources of variability and to generate ideas to improve clinic operations that would mitigate the undesirable effect of variability. norris modelA process map, that matched the process at the clinic, was developed and validated. Data on task times was collected by observing the process with stopwatch or from historical records. A simulation model corresponding to the process map was developed, and the output was validated. Several ideas to modify clinic operations were tested on the validated simulation model. The overall result was an improvement in both the mean and the standard deviation of patient wait time, as well as higher utilization of physicians’ time. The clinic has implemented several of their recommendations and experienced improvements consistent with model predictions.


Go to papergo to paper


Healthcare -- Public Health

Yale University"Choosing Number and Scheduling Priority of Warm-hand Offs: A DES Model"
Evelyn Cumberbatch
Yale University
Masters Thesis for Yale School of Public Health • May 2014


Abstract

The integration of behavioral health care into primary care is being promoted as a means to treat more people with behavioral health problems where they are most likely to be seen. Clinics with traditional behavioral health services may open slots among scheduled appointments to see these "warm-hand off" (WHO) patients identified by primary care providers (PCPs). The effects of giving priority for behavioral health appointments to either scheduled or WHO patients and of the number of appointments left open for WHO patients are investigated in this project.

cumberbatch modelA discrete event simulation model was built of a moderately integrated clinic. WHO patients arrive randomly, on average 4 per day per PCP, and wait to see behavioral health providers (BHPs) who also see scheduled patients. Simulations of four clinic sizes, with PCP to BHP ratios of 1:1, were run. Effects of queue discipline (priority is given to scheduled or WHO patients) and the number of open WHO slots (3 or 5) are analyzed. Outcomes include the percent of scheduled patients served, the percent of WHO patients served, and the percent of BHP utilization.

In clinics with 1 PCP and 1 BHP, for 3 and 5 open slots respectively, giving priority to WHO patients resulted in 80.6% and 81.0% of WHO patients served and 84.4% and 86.6% of scheduled patients served, however, giving priority to scheduled patients resulted in 97.8% and 98.1% of scheduled patients served, but 32.0% and 47.9% of WHO patients served. A similar pattern was seen for larger clinics, though the percent of WHO patients served increased for both 3 and 5 open slots with clinic size. Having 3 or 5 open slots led to few differences when WHO patients were given priority, but when scheduled patients were given priority, choosing 5 open slots rather than 3 open slots, increased the percent of WHO patients served by 15-20 percentage points across the clinic sizes. In either queue discipline, changing from 3 to 5 open slots reduced the percent of BHP utilization by approximately 8 percentage points for all clinic sizes. When WHO patients were given priority, the average wait time for scheduled patients increased from approximately 2-5 minutes to 13-19 minutes across clinic sizes.

These results might suggest to some clinics attempting to integrate primary care and traditional behavioral health services to choose to give WHO patients priority. However, it is recognized that there are costs associated with not seeing both scheduled and WHO patients, and clinics making this decision will have to weigh these tradeoffs. The analysis of these results provides one framework to assist in choosing between different arrangements for integration.


Go to papergo to paper

Download ExtendSim modeldownload ExtendSim paper and model


Other Publications by this Researcher

During Dr. Cumberbatch's psychiatry residency at University California, San Francisco:

Pole, N., Cumberbatch, E., et. al. "Comparisons Between High and Low Peritraumatic Dissociators in Cardiovascular and Emotional Activity While Remembering Trauma",  Journal of Trauma and Dissociation, 2005; 6(4): 51-67.


Information Security

Gjovik University"Analyzing Security Decisions with Discrete Event Simulation"
Magnus Felde
Gjøvik University
Masters in Information Security • June 28, 2010


Abstract

As organizations become increasingly more dependent on information security in order to succeed, the security decisions made by the Chief Information Security Officer (CISO) also becomes important and needs to be considered in the context of the organization. However, since the complexity of the organization's internal processes and the threats the organization is facing, the CISO needs a decision making tool or method in order to determine the effects of a specific security decision. Because of this, we have in this thesis determined the suitability of utilizing Key Performance Indicators (KPIs) and Discrete Event Simulation (DES) as a method to help the CISO make the "best" security decision for his organization.

Felde modelThe thesis is based on a health care specific scenario which has been constructed in collaboration with Akershus University Hospital (Ahus), Rheumatism Hospital at Lillehammer and Buypass. The scenario includes a patient treatment process and the processes related to the usage of smart cards and passwords as authentication mechanisms. Furthermore, KPIs which focuses on time usage and number of deviations has been identified, where deviations within this health care specific scenario relates to more traditional security incidents.

A case study was then conducted based on the scenario. The results of this case study indicate no statistical significant difference between the two authentication mechanisms with regards to the average time a doctor uses on a business activity. However, based on the number of deviations identified, smart cards were determined the preferred security measure of the two.

In order to determine the suitability of the simulation approach, a second case study was also conducted. This second case study was based on the same scenario, but this time with a non-simulation approach. By comparing the process surrounding the two case studies, the non-simulation approach were determined the most cost-effective approach and the approach which provided the most direct link between the input data and the results. Based on this, the non-simulation approach was also determined the most suitable approach. However, we did determine that for "what if" analysis, the simulation approach becomes the best choice of the two.

Should a "what if" analysis be desirable, we have in this thesis proposed a new methodology which modelers can utilize in order to reduces the complexity of the model building process. The methodology, called Minimalistic Model Design (MIMD), excludes the temporal relationship between the identified business activities within the business process. This exclusion helps to reduce total time used on the model building process, and enables better scalability.


pdf buttondownload paper

pdf buttondownload ExtendSim model


Kanban Approach in Software Development

National University of Ireland, Galway"Limiting Work in Progress but Liberating Progress in Work; Utilising Process Simulation and Theories of Social Science to Understand the Value of Kanban for Software Development"
James Moriarty
National University of Ireland, Galway
Masters in Information Technology • August 2014


Abstract

"The Toyota Production System" first utilised just-in-time with Kanban as a method of visualising the work being completed. Inheriting from lean and agile software development philosophies, Kanban emerged for software development. The Kanban approach has few rules yet it is a powerful tool to evolve a software process.

moriarty modelIn this study, ExtendSim simulation software is used to model a Kanban approach. The ease of model creation and the potential for adaptability are both assessed. In addition, literature from social sciences are examined for evidence to support the benefits Kanban can bring to people in the software development process.

This study succeeded in creating a simulation of the Kanban approach for software development. A Kanban board was simulated to serve as a foundation from which other processes could be modelled accurately through modification or evolution. Lead time and cycle time indicate performance depending on the changes to work-in-progress limits of the various stages.

Evaluation of social science literature provides support for the human benefits associated with the implementation of Kanban. The simulation model provides meaningful data, but lacking real data prevents the model being broadly valid. Kanban and simulation appear to complement one another which is worthy of future study.


Download ExtendSim modeldownload paper and ExtendSim model


Lean Manufacturing

Middle Tennessee University"Simulation of Lean Manufacturing Concepts using Extendsim Software"
Mark Young
Middle Tennessee State University
Engineering Technology • December 2012


Project description

In support of ET 6390 Lean Manufacturing, the visualization of core concepts for the class is important for increased comprehension. After polling the engineering and operations management curriculum about the use of simulation software, Mr. Young discovered it is not practiced on a frequent basis in either college at MTSU. 

Using simulation software, specifically ExtendSim, as a tool for visualizing the push/pull/constrained WIP modeling, Mr. Young made the work available and in a form that, if so desired, could be readily adapted to the curriculum of ET 6390.


Project outcome

Mr. Young produced 3 videos for the course ET 6390 - Productivity Strategies/Lean Systems using ExtendSim:

Simulating Little's Law with Penny Fab. Four identical tools in series.

Penny Fab 2. Four stations with different processing times.

Push & Pull Simulation. Side by Side Model.


Go to paperproject charter

YouTubeYouTube Channel with models


Manufacturing -- Creating a Production System

Lulea University of Technology"Simulation of a Manufacturing Process - Military Aircrafts"
Sandra Fors
Luleå Technical University
Master of Science in Engineering Technology • July 2016


Abstract

fors saabThe military aircraft manufacturing division at Saab Aeronautics is ready to enter the future of production systems and leap ahead in their field. Saab predicts an increase in demand of military aircrafts. To meet this new higher customer demand, this master's thesis is a part of a project to create a new production system.The military aircraft manufacturing division at Saab Aeronautics is ready to enter the future of production systems and leap ahead in their field. Saab predicts an increase in demand of military aircrafts. To meet this new higher customer demand, this master's thesis is a part of a project to create a new production system.

Ms. Fors' Master's thesis includes following steps, each of which she thoroughly validated:

A simulation model of a production system for military aircraft manufacturing.

A system analysis and identification of bottlenecks through investigation of the queues in the simulation model.

Testing of improvement suggestions.

A simulation model which tests different production volumes.

Recommendations for how to increase production volume by time, with solutions from two different development processes.

fors modelThe result of this thesis presents concepts on how to build a production system for military aircraft, including number of stations, improvement areas, and results of improvement suggestions on a general level. The simulation model created for this project presented the possibility to test a lot of scenarios and suggest improvements to find the best solution. The use of simulation software as an initial part of this project has proven successful and is recommended for use in the project in the future.


Download paperdownload paper

Go to papergo to abstract


Manufacturing -- Using RFID Technology for Routing

Arab Academy"Evaluating the Impact of Introducing RFID Technology in Manufacturing Systems Using Simulation"
Aly Mohamed Owida
Arab Academy for Science, Technology, & Maritime Transport
Master of Science in Industrial and Management Engineering • May 2011


Project presented at

ICCIE41st International Conference on Computers and Industrial Engineering
October 23 to 26, 2011

 


Abstract

Radio frequency identification (RFID) technology has significant impact on product tracking and identification in manufacturing systems. Most of the business cases that implement the RFID technology in their operations have achieved various benefits. RFID technology can reduce the operating errors that affect the efficiency of the operations which results in improving different performance measures such as cycle time, throughput, work-in-process, resources utilization, and average waiting time in queues. In addition, several benefits such as improved items monitoring, lower lead times, and better inventory control can be achieved by introducing RFID technology. Recent developments in RFID technology and other supporting technologies have created opportunities for real-time traceability and better visibility inRadio frequency identification (RFID) technology has significant impact on product tracking and identification in manufacturing systems. Most of the business cases that implement the RFID technology in their operations have achieved various benefits. RFID technology can reduce the operating errors that affect the efficiency of the operations which results in improving different performance measures such as cycle time, throughput, work-in-process, resources utilization, and average waiting time in queues. In addition, several benefits such as improved items monitoring, lower lead times, and better inventory control can be achieved by introducing RFID technology. Recent developments in RFID technology and other supporting technologies have created opportunities for real-time traceability and better visibility inshop floor operations.

owida modelThis paper investigates the effectiveness of introducing RFID technology in tracking and identification processes for products flow in a job shop manufacturing facility. A leading furniture manufacturer in Egypt has been selected as a case study. The manufacturer produces a large number of customized furniture products. Errors in tracking and identification usually occur due to the large number of products present on the shop floor. Introduction of radio frequency identification technology at different stages of manufacturing is proposed to overcome these errors. Different simulation models have been developed for the post-assembly processes in the facility. These models have been developed with an intent to capture all the features that characterize a real furniture manufacturing facility. Simulation is used to assess the impact of introducing the RFID technology on a number of performance measures. Analysis and comparison of simulation results for the base and proposed models show that RFID implementation can improve the overall performance of the facility.


Download paperdownload paper

Download ExtendSim modeldownload ExtendSim model


Other Publications by this Researcher

A. M. Owida, K. S. El-Kilany, and A. E. El-Sayed. "Analytical Hierarchy Process for Selection of RFID System: An Application in Retail Supply Chains". Proceedings of the 19th International Conference on Flexible Automation & Intelligent Manufacturing, 2010.


Neural Networks -- SimNeural Technology

Universities of Ottawa and Brasilia"SimNeural: Simulador Computacional de RNA para Aplicaçoes em Ensino e Pesquisas Medicas e Biologicas" ("SimNeural: Computer Simulation of ANN for Learning and Research Purposes")
Carlos Alberto Gonçalves
Joint collaboration through the Centre for E-Learning at both the University of Brasília and the University of Ottawa
PostDoc in Education and E-Learning • October 2013


Abstract

This research work is part of collaboration between the University of Brasilia (DF, Brazil) and the University of Ottawa (ON, Canada) and was carried out with other researchers from the University of Brasília: Bruno Bastos Neves, Jussie Marques Martins, and Wilson Henrique Veneziano. This project was designed to improve the quality of teaching materials of the former university with the know-how of the latter.

Computer simulation and pattern recognition by Artificial Neural Network (ANN) techniques has become an important tool for learning and professional purposes. This work reports the phases of design, development, and validation of the SimNeural software, that was conceived to be easy to use by professionals of Medical and Biological Sciences when working in research or learning purposes.

Computer simulation and pattern recognition by Artificial Neural Network (ANN) techniques has become an important tool for learning and professional purposes. This work reports the phases of design, development, and validation of the SimNeural software, that was conceived to be easy to use by professionals of Medical and Biological Sciences when working in research or learning purposes.

SimNeural was implemented in ExtendSim using the technique called component prototyping and recycling. SimNeural´s main objective is to be a framework where the user can easily configure and simulate models of ANN. For instance, it should be easy to implement hypothesis tests or to check the relationship among variables. To achieve that goal, the authors used ANN multilayer perceptrons and backpropagation learning. For the validation and test of this tool, they asked SimNeural to build and train some models.

This paper presents the following results:

The ANN training phase that was used for the validation of the software.

Sample analysis results.

Experimental test results designed for pattern recognition and classification of real biological and medical databases.The results showed that SimNeural can be a useful and effective tool for applications at research and learning.


Download paperdownload paper (available in Portuguese only) 


Other Publications by this Researcher

Venezuano, W.H.; Rocha, A.F.; Goncalves, C.A.; Pena, A.J.; Carmo, J.C.; Nascimento, F.A.O.; Rainoldi, A. "Confounding factors in water EMG recordings: an approach to a definitive standard". Medical & Biological Engineering & Computing, v. 44, p. 348-351, 2006.

line200

Goncalves, C.A.; Cardoso, I.P.; Carvalho, T.C.; Freire, V.D. "Action Potential Propagation/Propagate do Potential". (CD-ROM, Simulation Software)/Publisher: Editorial Universidade de Brasilia, DF, Brazil, 2005.

line200
Nogueira, E.L.; Correia, A.C.; Silva, R.A.; Goncalves, C.A.; Mota, Y.L. "Electromyographic assessment of the effect of age over paravertebral muscles/O efeito do envelhecimento nos mœsculos paravertebrais lombares investigado pela eletromiografia". In XI Congresso Brasileiro de Biomecica, 2005, Joo Pessoa. Anais do XI Congresso Brasileiro de Biomecnica, 2005. v. I. p. 1-4.

line200
Cardoso, I.P.; Oliveira, C.C.S.; Freire, V.D.; Goncalves, C.A. "A methodology for building hypermedia courses in Physiology/Uma metodologia para desenvolver aulas em formato hiperm’dia para o ensino de fisiologia". In FeSBE2004-XIX Reunio da Federao de Sociedades de Biologia Experimental, 2004, çguas de Linda, Brazil, 2004.

line200
Brasil-Neto, J.P.; Goncalves, C.A.; Lima, R.R.F.; Pessoa, V.F. "Development of a Computer-Based Method System for Studying Human Stereopsis: Contribution to the Study of Human Speed of Detection of Visual Depth". IEEE Computer Society, v. 1, p. 134-138, 1997.