Objective - To develop computer-based tools that facilitate exploration of decision alternatives for community resilience and that specifically address key challenges related to the scale and interdependencies of the system and the complexities of the planning process.
What is the new technical idea? Communities are places that function under the jurisdiction of geographical boundaries and a governance structure, such as a town, city, or county. All communities have social institutions to meet the needs of individuals and households, including family, economic, government, health, education, community service, religious, cultural, and media organizations. These social institutions rely in many ways on the built environment to function. NIST is developing community-centric guidance and tools to enhance the resilience of communities to hazard events. The approach recognizes the role that buildings and infrastructure systems play in supporting essential community functions over an extended time that includes disruptive events.
In this project, a range of methods will be applied to model the resilience system and to support community resilience decision-making. Data collected during field studies for past hazard events will be used to both inform and validate the models. The models and methods will be incorporated into software for greater accessibility by planners. These efforts will be coordinated with the NIST-funded Center for Risk-Based Community Resilience Planning (CoE), led by Colorado State University. The CoE simulation models and research findings will inform the models and tools developed in this project.
What is the research Plan? A first-generation integrated systems-level model of physical, social, and economic systems at the community scale will be developed to support resilience analysis and decision-making. In the development of plans, the formidable challenges posed by both the analysis and design of the resilience system must be addressed. The analysis must account for or include: a) stochastic phenomena (e.g., timing & severity of hazards, component failure); b) dependencies between and among the resilience subsytems; c) time aspects of resilience (e.g., preparedness through long-term recovery, aging infrastructure); d) connection of the built environment to the social services they support; d) significant uncertainties in knowledge (e.g., in recovery); e) metrics to quantify community resilience; f) large amount and diversity of data needed to characterize a community’s resilience system. The main challenge confronting design, or planning, for community resilience is the large scale of the resilience system. Decisions across the community ideally ought to be considered jointly given the system dependencies.
To address these significant challenges, the research plan adopts an operations research (OR) approach to the problem of community resilience planning. OR is a quantitative approach that is uniquely suited to large scale problems. A main tool of operations research is mathematical programming in which the problem is formulated; the formulation is called a math program. A search algorithm is then applied to solve the math program. The interactive formulation and solution of these models effectively provides a search capability to help decision-makers identify a set of cost-effective alternatives, in this case resilience-improving, that can be the basis of the solutions ultimately adopted.
The plan calls for gradually increasing the complexity of the math programming models as the project progresses. Early formulations will deal with only a subset of the resilience system and focus on a single hazard (e.g., floods, levee, single-family homes, population displacement, water system). Several math programming techniques will be demonstrated with the smaller system: multi-objective programming to explore tradeoffs; modeling to generate alternatives for more efficient exploration of the decision space. From here, the formulation will be extended to account for a broader resilience system (e.g., transportation, power, business interruption) and more realism (e.g., probabilistic failure, use of fragilities).
Software development is a key part of the research plan. The software will be compatible with the NIST Planning Guide and Economic Decision Guide. The software development will parallel that of the math programming model. The development of a software prototype will be centered on the initial models and methods. The prototype will fulfill two main purposes. First, it will support deliberations at NIST on the design goals of the tool. Input from a limited set of stakeholders from CoE testbeds and from communities using the guide will be solicited. In addition, the software prototype will inform a resilience roadmap workshop planned for FY19.
The software prototype will also serve as a basis for communicating with contractors on a fuller, more robust implementation. It is anticipated that the software architecture of the prototype will remain largely unchanged in the more robust implementation. It will include mathematical programming components, including a math program (prototype: written in the mathematical programming language AMPL) and a solver (prototype; FICO XPRESS), databases (prototype: MS Access), process models (prototype: MS Excel spreadsheets) and user interfaces (prototype: MS Excel spreadsheets). In the more robust implementation, it is anticipated that there will be, for example, connections to data systems (e.g., GIS) and, conditional on successful testing, open source software.
Certain collaborations are key to success of this project. A key challenge in this project is to make reasonable assumptions within the limitations of the mathematical programming framework, assumptions that capture the essential features of the resilient community systems while acknowledging data and other limitations. This project will rely in part on insights drawn from the work of the CoE. In addition, collaborations with the CoE on their testbeds (e.g., Galveston, Shelby MSA) and field studies (e.g., Lumberton) are key to informing and testing the models and software. Finally, the CoE simulation models can play a role in evaluating solutions found with the mathematical programming-based software tool.
In addition, as improving resilience is the driver of the math programming models, it is critical that the resilience metrics that are applied have a sound basis. For this reason, there is close collaboration with the Community Assessment Methodology project (7316004) that is focused on developing community resilience metrics and a corresponding assessment methodology. One of the tasks each year is to develop resilience metrics that work within the constraints of the math programming framework.
The timeline of the project reflects a gradual development of the decision support tool. FY17 focuses on assessing the validity of the math programming approach (“proof-of-concept”). FY18 will focus on the development of the software prototype. In FY19 and beyond, the tool will be further developed and tested, engaging an active stakeholder community for input on the tool and for validation purposes. Initial stakeholder involvement will be limited to the groups the Community Resilience team currently works with, which includes users of the Community Resilience Planning Guide and communities involved in the CoE testbed. The Community Resilience Panel is targeted as a venue for further stakeholder involvement.