Magnus Felde

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.

The 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.

Download paper
Download model (created in ExtendSim 7 so can be viewed in either 7 or 8)

N/A

Zongze Chen

A cross-dock facility enables the moving of products from a manufacturing plant and delivers them directly to their customers with little or no material handling in between. Cross-docking not only reduces material handling, but also reduces the need to store the goods in a warehouse. A typical cross-docking structure includes strip doors and stack doors. Strip doors represent the doors where full trailers are parked and unloaded. Any incoming trailer can be unloaded to any strip door. Stack doors represent the doors where empty trailers are put to collect freight for specific destinations. There are many issues with which to be concerned in the processing of goods inside a cross-dock. Based on our previous work on the assignment of facilities to locations, we are concentrating on the issues that determine the cost of processing and moving goods inside a cross-dock. We have shown in our earlier work that these costs can be minimized by appropriate assignment of receiving doors to incoming trailer trucks and stack doors to outgoing trailer trucks and have formulated this last problem as a Generalized Quadratic 3-dimensional Assignment Problem. We have also shown that the problem is also solved exactly by the simpler Generalized
Quadratic Assignment Problem.

In this project we collaborate with National Retail Systems (NRS) who provides us with realistic cross-dock situations and data to study and to evaluate our modeling and optimization result. NRS is a retail cross-docking giant in the Northeastern U.S. Their
North Bergen facility receives goods from multiple vendors, sorts and loads them onto outbound trailer trucks for a number of retail stores. We will develop a simulation model of two cross-docks in operation of their New Jersey facility - Buildings A and B. Since last November we visited NRS twice and got familiar with the daily processing in the cross-dock.

For my TCOM 899 independent study project, I developed a simulation based on ExtendSim software in order to model the processing within the cross-dock operations performed by NRS. 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 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.

Simulation nowadays is becoming a very useful tool when analyzing and testing the algorithm adapted in real situation. In this independent study period, we successfully received a research grant from ExtendSim who provide the main software for the simulation.

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.

	Download final report (664 KB) "CDAP Simulation Report" by Zongze Chen
	Download ExtendSim project files You may view these models in ExtendSim or download the ExtendSim Demo.

Institute of Industrial Engineering
Lean & Six Sigma Conference 2011

In addition to the paper's presentation, Mr. Shararah's modeling of value stream mapping concept was recognized in the conference's keynote speach by Tim Copes, the Vice President of Manufacturing and Quality for Boeing commercial airplanes due to its unique application.

Value Stream Mapping is an important tool in the implementation of lean manufacturing. It identifies the waste in the system, paving the way for a successful lean implementation. VSM is a paper and pencil tool that captures the state of the system at the state it was drawn. Simulation can be combined with value stream mapping to give it power and flexibility in order to dynamically capture the state of the system. Component-based modeling divides a simulation model into a number of smaller simulation models each encapsulated in a component resulting in a set of simulation building blocks. These blocks can be used for the purpose of developing value stream maps as they are designed to be generic, reusable, and appear exactly like the traditional VSM icons. This paper introduces the Value Stream Map Simulator using ExtendSim (VSMSx) as a powerful tool designed to facilitate the implementation of lean manufacturing by simulating the value stream map. Compared to traditional value stream mapping, this tool outputs more quantitative information about the system under study and various scenarios allowing for better decision making, thus paving the way for successful lean implementation.

Download paper

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. A 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.

Download final presentation

Purdue University has partnered with an outpatient medical facility operating 20 clinics. Their first project involves testing changes in the registration process at a primary care facility. Process flow maps have been developed to document current processes. They are modeling how reducing variability and excessive waiting early in the visit reduces variability and waiting throughout the patient visit. Their second project specifically addresses the routing and staffing of phone calls at the same clinics. Purchased call management software provides best practices for call routing. Research is focused on the cost benefit trade off of staffing levels and providing recommendations for improving the routing of phone calls.

February 1, 2008 - Unfortunately my funding as a PhD student ran out at the time I received the Extend research grant. My professors accepted my work to date for my dissertation and I defended in August of 2007. Results of the project, a manuscript which is under review at IIE transactions, is confidential until it is approved for publication.

Thank you for the assistance throughout the program.

Bob Batt

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.

Access paper's abstract *
Download model of ED Boarding
Download model of the whole hospital

* The American College of Emergency Physicians holds the copyright to this paper. It is only available through the Annals of Emergency Medicine.

November 21, 2011

URL

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.

Download complete report (in Croatian)
Download results report (in English)
Download model files

Selected Publications:
(for the entire list, please go to http://bib.irb.hr/lista-radova?autor=300571)

	Granic, Andrina; Cukusic, Maja; Tzanavari, Aimilia; Papadopoulos, George A. Employing Innovative Learning Strategies Using an E-Learning Platform. Adult Learning in the Digital Age: Perspectives on Online Technologies and Outcomes ; Kidd, Terry T., Keengwe, Jared (ed.); Hershey, PA, USA: Information Science Reference. (2010)
	?ukuši?, Maja; Alfirevi?, Nikša; Grani?, Andrina; Gara?a, Željko. e-Learning process management and the e-learning performance: Results of a European empirical study. Computers & education. (2010)
	Jadri?, Mario; Gara?a, Željko; ?ukuši?, Maja. Student Dropout Analysis with Application of Data Mining Methods. Management : journal of contemporary management issues. (2010)
	Grani?, Andrina; Mifsud, Charles; ?ukuši?, Maja. Design, Implementation and Validation of a Europe-wide Pedagogical Framework for e-Learning. Computers & education. (2009)

Aly Owida

Radio frequency identification (RFID) technology has significant impact on product tracking and identification in manufacturing systems. Most of the business cases that implement RFID technology in their operations have reduced operating costs,
such as labor and inventory. Also, RFID technology minimizes operating errors that affect efficiency of operations appearing in some key performance indicators such as cycle time, work in process, and resource utilization. In addition, several benefits such as improved item monitoring, shorter lead times, higher customer satisfaction, and better inventory control can be achieved by introducing RFID technology in different phases of production. In particular, recent developments in RFID technology and other supporting technologies have created opportunities for real-time traceability and better visibility of shop floor operations.

This paper investigates the effectiveness of the iintroduction of RFID technology in tracking and identification processes for product flow on the shop floor of a job shop manufacturing facility that produces a large number of customized furniture products in order to improve product tracking and identification. The former identification system depended on metal tags which would be replaced by radio frequency tags. Simulation was used to assess the impact of introducing the RFID technology on a number of performance measures to that manufacturing setting; such as output, throughput, cycle time, work in process, resource utilization, and average waiting time in queues. Analysis and comparison of simulation results for the base and proposed models show that RFID implementation maintains the value of most of the measures while improving the remaining measures. In addition, a cost analysis was conducted to estimate the required investment associated with RFID technology adoption, as compared to that of the current identification system, and whether the firm could return this investment or not. Results of the study shows that RFID technology can improve most of the selected performance measures of the system at the shop floor level with an acceptable cost.

Download paper
Download model files (created in ExtendSim 7 so can be viewed in either 7 or 8)

Geoffrey Ryder

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.

Download paper

Many service providers offer customers the choice of either waiting in a line, or going offline and returning at a dynamically determined future time. The best known example is the FASTPASS® system at Disneyland. To operate such a system, the service provider must first make an up front decision on how to allocate service capacity between the two lines. Then, during system operation, he must dynamically provide estimates of the waiting times at both lines to each arriving customer. The estimation of offline waiting times is complicated by the fact that some offline customers do not return for service at their appointed time.

Managing Service Systems with an Offline Waiting Option and Customer Abandonment shows that when demand is large and service is fast, for any fixed capacity allocation decision, the two-dimensional process tracking the number of customers waiting inline and offline collapses to one dimension, and characterize the one-dimensional limit process as a reflected diffusion with linear drift. Next, we use the one-dimensional limit process to develop approximations for the steady-state distribution of the number of customers waiting inline and offline, the steady-state probability of abandonment from the offline queue, and to dynamically estimate inline and offline waits for each arriving customer. We conclude by considering a cost model, and optimize the up front capacity allocation decision.

Download complete report

November 21, 2006 - In an effort to improve the customer experience in service systems, we study a queueing system in which some customers immediately join the queue and some return at a later time for immediate service. We develop methodology to estimate real-time delays and return times at which immediate service can be guaranteed. One factor complicating the estimation procedure is that some customers promised immediate service will not return. We will use ExtendSim to build a simulation study verifying the accuracy of the proposed approximations.

The development of such methodology has applications for customer contact centers and hospital emergency rooms. In the customer contact center setting, callers may be given the option between immediate service and postponed service. In the hospital emergency room setting, patients in critical condition should receive immediate service, but patients in non-critical condition can be given an accurate “appointment” time later that day. ExtendSim’s rich discrete event simulation tool set will help us to refine our methodology for these applications contexts.

February 21, 2007 - We studied a queueing system in which customers have the choice to either join the queue immediately or to come back later. At a first step, the simulation we built works under the assumption that the waiting costs for the customers are equal for both lines and depend only on the ratio of service rate between the two queues. We run the simulation for possible values of this ratio and for different probabilities of abandonment of the offline queue (the abandonment time was assumed to be Exp(gamma)). We collected the results for the mean queue length of the two lines, the probability of abandonment and the queue ratio for the n-th system in a 95% confidence found after running the simulation enough times to generate over 1,000,000 arrivals. The results then were compared with our approximations that were proved to be quite accurate. (Click on the table to download the complete Excel table.)

As a next step, we are thinking of making our model more complicated by removing the assumption of equal waiting costs and by adding another similar queue. This model will be applicable in circumstances like entertainment parks were customers want to enjoy two rides and have to choose whether they will join the standard queues or they will take a fast pass to come back later.

June 2, 2007 - Unfortunately, the last quarter we didn't make any progress in our ExtendSim model because we were working on the mathematical background of the model trying to prove the computational results. We are currently working on an extension of the model; we study a queueing system in which customers have the choice to either join the queue immediately or to come back later, but now we allow the abandonment times to be generally distributed. We will soon have the ExtendSim model for this generalization and the simulation results.

August 26, 2007 - This quarter we worked on a generalized version of our model. More specifically, we worked on a single-server queueing model in which customers may choose between waiting for service in real-time, and returning for service at a dynamically specified future time point. Customers waiting for the delayed service option may not return for service. Customers arrive at the system at a rate lambda and are served with rate mu and their service follows a generalized processor sharing discipline. More details on the model can be found in "A Single-Server Queue with a Delayed Service
Option and Exponential Abandonment Times". Now we are working with the simulation model to look at the accuracy of the approximations by the convergence results suggested in Theorems 1-3 of the paper.

February 6, 2008 - Paper submitted in the Fall of 2007 and is being revised. A minor change was made to the model so are currently rerunning some of the simulations based on the new model. Results will be incorporated in the new version of the paper.

Variability 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.

	Download final report (664 KB): "Work-In-Process Buffer Design Methodology for Scheduling Repetitive Building Projects" by Vicente González and Luis F. Alarcón
	Download Extend OR 6 model files (5.7 MB) You may view these models in ExtendSim or download the ExtendSim Demo.

April 18, 2005 - "The specific topic related to my PhD Research is buffer management in repetitive building projects. Like buffer I have been using work in process between dependent activities or production processes in building projects. The use of discrete simulation and the process interaction strategy is fundamental in my research to model the effect of different buffer sizes in the variability of production processes and the performance of projects construction (schedule and budget).For this reason, I have been working with ExtendSim and its powerful capacities to understand the effect of the buffers and to develop heuristic and analytic expressions at different levels of construction projects development to the decision making process."

February 16, 2006 – "During February, I'll be preparing an ISI paper and another one for a specific conference (International Group for Lean Construction, XIV Conference, Santiago, Chile, 2006)....so, I expect to finish these papers at the end of February or at the beginning of March, but these versions will be draft papers..the definitive paper should be ready in May or June..."

July 5, 2006 – Presented paper about WIP Buffer Design and Management at the 14th International Conference for Lean Construction, Santiago, Chile, July 2006.

December 29, 2006 - "My research is finishing. Next July (2007) I'll have my final exam about my research with the international advisors. About my models, I have developed a generic simulation modeling approach to size WIP Buffers in repetitive projects. By using Simulation-Optimization, I can find the optimum WIP Buffer size that minimizes the project schedule and maximizes the average productivity. During January and March I'll use neural networks to model the WIP Buffer size by using only a few parameters. These parameters are production rates, variability of cycle times, production amount, and numbers of processes."

January 27, 2008 - Research has been completed. It was developed using a methodology to manage uncertainty in repetitive building projects using Work-In-Process (WIP) buffers (Bf). The methodology has three hierarchical levels of application in production: strategic, tactical, and operational. At strategic level, a multiobjective analytic model (MAM) based on concepts of Pareto Front and simulation-optimization (SO) modeling was built. At the tactical level, SO modeling is directly used. At the operational level, a statistical model called Rational Commitment Model (RCM) was applied (in this level, simulation was not used). The methodology was applied in real and hypothetical cases demonstrating improvements in cost and productivity.

González, V., Alarcón, L.F. and Gazmuri, P. (2006). Design of WIP Buffers in Repetitive Projects: A Case Study. 14th International Conference for Lean Construction, Santiago, Chile, July 25th – 27th.

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

González, V., Rischmoller, L. and Alarcón, L.F. (2004). 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 5th – 8 th.

González, V., Rischmoller, L. and Alarcón, L.F. (2004). 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 1st –2nd.

González, V. y Alarcón, L.F. (2003). 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

Brandon Wolfe & Kyle Naylor

A Senior Project presented to the Faculty of the Graphic Communication Department

As the shift towards offhore business within the screen printing industry increases, American businesses must find a way to remain competitive. Due to manufacturing costs and regulation it is impossible for Americans to compete with the low per unit cost of foreign suppliers. By decreasing the time to market, American businesses will be able to remain competitive and recover lost business. In order to achieve this, a flexible scheduling model must be developed in order to increase throughput of a production process. This model serves to account for variables
and variances that occur within the screen printing process, allowing for effective continuation of production.

Download paper

July 8 to 9, 2010 - Paderborn, Germany
Application of Re-estimation in Re-planning of Software Product Releases
International Conference on Software Process (ICSP) 2010
Ahmed Al-Emran, Anas Jadallah, Elham Paikari, Dietmar Pfahl, Günther Ruh

Ahmed Al-Emran

Context -- Uncertainty is an unavoidable issue in software engineering and an important area of investigation. This paper studies the impact of uncertainty on total duration (i.e., make-span) for implementing all features in operational release planning.

Objective -- The uncertainty factors under investigation are:

	The number of new features arriving during release construction
	The estimated effort needed to implement features
	The availability of developers
	The productivity of developers

Method -- An integrated method is presented combining Monte-Carlo simulation (to model uncertainty in the operational release planning (ORP) process) with process simulation (to model the ORP process steps and their dependencies as well as an associated optimization heuristic representing an organization-specific staffing policy for make-span minimization). The method allows for evaluating the impact of uncertainty on make-span. The impact of uncertainty factors both in isolation and in combination are studied in three different pessimism levels through comparison with a baseline plan. Initial evaluation of the method is done by an explorative case study at Chartwell Technology Inc. to demonstrate its applicability and its usefulness.

Results -- The impact of uncertainty on release make-span increases – both in terms of magnitude and variance – with an increase of pessimism level as well as with an increase of the number of uncertainty factors. Among the four uncertainty factors, we found that the strongest impact stems from the number of new features arriving during release construction. We have also demonstrated that for any combination of uncertainty factors their combined (i.e., simultaneous) impact is bigger than the addition of their individual impacts.

Conclusion -- The added value of the presented method is that managers are able to study the impact of uncertainty on existing (i.e., baseline) operational release plans pro-actively.

Keywords: Operational release planning; Uncertainty; Impact analysis; Discrete-event simulation; Heuristic optimization; Explorative case study

The planning of releases is a key factor for the success or failure of incremental product development. The challenge here is deciding which features of a product are to be offered in which release (i.e., strategic planning) and how to carry out the feature development tasks within in each product release (i.e., operational planning). This planning process is extremely important since a bad release plan may cause late delivery of high-value features, unsatisfied customers, budget overrun, and decreased competitiveness. Existing research in release planning mainly takes place at strategic level. Thus, the research area is still in need of developing efficient and effective methods and tools that consider the operational feasibility of a proposed strategic release plan. The primary focus is allocating resources to feature development tasks (e.g., design, implementation, and test) with a “reasonable” release make-span. Three research objectives have been defined for the proposed research project in the area of operational release planning:

July 20, 2009 - We have recently submitted a journal article that studies the impact of uncertainty in operational release plans. We present a simulation-based method that considers planning parameters as the primary source of uncertainty. The method measures the impact of the planning parameters’ variations on total schedule time (make-span). More precisely, with our method product managers can analyze proactively the impact of both individual and combined variation of planning parameters using various levels of pessimism. We are going post the link to the article as soon it is accepted and published to the submitted journal.

February 26, 2009 - A re-planning method called H2W has been proposed to answer the "how?", "when?", and "what?" questions to re-plan of an existing product release strategy. At each of the re-planning iterations, H2W either provides a new improved plan or states that an improvement does not exist. As an instantiation of this method, an ExtendSim prototype has been developed as shown. The related research work has been published in International Conference on Software Process (ICSP) 2009.

We have further restructured the original research project during my comprehensive examination. At first, we thought we will be working on all of the parts in parallel. However, we revised our strategy and decided to work on the simulation part exclusively first and finish it completely. This project will culminate in two conference papers and one journal article which we would be happy to share with you upon their acceptance.

Vera Schmidtmann

In Germany providers of so-called "dual systems" organize the collection and recycling of waste glass under the well-known license symbol "Der Grüne Punkt" ("The Green Dot").

Providers of dual systems coordinate which quantities the glass recycling plants obtain from the locally collected glass and which quantities the glass recycling plants deliver to which glassworks. The companies have to consider the transportation and storage costs. The problem is to determine the optimal distribution of the quantities in the network.

For the implementation of the simulation model, the simulation software ExtendSim AT was selected.

Download paper

URL

Modelling of discrete operational events that impact the logistic support ramifications of unmanned aerial vehicle availability. Logistical studies of unmanned aerial vehicles (UAVs) are quite immature at this stage. This is due to the fact that most UAVs are brought into service under Advanced Technology Concept Demonstration (ATCD) programs and this method of introduction into service sees little effort put into precursory logistic planning - it is usually an after-thought. As such, while there has been a great deal of modeling of standard systems engineering logistic support, while Major Joyce has found none that he can identify in the literature pertaining to rapidly acquired aircraft, let alone unmanned aircraft systems.

March 14, 2011 - What a temultuous last twelve months!

My day job (Australian Army) has picked up to full speed and subsequently I negotiated a program amendment with UNSW@ADFA to roll back from full-time to part-time. The university candidature board (last November) has approved my project's continuance and I am now working towards the completion of Phase 1 of my doctorate studies. This being that I have completed the model and will work over coming weeks to validate it against the 2004 RAND Corp research. From here I will update the assumption based 2004 work with 2011 data and then publish my first paper.

March 30, 2010 - The project has been slightly delayed due to a stolen laptop. The project will be completed by September of 2010 and a project update will be posted in May of 2010.

October 2009 - Continuing to develop model (my work has been quite slow this month due to work commitments). I envisage having the first iteration of the model ready over the Christmas break and will commence validation of the model at this time.

September 2009 - Compiling data for use within model.

August 2009 - Overseas gathering data from a variety of sources (NASA, RAND, US Air Force, US Army, US Navy and contractors attending the UV Europe (Wales) and AUVSI North America (Washington DC) conferences).

July 2009 - Model development commencing with conceptual analysis and then beginning to map the ExtendSim blocks into their required functions.

June 22, 2009 - Unfortunately, progress in my project (UAV logistic support modelling) has been hampered over the last three months, but is getting underway. The plan from here is to develop the model over the next couple of weeks and then begin the analysis phase. I will be able to report much more in my progress report for the next quarter.

No publications; however, Major Joyce has presented the following technical papers:

	AUVS Australia February 28, 2011
	TUAV Capability Implementation Team Army Headquarters, Australia
	AIAC14 March 1, 2011
	Deconflicting the Unmanned Sky: An Analysis of UAV Deconfliction Data Australian Army, Australia
	The Australian Army and Unmanned Aircraft: A Short History Australian Army, Australia

Nana Bortsie-Aryee

Ismail Fahmi

URL

January 18, 2012 --We automated the simulation-optimization approach for the BTCC investment problem. We designed the framework such that in every time step, we can collect the data and use them for optimization, and then send the optimization result back to ExtendSim simulator. We utilized the Pause block capability of ExtenSim to pause the simulation in each time step. To Start and Resume the simulation, we used Excel VBA as our client. Our current framework was based on continuous-time = simulation and ExtendSim is used to simulate one time step only and mainly everything else is controlled in Excel VBA environment. We will investigate the feasibility of using discrete-event simulation capability of ExtendSim.

September 13, 2011 -- Our research goal is to use a simulation-based optimization (SIMOPT) approach to investigate the impact of the scheduling of investments on capital and research & development on the evolution of shifting chemical process industry from using petroleum-based sources into using biomass feedstock. The optimization framework is drawn from the analogies on graph theory and R&D pipeline management. One of the features in this framework is the incorporation of two-factor learning curve representing the relationship between the unit expansion cost and the cumulative capacity and R&D expenditure. The optimization framework is expressed in a form of mixed-integer nonlinear programming (MINLP) formulation and solved using a software called General Algebraic Modeling System (GAMS).

Our algorithm uses GAMS to perform the optimization and feed the resulting investments schedule to ExtendSim. We are to use ExtendSim to simulate the optimum result with uncertainties in the learning curve elasticities and market demand. The simulation framework that we built in ExtendSim computes the resulting total cost. We also set triggering criteria, which are whether the demand exceeds the realized capacities or whether the realized total cost exceeds the forecasted optimum total cost by a certain level of tolerance. If either of these two events occur, the simulation is halted and pertinent variables are sent back to GAMS such that the problem can be re-optimized for the remaining time periods. This loop between the simulation in ExtendSim and optimization in GAMS is performed until the end of the prescribed time period and this act results in the realization of one simulation timeline.

We will then generate many unique timelines and analyze the resulting optimum investments schedules. The details of our progress during the first three months of this research grant to use ExtendSim are as follows:

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

Fahmi, Ismail and Selen Cremaschi. 2011. "Stage-gate Representation of Feedstock Development for Chemical Process Industry" Foundations of Computer-Aided Process Operations 2012

URL

This research work is part of collaboration between the University of Brasilia (DF, Brazil) and the University of Ottawa (ON, Canada). It aims to improve the quality of teaching materials of the former university with the know-how of the latter.

The teaching of Physiology - an important segment of Biology - strongly depends on dynamic simulations and experimentations. This need is specially augmented in e-learning environments where the quality of the didactic material is a factor. Most teachers of Physiology are unable to produce high-quality simulations because they are generally non-acquainted with code-based programming. To deal with this problem, this twofold project intends to first develop some lessons using block-based programming tools and, second, in the case of lectures, scientific papers and presentations in scientific congresses - to encourage average teachers to use those tools more frequently.

A broad review of the currently available block-based programming tools pointed to ExtendSim as one of the most suitable simulation software to meet our purpose. We plan to build lessons in biology and health sciences for students of the University of Brasilia. During the first term of the project, we will produce lessons for a course entitled "Introduction to Biostatistics". The point here is to overcome the natural resistance of those students vis-a-vis biostatistics, by stressing the use of 2-D simulations. Concurrently, this phase of the work will allow us to get used to ExtendSim.

During the second term, we intend to develop more complex 3D simulations to support classic physiology lessons. The point here is to help teachers and students to better handle the high dynamic and complex features of physiological systems.

January 4, 2010 - Due to health issues project completion date extended to July 2011.

July 2, 2009 - Name of the model: Average estimate
This first simulation aims to help students to better understand the differences between three popular statistics parameters used to estimate the average of a sample. The user is initially stimulated to change the distribution of a sample and to observe the plot of the new distribution. Then he may assess the new mean, median, and mode computed by the model. This simulation is being developed using some ExtendSim blocks such as sliders (to allow changes in the sample distribution), equations (to compute the parameters), and plot (to display the shape of the sample). I suppose that this simulation should be completed in about three weeks. Therefore, for the next quarter update, I would present the screen shots of the model. The development of a new simulation should be in progress as well.

March 31, 2009 - This is my first update. I've received my license approximately one month ago and since that I have been studying the manual, especially on those subjects concerning data managing and plotting tools. I am finishing the theoretical design of my first model, which will show up how parameters like mean and median change as the user changes the sample distribution.

VENEZIANO, W. H. ; ROCHA, A. F. ; GONCALVES, C. A. ; PENA, A. J. ; CARMO, J. C. do ; 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.

GONCALVES, C. A. ; CARDOSO, I. P. ; CARVALHO, T. C ; FREIRE, V. D. . Action Potential Propagation/Propagate‹o do Potential de A‹o /(CD-ROM, Simulation Software)/Publisher: Editorial Universidade de Brasilia, DF, Brazil, 2005.

NOGUEIRA, Edna L’via ; CORREIA, Andressa da Costa ; SILVA, R. A. ; GONCALVES, C. A. ; MOTA, Yomara Lima. 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 Biomec‰nica, 2005, Jo‹o Pessoa. Anais do XI Congresso Brasileiro de Biomec‰nica, 2005. v. I. p. 1-4.

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 Reuni‹o da Federa�‹o de Sociedades de Biologia Experimental, 2004, çguas de Lind—ia, Brazil, 2004.

BRASIL-NETO, J. P. ; GONCALVES, C. A. ; GON‚ALVES, C. ; 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.

URL

Although there has been extensive research and simulation conducted within the production and supply chain domain, in respect to the United States economy, research and simulation related to the services sector lags behind. This is somewhat paradoxical, since the service sector represents over 75% of the United States labor force and approximately 75% of the United States Gross Domestic Product. This research project is unique in that it completely focuses on the services sector, and more specifically targets the leisure and recreation industry.

This research and simulation project proposes to model, in a two phased approach, the service activities associated with a major ski resort in Utah by utilizing granular RFID data from the resort’s lift access system. The resort has many years of experience with RFID lift access systems, and is a notable industry leader. The RFID data will support the development of daily skier visit forecasts, frequency distributions and transition probabilities associated with the resort’s extensive lift system.

Phase 1
The first phase of this research project is to create a discrete event simulation model, of a major ski resort, to measure and possibly optimize the effects of a set of decision variables on a specific group of metrics such as inline and offline queuing, lift capacity upgrades, and real time queue information. Currently there are five proposed decision variables each having several experimental levels. Servicescape constructs and SERVQUAL measures will be evaluated.

Objective: The main goal is to make managerial recommendations of reasonable and appropriate alternatives related to demand and capacity management of the ski resort. The simulation will support a business case for capacity management recommendations. The management will market any implemented enhancements as competitive advantages to their customer base.

Phase 2
The second phase of this research project is to construct an agent based simulation model that replicates the behavior of several skier personas with as many as 5,000 agents active during one simulated day and then aggregate the results over a full ski season. An agent based simulation makes it possible to analyze the relationship between individual micro behaviors with the aggregate patterns of many agents that eventually emerges over a day or ski season. An agent based model is applicable to skier behavior because “skiers” are intelligent, act upon their own local information, have memory, and are goal oriented.

Objective: Due to this ski resort’s unique ability to passively gather RFID data on each skier during each day, the resort has the ability to leverage the data into numerous highly targeted customer relationship campaigns. The return on investment on CRM type projects are in general difficult to quantify, so most ski resorts in North America are not fully utilizing their available CRM system. This agent based simulation will support a business case for additional marketing resources.

April 19, 2011 - I recently returned from my annual data collection trip. I was able to obtain updated inter-arrival time data as well as demand data. I have been unable to obtain demand data during previous trips, but was successful in obtaining 10 years of daily skier day demand data. This data is invaluable with respect to modeling daily demand and thus inter-arrival times for the simulation model. Work continues with refining the queue equation block to better model priority based an entity attributes. Verification and validation is still pending.

November 10, 2010 - Current work includes the addition of a new process shut down logic, variable speed travel logic, through put metrics, average cycle time metrics, sensitivity analysis and a factorial analysis of my decision variables.

One key task that I have completed is correcting a system shut down problem by adding an equation block that reads data from the convey process, with the logic of “if the convey process is empty and it is past system shut down time then shut the system down.” I discovered that I was leaving “skiers” “stranded on the lift after slope closing.” This is done by sending a “1” to the count connector on the Executive.

Future work will focus on refining the queue priority model that I am developing using the queue equation block to model complex priorities. Also, I plan to optimize the model after the model has been verified and validated on site.

July 14, 2010 - After an extensive period of data cleansing and scaling, I have been able to create interarrival schedules for each "process" in my proposed model. These "processes" of course, are the individual ski lifts at the resort. I built a database for the schedules and created a data feed into the model and ran the simulation for one day. I quickly noted that I had a major computer over utilization problem. After some model modifications to reduce CPU utilization, I still have a "hardware" problem. I have several options:

Currently, I am requesting a much more powerful computer with 8 GB of RAM for temporary loan from our computing staff.

While I am waiting during summer, I am validating more RFID data and also preparing for a quick trip back to the resort in August to obtain 2009/2010 data.

January 10, 2010 - I have recently returned from Utah were I was able to obtain complete RFID gate access data for the year 2008 - 2009. This was a major research goal accomplishment. The resort is in a growth phase and it took six months to arrange a meeting. I briefly met and talked with the president and then worked with the marketing director for almost a day downloading detailed data reports on skier days and lift rides. I will now build out the full resort model and add the interarrival times for each lift for each day of the season. The next steps are model verification and full factorial experimental model design.

August 19, 2009 - Over the last few months, I have been able to develop custom cumulative discrete distributions for skier arrival for each ski lift at the research site. A model for each lift has been created that includes balking behavior, a random batching process, entity persona attributes, a specific ride time, and an unbatching process. The current model has been constructed in a hierarchical design to force detail to lower levels of the model.

The next step is to travel to Utah and gather skier transition data between each lift, so a transition probability distribution can be created for each lift pair. This task will be by far the most time consuming and probably the most interesting, and in the end the most valuable to creating a valid ski resort discrete event model.

King, Michael A., Arnette, Andrew N. “Integrating Geographic Information Systems Into Business School Curriculums: An Initial Example.” The Design Science Institute Journal of Innovative Education, The Decision Science Institute, forthcoming September, 2011.

King, Michael A. “A Realistic Data Warehouse Project: An Integration of Microsoft Access® and Microsoft Excel® Advanced Features and Skills” Journal of Information Technology Education, The Informing Science Institute, Vol. 8, 2009

King, Michael A. “The Implementation of DSpace at the Intercontinental Hotels Group: A Knowledge Management Project Success.” Journal of Cases on Information Technology, The Idea Group, 10(2).

Conference Papers:
King, Michael A., “A Strategic Assessment of The Higher Education Industry: Applying the Porter’s Five Forces for Industry Analysis.” Proceedings of the annual Southeast Conference of the Decision Sciences Institute, Charleston, SC, February 2009.

King, Michael A., Zobel, Christopher. “Applying the R4 Framework of Resilience: Information Technology Disaster Risk Management at Northrop Grumman,” Proceedings of the annual Southeast Conference of the Decision Sciences Institute, Orlando, FL, February 2008.

Arnette, Andrew N., King, Michael A., “The Business Value of Geographical Information Systems In Business School Curriculum.” Proceedings of the annual Southeast Conference of the Decision Sciences Institute, Orlando, FL, February 2008.

King, Michael A., et. al. “Information Technology Project Retrospectives: The Northrop Grumman Sperry Marine Story.” Proceedings of the annual Southeast Conference of the Decision Sciences Institute, Savannah, GA, February 2007.

King, Michael A. “The Evaluation and Implementation of DSpace at the Intercontinental Hotels Group: A Project Portfolio Management Approach.” INFORMS Annual Conference, Pittsburg, PA, November 2006.

King, Michael A. “Is the Value of a Ph.D. Always Negative? A Case Study Comparing Real Option and Stochastic Discounted Cash Flow Valuation Techniques.” Decision Science Institute Annual Conference, November 2005.

URL

The project will be the design and operation of integrated sites under uncertainty and disruption.

We will consider both existing and newly integrated networks of production processes.

	For existing networks, we will examine modifications to either produce new products or increase current production capacity.
	For new networks, we will examine the overall design to meet reliability and capacity targets.

We will consider three major types of uncertainties:

Conceptually this is a very difficult problem because the uncertainties in the demands are continuous, typically represented by probability distribution functions, while the uncertainties in the availability and supplies are discrete in nature and correspond to major disruptions. To address this problem we intend to first investigate a probability measure that incorporates both the discrete and continuous uncertainties based on some of our previous work (Straub and Grossmann, 1990). We also intend to investigate the use of semi-analytical models for inventories in supply chains under disruptions (Snyder and Shen, 2006). We then intend to incorporate these elements in a multiperiod MILP model for optimizing the design and operation in the face of the three types of uncertainties. While our initial goal is to address design problems using somewhat more simplified models (eg linear mass balances for processes, neglecting changeovers in multiproduct plants), we intend to develop in a second phase more detailed models for operations that can be validated with plant data and used for day-to-day decision making of operating sites.

February 14, 2011 - As an update this quarter, I would like to upload a manuscript that we are also submitting to a scientific journal for publication. I cannot upload this paper at present since it is undergoing a review process required by Dow before it is released to the public.

I appreciate your patience.

November 16, 2010 - During the last quarter we completed a manuscript that describes the use of discrete rate simulation as part of the strategy to optimally design integrated sites. The release of this manuscript for external publication is being evaluated by Dow Chemical. If the paper is released and positively reviewed, we will be able to share it in future updates.

The basic idea of the manuscript is to use an optimization strategy called two-stage stochastic programming to obtain the optimal design of an integrated site considering random process failures. The technique consists in selecting failure scenarios and determining the integrated design that results in the best average performance over all scenarios. The formulation requires computing probability of occurrence and distribution for residence time for each scenario. In order to obtain this information we use the discrete rate model of the system including all the candidate production units (know as the superstructure in the optimal process synthesis field). More specifically, we run the model for 10 years of simulation time and record all process failures and their corresponding repair times. We use this information to construct the failure scenarios for the two-stage stochastic program. The manuscript includes the verification of the performance of the optimal design by repeated simulation runs, as it has been described in previous updates.

In the following quarter we expect to work on improvements on the manuscript, as well as spend some time discussing the possibility of applying the technique (and thus, discrete rate simulation) to the optimal design of discrete and batch manufacturing processes.

August 13, 2010 - During the last quarter I was doing an internship and I was not able to make any progress on the project.

May 27, 2010 - During the last quarter, we continued working on combining specialized tools for mathematical optimization and discrete rate simulation in ExtendSim for the design of integrated manufacturing sites. In previous updates, we explained how simulations were run after using equation-oriented, mathematical programming software to obtain the optimal configuration of an integrated site. Recall that the model used in the equation-oriented software is an approximation of the model used in ExtendSim. We exploit state-of-the-art optimization technology for initially reducing the search space for the design of integrated sites. The idea of running the simulation after the optimization is to validate and fine-tune the design obtained in the optimization step using an approximate model.  

In the past months, we developed an idea to take advantage of ExtendSim in a new way. Namely, we used discrete event modeling tools embedded within the discrete rate model of an integrated site to construct samples of random events. The ExtendSim setup for representing the occurrence of random failures in the processing units of an integrated site was developed at the Dow Chemical Company and it’s summarized below. 

Unplanned failures are generated randomly from time-between-failures (TBF) and time-to-repair (TTR) distributions. Each failure mode is represented by a single item. For a unit with multiple trains (and same failure parameters), the item is duplicated so that there is one item per failure mode per train. A Queue Equation block is used to correctly model multiple failures. If a failure is generated when the unit is already down, that failure is held back until the unit is back in operation. If the unit is experiencing a rate cut, then the new failure is allowed to proceed, and the minimum rate of all active failures is used to determine the rate for that unit. 

The modeling technique just described was used to generate, for each simulation run, a list of exact times when failures occurred and the time they took to be repaired. This list was used to define failure scenarios in the equation-oriented optimization model. In this way we were able to obtain a closer approximation of the detailed model in ExtendSim and the equation-oriented mathematical programming model that we use for getting an initial design. 

We have begun to write a journal paper that will illustrate the way we use ExtendSim as part of our research on integrated site design.

February 2, 2010 - Last quarter I had to prepare for some conferences and wrote and corrected a research paper that we are about to submit to a Chemical Engineering Journal. Hence, I do not have any significant progress to report to you for the last quarter.

November 6, 2009 - In our last update we described how we used discrete rate models to validate the results obtained in the design optimization step. As a result of this procedure we obtained Pareto curves of minimum cost and maximum robustness against uncertainties. In that update we reported that the results predicted by the optimization model were in reasonable agreement with the simulation. However, we considered that the match between the mixed-integer optimization model and the discrete rate simulation could be improved. The results of our efforts during the last quarter show a better agreement between the mathematical programming model and the discrete rate simulation. For illustrating our results we used the following system: an integrated site made up of 9 plants that serve two external customers modeled as product sinks. We assume there is a given demand rate for the two products represented as the maximum total flow in the discrete event simulation. For confidentiality reasons we do not provide any more details on the process.

We developed a new mathematical programming model that corresponds to this integrated sites, with the objective of maximizing the probability of delivering the products at the demanded rate consistently. The optimization chooses the production units, their capacities, and the size of their storage tanks. One of the key ideas this last quarter was to make the objective function of this optimization similar to that of the network flow model embedded in the discrete rate simulation. The results of the optimal design predicted by the mixed-integer model are in very good agreement with the performance obtained by simulating the design using the discrete rate model.

Another key idea implemented during the last quarter was to add operational logic to the way that intermediate storage is managed. The optimal design of the integrated site installs extra plant capacity to keep the intermediate storage tanks from emptying. However operating at full plant capacity all the time disrupts the flow balance in the network. To avoid this effect, we added a valve between each plant and its storage tank. The valve allows the use of the full plant capacity only when the inventory falls below a set-point. The average inventory predicted by the mixed-integer optimization model is in very good agreement with the simulation results and the standard deviation is in the same order of magnitude.
Next quarter we will work on how to use ExtendSim to make improvements on the design obtained by mixed-integer optimization.

August 4, 2009 - The research project “Design and operation of integrated sites under uncertainty and disruption” represents a collaborative effort between Carnegie Mellon University and The Dow Chemical Company. The research problem can be stated as follows. We are given a set of finished products for which there is a demand that can be deterministic, or given by a specified probability distribution. The raw materials and the intermediate products involved in the integrated production site are known. The supply of raw material can also be deterministic, or given by a specified probability distribution. There are a predetermined number of steps involved in the transformation of raw materials to intermediate and finished products. Each of these steps is carried out by a different plant that can have multiple production units. Plants are connected through the flow of intermediate products between them. The network formed by the plants and their connections represents an integrated site. The plants in the integrated site are subject to random failures that result in corrective maintenance. As a result, plants experience some amount of down time during which production is decreased or stopped all together. Since plants are interconnected in the integrated site, a failure event will propagate downstream or upstream, forcing some other plants to decrease or stop their production. If these events are not considered while designing the integrated site, there is a risk that product demand will not be consistently met. The objective of this work is to develop a systematic method to determine the trade-off between capital investment and process robustness in the design of an integrated site subject to discrete and continuous uncertainties. The main design decisions considered are increases in process capacity, addition of parallel units, and addition of intermediate storage.

Discrete rate models are frequently used for simulating the design of integrated sites at Dow. At Carnegie Mellon we are proposing to leverage these modeling techniques as a part of an optimization & simulation approach for the design of such process networks. The optimization step is performed using rigorous mathematical programming techniques. However, in order to build an optimization model of practical size we are forced to make a set of approximations to the detailed behavior or the integrated site. Our goal is to use ExtendSim discrete rate modeling capabilities to correct and improve the initial design provided by the optimization formulation. For instance, information provided by the optimization technique, such as sensitivity around the optimal solution, could be used to design a set of simulation experiments in ExtendSim.

So far we have used ExtendSim discrete rate simulator to compare the actual performance of different designs obtained in the optimization step. As a result of this procedure we obtain Pareto curves of minimum cost and maximum robustness against uncertainties. We are working on developing a systematic procedure for using the full capabilities of ExtendSim, not only to validate, but also to correct and improve the optimal design obtained through the mathematical programming formulation

n/a

URL

Uncertainty is an inherent part of production systems. In construction processes, production variability emerges as one of the most typical representations of uncertainty. The negative impacts of variability over construction processes demands effective solutions to mitigate its effects on the accomplishment of projects. One of the tools to deal with variability in construction processes is the incorporation of buffers, which can help reducing the negative impacts of variability on projects. Despite the fact that buffering strategies have been developed for production environment in construction, there is no evidence of specific applications of these strategies to highway projects. Therefore, this study presents a new approach of buffering strategies applied to transportation construction projects.

Specifically, this research searches to study and analyze different types of buffers in order to figure out which are more relevant in road projects. Then, the buffers, which were chosen in the first part of this research, will be modeled with ExtendSim and, subsequently, validated through a case of study.

Full details on Mr. Forcael's proposed research can be downloaded here

n/a

The Conceptual Model - The model has been built based on the activities which are present in a transportation construction project. A schematic to represent the simulation model used for n linear sequential activities or processes was developed.

The Proposed Model - Based on the previously-mentioned conceptual model, ExtendSim was used to create a model to represent the most significant transportation construction activities or processes. The proposed model has three activities as an example (the full model includes about eight activities).

Capacity buffer and material buffer are simply modeled as queues which store the number of workers, machines and material available. However, the calculation of WIP is not directly obtained; therefore, it is necessary to calculate it using an alternate algorithm.

Preliminary Conclusions - This current research is now focused on testing the model. The next step will be to validate the model using three case studies, which outputs will be including within the next progress report.

Download March 31, 2010 progress report

Frederick Abiprabowo

URL

October 18, 2011 -- In this third phase of research, Mr. Abiprabowo spells out step-by-step explanations of the new changes in "Create Truck and Items" H-Block, which now enables the random creation of items according to a set of constraints.

Previously, items were created according to a schedule that had to be manually typed into ExtendSim. This was not deemed efficient, since for a large amount of data, the process would be very tedious and time-consuming. Moreover, having a pre-determined schedule would not add to the randomness of the simulation.

Download October 18, 2011 report
View Cross-Dock model

In phase 3, items are now randomly created. When running the whole cross-dock simulation model a number of times, a wide array of outputs can be observed. These outputs can then serve as one of the measures of how well the cross-dock model performs as a whole.

July 6, 2011 -- Basic cross-dock operations have been modeled. More complicated operations will be implemented as the project progresses.

This progress report contains a step-by-step explanation of the current model. Diagrams of each process are also included for clarity.

Download July 6, 2011 report

At the end of this report, Mr. Abiprabowo proposes some ideas for the future model.

March 30, 2011 -- The current model tries to simulate National Retail System's (NRS) "Building A" cross-docking operations in North Bergen, NJ. The size of the current model is 16x8x8, while the actual cross-dock is much bigger. The reason for the smaller size is that we are trying to have it run with the correct operations and smaller runtime first; when all the operations are correctly laid out, enlarging the model is a much less complicated problem. In addition, the model is driven by specific volume and distance matrices. The data contained in the model is artificially generated to test the model and does not represent actual NRS operations.

In words, I will describe the logic and constraints of our current model step by step:

The model keeps track of when the items arrive at any point in the cross-dock. This measurement might be useful for calculating the amount of time needed for an item to go through the process or estimating the amount of resources, such as forklifts or workers, needed to operate in the cross-dock at any given time (with help of an external program). In the future, when priority and due time are introduced, the model can also try to prioritize items using sophisticated considerations in order to minimize the risk of sending late items.

N/A

James Woodward Saunders

URL

jim.saunder@pixie.co.za

February 23, 2011 - For the last year I have been working on the math related to solving over and under-determined matricies which are reflections of the ecology of the Knysna Catchment.... When I reach a point of acquiring a distinct process which I can demonstrate credibly I will prepare an article for publication in a peer reviewed Journal. A summary of the article will be forwarded to you after it is accepted for publication.

William John Rockett

URL

(10) Wadhwa L, Mitra S, Rockett J, Donnelly P. Systems analysis of patient flow and theatre utilisation. In: Donnelly P, Wadhwa L, editors. Access to surgery. Brisbane: University of Queensland, 1997: 63-81

Marek Lubicz

URL

August 30, 2011 - Due to organizational reasons and problems with data collection activities the project completion date has been extended to 21 May, 2012. Time schedule for modelling activities had been corrected accordingly, which in particular means that ExtendSim modelling work has been postponed until the last quarter of 2011.

June 2, 2011 - The bibliometric and data collection parts of the project are well under way; the publications are in preparation, in particular for the 37th Annual Conference of the EURO Working Group Operational Research Applied to Health Services (Cardiff University, 24-29 July).

According to the strategic assumptions of the research project, the simulation modelling part of the project had been started in the end of May with technical experiments to investigate requirements and premises for modelling a large regional system (population 3 million objects).

URL

This research is applicable to model the difficulties of a high volume
manufacturing environment with research findings focused on flexible work
cells. This consists of high volume high mix manufacturing with strict
process and quality requirements meeting international standards. This
requires modeling production for a mix of over 72 products from simple to
complex which impact cost and delivery rate. In detail the multivariates
include configuration of sub-components, products, testing, training, and
quality processes. Specifically this includes modeling the processes of:

Specifically this includes modeling the processes of:
	Demand forecasting of products and sub-assemblies.
	Product components and vendor logistics such as MRP.
	Subassembly: Build rate   Automated test rate   Items Inventory   Value
	Final Product: Build rate   Automated test rate   Items Inventory   Value   Labeling   Shipping
	Product tracking (barcode/RFID + database) through multiple processes including work in progress, automated testing and shipping.
	3D Modeling of specialized equipment & training required for production, quality and testing.

November 4, 2010 - On track for completion of model and results of PCB parallel PCB testing end of 2010.

July 10, 2010 - Continuing development of test model for capacity planning and statistical analysis. Continuing changes in model based on results of PCB parallel PCB testing - completion is still on target for 2010.

March 8, 2010 - Continuing to develop test equipment model for capacity planning and statistical analysis. New Changes to the model for implementation of parallel PCB test are in implementation. The dual test model is requiring much more effort than originally anticipated, however completion is targeted for 2010.

July 23, 2009 - Continuing to develop test equipment model for capacity planning and statistical analysis. Delays continue with introduction of the new product line.

February 27, 2009 - Continuing to create a model for evaluation of test equipment, which has been delayed with the introduction of new product line. This will be used to determine factory capacity planning & statistical analysis for dual product test system.

February 20, 2008 - Project scope has changed slightly. Since I cannot use most of my previous employer's information/ process, I now plan to create a model for the multivariate's of products and process of my new employer. Plan to incorporate ExtendSim's 3D environment to evaluate specialized test equipment to determine factory capacity planning & statistical analysis. One specific case is our unique new test station concept where a single operator must use one PC with two LCD displays and manually operate two test adapters connected on this dual product test system.

November 15, 2007 - Continuing to model processes waiting on next Kaizen event to map external processes of quotation, purchasing, and engineering.

July 11, 2007 - Continuing to model manufacturing process based on Value Stream Mapping effort (Kaizen) to validate and improve competitive manufacturing methodologies with simulation.

March 29, 2007 - Research along with work related activities identified Kaizen for (continuous improvement) using operational excellence tools such as LEAN Value Stream Mapping to develop the framework for the ExtendSim simulation model. The goal is to model all processes for identification of improvement while eliminating waste and variability. A Kaizen event was completed to map all internal processes from receipt of raw materials through delivery for a specific product and customer (due to time constraints the external processes of quotation, purchasing, and engineering were excluded for a separate event). This LEAN exercise resulted in development of the current state and desired future state maps that provide the flow with batch and queue process for the internal simulation model that is currently under development.

December 29, 2006 - Research for previous work performed in this field is being performed using electronic databases to locate texts, journals, white papers and other publications. This will be used as a guide to compare, contrast or expand on the use of modeling to improve and create competitive manufacturing methodologies with simulation.

The research will help develop the framework for the simulation models. Working through the User's Guide tutorial to gain improved knowledge and skill of the ExtendSim tool. After completion will also review the creation of blocks as this may be necessary to create the final manufacturing solution libraries.

URL

“Determining Capacity in a Healthcare Organization with Changing Client Mix,” a presentation for the Decision Sciences Institute 2009 Annual Meeting.

May 9, 2011 - Research continues at the vocational rehabilitation center. Recently the center underwent a major audit, from which changes were made that will have a significant impact on the current model. The changes that were made will have an impact in the services that the center delivers as well as the overall utilization of the centers resources. Like many other facilities that are supported by state and federal dollars the rehabilitation center has experienced a change in resource allocation due to the changes in the economy. As a result of the changing environment within the rehabilitation center, and the changing economy, administrators of the facility would like to utilize the simulation model once updated to reflect the current environment within the center to test additional operational scenarios. With the help of the centers administrators and employees we will continue to utilize ExtendSim OR to assist the center in improving service delivery, and helping their client to achieve a higher quality of life and standard of living.

September 24, 2010 - Research at the rehabilitation center continues to progress as expected. The simulation model was presented during a poster session at the POMS Health Care Mini Conference in Vancouver Canada. Since that time we have further refined the model to show the utilization of resources and capacity limitations within the center. Work continues to further define and evaluate the test factors for the simulation of the center.

The ExtendSim OR package has allowed us to build a model that employees and executives can understand. We will continue to use ExtendSim to complete the project, and as the platform of choice for future simulation research.

March 10, 2010 -Since the inception of the research grant I have developed the preliminary model of the rehabilitation center, and converted the model from the initial platform to ExtendSim OR. This conversion will allow us to collect additional data from the model, and can graphically depict the initial data and results. Together with members of the rehabilitation center staff, I am working to refine the model and conclude the initial phase of the research project by the end of April 2010. The initial phase includes developing the model of the facilities processes, including client scheduling, and revising the key performance measures. Research from this phase of the project will be presented at the POMS College of Healthcare Operations Management mini-conference in Vancouver 2010 Annual Meeting. 

URL

I am a professor at UNEFA University in Venezuela, at the Department of Post Graduate studies in Public Policy advising several post graduate students in the development of a model to analyze public resources allocation.

	Brief description of the problem/issues
	National oil companies (NOCs) face multiple challenges associated with a rapidly changing business marketplace, as well as challenges related to the development of local content. Plans developed for the development of local content often compete with the managerial urgency to complete projects on time to achieve the desired financial performance. Power point based plans to develop local content fail frequently, among other reasons, because of the lack of integration of such plans with the NOCs supply chain performance, and the lack of understanding of nonlinearities, delays, and capacity building issues related to the NOCs operations, priorities, and plans. A system dynamics based simulation model that links local content development strategy to the NOCs structure may help the definition of strategies to enhance the development of local services and manufacturing industries, the integration of local content in the core business, and consequently, help increase the impact of local companies on regional development, which is the ultimate objective of the related public policy.
	Question/s that will be answered by your research.
	Whether prioritization of capacity building projects over time, in combination with a clear understanding of the role of delays in decision making within NOCs, drives the difference between desired and real local content development, significantly.
	How research strategy will be designed to answer these questions.
	Having already designed the general structure model for a NOC (Exploration, Production, Storage, Transportation, Terminals, Refining, Commerce, Supply, and Finance) and having identified some of the entry points of local content providers in the supply chain, I am ready to pursue the development of a simulation model to accomplish the described task.
	How ExtendSim will be utilized in this project.
	Having tried other simulation programs extensively (Powersim, Vensim, I Think), and after having examined ExtendSim specifications, I find this simulation software appropriate to tackle the complexity associated with public policy formulation using simulation engines, which is the combination of the dynamic, discrete, and agents simulation paradigms, all present in ExtendSim; last, but not least, the user interface approach managed by ExtendSim increases the communication between non-technical policy makers and model designers by using animation.
	Project uniqueness
	Regarding the representation of the structure of NOCs using system dynamics, I have recently completed an effort for a major NOC; while on the side of local content policy, I designed the strategic plan for a NOCs local content support in 2008. Both efforts have given me an overview of the pros and cons of different strategies associated with the subject. I acknowledge, though, the following contributions as unique in the field: Dynamic Planning and Control Methodology: Monseo Park and Feniosky Pena Executing major projects through contractors using: Nicholas McKeena Different case studies on local content development strategy in Europe.
	Why the interest in the indicated subject?
	My interest is focused on this subject because I am a professor at the Public Policy Post Graduate School of the Armed Forces University in Venezuela, where I teach different subjects in Public Policy to Master's students.
	What will be learned from your research?
	The identification of policy levers to help improve the integration of local content companies in the core oil and gas business related to NOCs.
	What specific contribution would you like to give to the existing current state of knowledge in your subject area?
	The design of a formal policy framework for the development of local services and manufacturing industries related to National Oil Companies.

Full details on Mr. Chique's proposed research can be downloaded here

November 13, 2009 - Sending the quarterly update of research results before the end of this month as I am just wrapping up, checking the model, and writing the paper. I am sure that you will like it and recognize it as an innovative project using extendsim. Again thank you so much for your support.

n/a

Krishna Raghavan

URL

Gupta, C.L., Raghavan, K. and Daniel, B., 1987. Computer simulation of solar heated Biogas Plant - comparison of five solar heating techniques using thermal network analysis. Solar Energy Unit, Sri Aurobindo Ashram, Pondicherry, India.

Raghavan, K., 1992. End-user survey of Wind pumps and other pumping systems in South India (mechanical windpumps, diesel pumpsets, electric pumpsets, SPV pumps) for DGIS, the development co-operation department of the Netherlands govt. (1992)

Raghavan, K., and Kishore, V.V.N., 2000. Action Plan for 100% Renewable Energy supply on the Lakshadweep Islands. Working Paper. The Energy and Resources Institute, New Delhi.

Raghavan, K., 2003. 100% Renewable Energy Islands in Tuvalu, Fiji and Tonga - a Case Study. United Nations, Department of Small Island Developing States (SIDS).

Binger, A., Raghavan, K. and Ronneberg, E., 2005. Development of a Biofuels Industry in Fiji Ð a Cabinet Strategy Paper. Govt. of the Fiji Islands, United Nations Development Program (UNDP) Capacity 2015 and the South Pacific Applied Geoscience Commission (SOPAC).

Binger, A. and Raghavan, K., 2006. Strategy for the Development of an Agro-Energy Program for 16 Island States in the Caribbean Region. Inter-American Institute for Co-operation in Agriculture (IICA), Costa Rica.

Doug Lundin

URL for a complete abstract of this project

Authored by Doug
2018 which now is Raytheon Proprietary

Dinu Corbu

URL

As representative for supporting this piece of research, there can be mentioned the article “Product warranty: modelling with 2D-renewal process” by Dinu Corbu, Stefanka Chukova, and Jason O’Sullivan, published in International Journal of Reliability and Safety 2008 - Vol. 2, No.3 pp. 209 - 220.