Research

Participatory Modeling of Complex Urban Infrastructure Systems

The fitness and function of infrastructure in urban areas (particularly infrastructure for water, energy, and transportation) is critically important for the survival, sustainability, resilience, and success of cities. But because infrastructure systems generally are viewed as independent from each other, we often fail to recognize the strengths, weaknesses, opportunities, and threats of the interactions and interrelations between systems. This balkanization is compounded by cities' histories of centralized infrastructure creation and control that has led to fewer, but bigger, disconnected systems that have proven to be susceptible to failure, and may be unsustainable moving forward. The central hypothesis of this project is that interconnected and decentralized infrastructure systems are more resilient than isolated and centralized infrastructure by increasing response diversity. A secondary hypothesis is that decentralized infrastructure systems are more scalable and adaptable to change. The means to assess these hypotheses, however, are not readily available. While metrics and models exist to evaluate the subsystems, there is no way to consider their performance and function working together as a whole and in the context of social, behavioral, and economic decision making (SBEDM). This project will create that capability and then use it to develop the necessary comprehensive understanding of the resilience of centralized versus decentralized water, energy, and transportation (WET) systems at the metropolitan city and community level using Atlanta, GA as a test bed. 

There are 4 main research elements in this project. First, it will develop a systems dynamics model for the WET infrastructure, and the model will be used to assess how the system responds and adapts to exogenous and endogenous stressors for two alternate urban growth scenarios. The systems dynamics model will integrate the challenges and impacts of technology implementation with SBEDM. Second, a model will be developed to quantify the resilience of the WET infrastructures. The model will adopt an ecological engineering conceptualization of resilience and engage a demographically representative cohort of stakeholders in the process. The resilience of the proposed WET infrastructure will then be benchmarked to the expected range of climate change stressors, and compared to different measures of resilience as defined by theories of complex engineered and ecological systems. Third, a set of tools will be developed to simulate decision-making in the SBEDM environment. An agent-based simulation tool will capture the level of service impacts within the stakeholders, and a Bayesian network model will examine infrastructure investment decision-making. This tool will interact with the systems dynamics model to inform the stakeholders about the system performance under stressors and convey the decision back to the systems model providing directives for future development/ rehabilitation. The level of service impacts will be determined with the agent based model and the investment decisions will be conducted with a probabilistic approach. Lastly, a Pareto optimality model will be created of resilience and sustainability in WET infrastructure to assess the effect of climate change stressors like extreme heat events, droughts and floods on the water-energy-transportation nexus. This research represents a new system-of-systems approach to engineering the resilience of critical urban infrastructures in the context of their physical and socio-economic environments. It will develop fundamental theories about interdependent infrastructure systems based on a complex systems engineering approach as well as from the study of ecological systems. The insights developed here will be useful in creating tools and methods for designing and evaluating the resilience of complex urban infrastructure systems, examining the value of engineering vs. ecological approaches, pioneering methods to bridge the gap between social decision making and urban design, and contributing to the creation of a national research agenda for integrating urban resilience and sustainability into urban planning by identifying necessary data and methodologies. Advanced course modules and curriculum materials (undergraduate and graduate) related to complex engineering systems will be developed, promoted, and widely distributed through the Center for Sustainable Engineering. Results and methods will be integrated into existing programs at Georgia Tech, including the Center for Education Integrating Science, Math, and Computing that specifically target scientific interactions with African American high school students. The team will engage stakeholders from the private and public sectors to: 1) develop model structures and secure data for model implementation and validation by industrial partners; 2) explore the full breadth of decision space thus making the analyses and models relevant to stakeholders; 3) increase model transparency and stimulate group learning through interactive model development including a participatory game; and 4) make science and engineering results more accessible (e.g., through visualization).

Faculty: John Crittenden (PI), Jennifer Clark, Richard Fujimoto, Marc Weissburg, Baabak Ashuri

Sponsor: National Science Foundation


From Learning, Analytics, and Materials to Entrepreneurship and Leadership (FLAMEL)

This Integrative Graduate Education and Research Traineeship (IGERT) award prepares Ph.D. students at the Georgia Institute of Technology with the tools to improve the efficiency, design, and manufacturing of new materials. While providing students with interdisciplinary training in computing, mathematics, and material sciences, the program promotes entrepreneurial approaches to developing new materials into products for the global market. The goal of this training program is to produce scientists and engineers who will develop and use data analytics, modeling, and simulation methods to advance the fields of materials design and manufacturing. Trainees will quantify the microstructures that comprise materials and will develop algorithms and software to represent and transform these microstructures. Additionally, trainees will explore application challenges such as creating light-weight materials for energy-efficient vehicles and accelerating the development of new technologies such as additive manufacturing. Through internships in industry and a curriculum that combines business with materials science, mathematics and computing, trainees learn how to apply their research to commercial product development. By increasing the efficiency of materials development and manufacturing, this program contributes to America’s competitiveness in an increasingly globalized marketplace and contributes to the creation of new and sustainable materials and technologies. Moreover, this traineeship program helps students develop the skills and tools needed to pursue careers at the intersection of materials science, mathematics and computing.

Faculty: Richard Fujimoto (PI), Terry Blum, Surya Kalidindi, Wendy Newstetter, Hongyuan Zha

Sponsor: National Science Foundation


Dynamic Data Driven Application Systems (DDDAS)

This project addresses several interrelated areas in DDDAS including the realization of predictive frameworks to manage systems using online data-driven distributed simulations, power-aware distribution of data, utilization of crowd sourced data, and advanced image processing algorithms. These technologies offer the possibility of realizing dynamic data driven application systems (DDDAS) with much greater capability and reduced expense compared to systems using traditional command-and-control hierarchies. The project is focusing specifically on use of DDDAS in transportation systems, e.g., to track vehicles in complex urban environments. Prototype implementations of these elements are being created, integrated, and evaluated using an experimental testbed that has been deployed in the midtown area of Atlanta, Georgia.

Faculty: Richard Fujimoto (PI), Michael Hunter, Haesun Park

Sponsor: Air Force Office of Scientific Research


Embedded Distributed Simulations for Transportation System Management

Unexpected events ranging from traffic accidents on freeways to major catastrophes such as hurrucanes require sophisticated system reconfiguration and processes. The complexity of such systems mandates the exploitation of sophisticated online decision support tools to effectively manage major crises. This project addresses research challenges concerning the effective realization of autonomously reconfiguraable engineering systems enabled by cyberinfrastructure. The project explores a new approach to real-time system simulation termed ad hoc distributed simulation. Ad hoc distributed simulations feature dynamic collections of autonomous simulators interacting with each other and real-time data to in a continuously running real-time distributed simulation environment.

Faculty: Michael Hunter (PI), Richard Fujimoto, Randall Guensler, Christos Alexoppoulos, Frank Southworth

Sponsor: National Science Foundation

© richard fujimoto 2021