20th MIT-UAlbany System Dynamics Ph.D. Colloquium
Friday, April 30, 2010
MIT
Building E51, Room 095

Organizers: Joe Hsueh (MIT) and Junesoo Lee (UAlbany)

George Richardson and Joe Hsueh

From left: Nitin Joglekar, Paulo Figueiredo, Ken Parsons, Chintan Vaishnav, John Lyneis, Navid Ghaffarzadegan, Rhonda Jordon, John Sterman, Rogelio Oliva, David Andersen, Andy Whitmore, Mariana Medvetchi Dahan, Burcu Tan, Ignacio J. Martinez- Moyano

Absent in the picture: George Richardson, Joe Hsueh, Kawika Pierson, Dimitris Papanikolaou, Junesoo Lee, Katrina, HiSook Kim

From left: Joe Hsueh, and George Richardson

Time Presentation Speaker
10:00 AM Cycles In Casualty: Balancing Loops in the Insurance Industry Kawika Pierson, MIT Sloan School of Management
10:45 AM Optimal Public Warning Decisions in a Dynamic Context: Calibration for Criminal Incidents in A University Navid Ghaffarzadegan, University at Albany, SUNY
11:30 AM Using Binomial Decision Trees and Real Options Theory to Evaluate System Dynamics Models of Risky Projects Burcu Tan, University of Texas, McCombs School of Business
12:15 PM Lunch .
1:00 PM Modeling the Diffusion of Innovations in the Telecommunications Industry using System Dynamics Mariana Medvetchi Dahan, ESCP Paris Business School, France
1:45 PM Retaining Extreme Value Outcomes in an Aging Chain with Co-Flow Structure. Paulo Figueiredo, Boston University, School of Management
2:30 PM Sustainable transition; estimating marginal emissions abatement costs for the US electric power sector Ken Parsons, WPI
3:15 PM Break  
3:30 PM Evaluating Performance of Dynamically Priced Oneway Vehicle Sharing Models Using System Dynamics Dimitris Papanikolaou, MIT Media Lab
4:15 PM What Can Formal Modeling Add to Qualitative Systems Thinking? A Case of Shift the Burden Archetype Joe Hsueh, MIT Sloan School of Management
5:00 PM System Dynamics Model Documentation Tool (SDMDoc) Dr. Ignacio J. Martinez- Moyano, Argonne National Laboratory
5:15 PM Closing Remarks .

Abstracts

Cycles In Casualty: Balancing Loops in the Insurance Industry
Kawika Pierson, MIT Sloan School of Management

Many of us are familiar with how capacity acquisition delays cause cycles in the profit of some industries. There are many other balancing loops with delays that exist in the economy however, and in this presentation I will discuss a work in progress that attempts to model how these loops interact in the insurance industry causing cycles in a case where there are only insignificant delays to adjusting productive capacity.

Optimal Public Warning Decisions in a Dynamic Context: Calibration for Criminal Incidents in A University
Navid Ghaffarzadegan, University at Albany

A normative approach to security problems can give important insights into what the most proper security policies are for a given set of circumstances. We develop a simulation model that enables the determination of optimal private and public warning decisions. We use the model to examine dynamics of adversaries’ attacks on public targets and effects of public and guards readiness to react. Then we conduct a set of static and dynamic optimization procedures. Through the extended model we show that optimal solutions are sensitive to two major assumptions: sensitivity to false alarms and terrorists’ perceptions of public sensitivity. The results show that an underestimation of these effects can result in biases in optimal solutions. Furthermore, we analyze the effects of intelligence systems capabilities on optimal warning solutions under different scenarios.

Using Binomial Decision Trees and Real Options Theory to Evaluate System Dynamics Models of Risky Projects
Burcu Tan, University of Texas McCombs School of Business

Many important risky projects are characterized by stochastic processes embedded in nonlinear, feedback structures with delays. System dynamics models may be used to estimate the cash flow resulting from these projects. If these projects include managerial flexibility (real options), a correct financial evaluation of these cash flow requires the use of real options methodology. We adapt prior work on real options valuation in the decision analysis literature to develop a methodology that avoids the need to estimate a risk-adjusted discount rate for the project with options. We illustrate this approach with a model drawn from the wind power industry, which is characterized by numerous uncertainties and high managerial flexibility. We conclude with a discussion comparing this methodology to the previous methods and describe under what conditions each one might be a more appropriate choice.

Modeling the Diffusion of Innovations in the Telecommunications Industry using System Dynamics
Mariana Medvetchi Dahan, ESCP Paris Business School, France

This research work explores the factors at play in the international diffusion of innovations in the telecommunications sector, by adopting a system dynamics approach. Aspects such as countries’ socioeconomic and cultural characteristics are integrated to the analysis, along with traditional diffusion factors directly related to the product offer and to the consumer demand. We first build a model using former research on system dynamic modeling and field interviews with industry experts. We then calibrate the model on 2G and 3G mobile telephony data across a number of countries, both in the developed and developing world. Finally, we test the diffusion model on a recent innovation that has reshaped its product category: Apple’s iPhone.

Retaining Extreme Value Outcomes in an Aging Chain with Co-Flow Structure
Paulo Figueiredo, Boston University School of Management

In many situations, System Dynamics modelers have to capture attributes of items that are tracked in an aging chain. The outflow of items from the stocks in these chains usually depends on the attributes that are tracked in the co-flow. But these well known, classic models fail to account for a specific phenomenon, the screening of items. This study presents a new application of co-flows in aging chains: A co-flow that enables the process of screening, i.e. the process of either terminating or approving items depending on an attribute. An application to New Product Development (NPD) is developed. We model a two-stage product development pipeline with a co-flow structure that tracks the number of projects and the corresponding net present value of payoff. Managers at each stage must decide on capacity utilization and thresholds for minimum value of projects. Simulation results illustrate that screening can eliminate the backlog bullwhip effect in the pipeline.

Sustainable transition; estimating marginal emissions abatement costs for the US electric power sector
Ken Parsons, WPI

This presentation will highlight the author’s proposed research to estimate an emissions allowance cost schedule necessary to achieve accelerated large-scale adoption of low emissions electric power generation technologies by US Independent Power Producers to satisfy prospective US GHG regulations (H.R. 2454). Exploration of the dynamic costs of physical and financial capital as a function of the rate of technology investment required to satisfy prospective CO2 emissions reductions will be central to the research agenda. Contemporary prior art addressing a similar research agenda employ multi-sector, multi-region computable general equilibrium and or optimization models that assume seamless capital transition costs and full rational expectations. [NEMS, MARKAL, NEEMS, IPS] Additionally, these same research efforts rely on relatively short time horizons, through 2050 at most, from which to draw their respective policy insights. The author’s proposed research will reflect endogenous feedback process, realistic sub-optimal agent decision making, robust assumptions regarding technology availability / cost across a research time horizon sufficient to suitably explore plausible allowance price dynamics in context with productivity and climate goals. A principal goal for this research is to illuminate potential policy resistance as might manifest through agents’ reaction to unexpected transition costs, or ‘speed bump’, which could substantially undermine the efficacy of climate mitigation policy in the US electric power sector.

Evaluating Performance of Dynamically Priced One-way Vehicle Sharing Models Using System Dynamics
Dimitris Papanikolaou, MIT Media Lab

One-way vehicle sharing systems are decentralized urban mobility networks of vehicles and parking stations; users can pick up a vehicle from any station and return it to any other station. However, due to asymmetric demand patterns, eventually all vehicles are ending at the stations with no demand. Existing policies use trucks to redistribute vehicles, which is a complex, inefficient, and unsustainable solution. We explore a new strategy that uses price incentives to motivate users to redistribute vehicles. Similarly to a market economy, prices adjust to parking needs and user behavior. Essentially this is a self-organizing system operated by its users for its users. This work provides a methodical look on how users make decisions in dynamically priced mobility systems, under which circumstances their actions can make up a self-regulating economy, and how this economy dynamically behaves based on a given demand pattern. Is there a pricing policy that can make a one-way vehicle sharing system self-sustainable? We develop a System Dynamics framework that explains how equilibrium levels and service rates change for different demand patterns which will later be used to determine optimum pricing policy, decision making, number of parking stations, and number of vehicles for having a stable yet profitable system.

What Can Formal Modeling Add to Qualitative Systems Thinking? A Case of Shift the Burden Archetype
Joe Hsueh, MIT Sloan School of Management

While qualitative-based system archetypes have helped popularize the application of systems thinking since the publication of "The Fifth Discipline", some argue using the archetype alone, without the knowledge derived from working directly with formal models, can be dangerous. This paper presents a formal simulation model of Shift the Burden (StB), one of the most popular system archetypes. I seek to create a model as parsimonious and generic as possible while grounding the formulations based on behavioral decision processes. I conduct partial model testing and show the intended rationality of the StB dynamics. A set of leverage points are identified that prevents the system from tipping into a vicious circle where more symptomatic solution leads to erosion of fundamental solution that augments problem symptom. Counter to the common advice of eliminating StB whenever possible, I demonstrate when and why StB could be useful or even necessary under certain conditions of external shocks and internal managerial policies. The appropriate combination of Shift the Burden Sensitivity, Fundamental Solution Sensitivity and Slack Buildup as managerial policies can help dampen oscillation in fundamental solution, prevent vicious tipping dynamics and sustain high fundamental solution in face of volatile exogenous shocks. I hope to contribute to the systems thinking literature by demonstrating how formal modeling can deepen the insights of qualitative system archetypes and call for similar efforts towards other archetypes.

System Dynamics Model Documentation Tool (SDM-Doc)
Dr. Ignacio J. Martinez-Moyano, Argonne National Laboratory

The System Dynamics Model Documentation Tool (SDM-Doc) developed at Argonne National Laboratory (ANL) creates HTML-based documentation of models developed using Vensim. The model documentation created by the SDM-Doc tool allows modelers to navigate through model equations and views in a very efficient and practical way creating documentation of the model sorted by variable name, type of variable, group, view, module, or module/group/name. The tool also verifies some desirable characteristics of models helping in the model development process and creates a view summary that allows the modeler to visualize how the model is organized in the different views.


[ Back to the previous MIT-Albany colloquia ] [ Home ]

Last edited by NG 06/21/2010