Development of a First-Generation Community-Resilience Systems Model

Summary

Community resilience is a complex, multi-dimensional, multi-disciplinary problem with the goal of improving the way communities prepare for, resist, respond to, and recover from disruptive events, whether those events are due to natural or human-caused hazards. This project will develop systems-based models to analyze the response and recovery of physical, social, and economic systems to disruptive hazard events at the community scale, considering the dependencies that exist among buildings and infrastructure systems and the social and economic systems that they support.  The overarching model being developed in this project, NIST ARC (Alternatives for Resilient Communities) will support decision-makers with interactive tools for identifying cost-effective, resilience-improving alternatives to inform community resilience planning.  Where there are opportunities to do so, ancillary tools will also be developed that have application to specific parts of the community resilience system (e.g., transportation resilience investment prioritization).

Description

Objective - By 2028, develop and test an efficient, interactive computer-based tool, Alternatives for Resilient Communities (ARC), to facilitate exploration of decision alternatives at the community scale for infrastructure systems (e.g., buildings, water, power, transportation) and their impacts on the community functions they enable (e.g.s., employment, public safety, health, educational outcomes) for a range of natural hazard events.

What is the new technical idea?  

All communities have social institutions to meet the needs of individuals and households, among these family, the economy, 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 guidance and tools for communities–defined here as places that function under the jurisdiction of geographical boundaries and a governance structure such as a town, city, or county–to increase their resilience to hazard events to minimize the costs, economic and otherwise, borne by the community and state and federal government.  The approach recognizes the role that buildings and infrastructure systems play in supporting essential community functions over an extended time that includes disruptive events.  The goal of ARC is to support community resilience decision-making by addressing a range of problematic complexities.  Data collected during field studies for past hazard events will be used to both inform and validate ARC.

The ARC software is being developed to be complementary to the IN-CORE computational environment developed by the NIST-funded Center for Risk-Based Community Resilience Planning (CoE).  The developers of ARC regularly meet with CoE researchers involved in developing the optimization algorithms for IN-CORE.  ARC’s generally places more emphasis on interactivity,  whereas IN-CORE’s places more emphasis on higher-fidelity modeling.  ARC solutions can be tested with the higher-fidelity models of IN-CORE, and can also be used to inform the starting solution of IN-CORE heuristic search algorithms for optimization.
 

What is the research Plan? 

An integrated systems-level model will be developed to support resilience analysis, planning, and decision-making.  Modeling challenges include the need to address the many relevant systems (“system of systems”), the complexities of the planning process, and the political, budgetary, and other concerns of the community that constrain options. In the development of community resilience plans, there are additional formidable modeling challenges:  a) stochastic phenomena (e.g., timing & severity of hazards, component failures); b) dependencies between and among the physical, social, and economic subsystems impacting resilience; c) system dynamics (e.g., aging infrastructure); d) need for a mathematical description that connects the built environment to the community functions they support; d) significant uncertainties in knowledge (e.g., in drivers of recovery time); e) metrics to quantify community resilience; and f) the large amount and diversity of data required to characterize a community to support analysis of its resilience.

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 primary tool of operations research, mathematical programming, is applied here.  Analysts develop formulations, or “math programs”, that describe with mathematical expressions the specific problem requiring solution.  An optimization modeling environment (OME) assists the analyst in writing math programs and handles communications with softwares (“solvers”) that can solve the math program, returning solutions to the OME.

The modeling process to support decision-making is iterative, and this is the reason for ARC to be designed for maximum interactivity. For each solution found, decision-makers and stakeholders typically find attractive and unattractive elements. Their concerns can be folded into a modified math program through new constraints or objectives being added. This interactive, iterative modification of the problem provides decision-makers the ability to identify cost-effective, resilience-improving alternatives, that also address other community objectives and constraints.

As the project progresses, the complexity and breadth of the math programming models will increase.  Early formulations will focus on a single hazard and address a few selected systems and (e.g., floods, levee, single-family homes, population displacement, water system).  The formulation will then be extended to address new systems (e.g., transportation, power, business interruption) and new hazards (e.g., seismic, wildfire, tornado, multi-hazards).  Also, general math programming techniques will be applied, including multi-objective programming to explore tradeoffs, and modeling to generate alternatives to explore the flexibility in meeting objectives.

This line of research intersects with the research of the other projects in the NIST Community Resilience (CR) Program, including the Community Assessment Methodology, and the Economic Decision Guide projects.  The determination of the math program objectives, which include resilience and cost, will be informed by the metrics and economics research.  Further, the solutions found with ARC will be evaluated from a resilience metrics and economic perspective.  For this reason, ARC is being developed in close coordination with these projects.

The annual software advancements will involve making user enhancements (e.g., data visualization, automated trade-off analysis), testing new math programs with ultimate incorporation into the publicly available version of ARC.  Input from stakeholders, including from the CoE testbeds and from communities using the CR guide will be solicited.  In addition, the software will be developed to interact with other CR tools, including NIST-funded tools such as the EDGe$ online tool, TraCR, and CoE’s IN-CORE.

Summary of Progress to Date and Next Steps:

The timeline of the project reflects the growth in the capabilities of the decision support tool.  In FY17, FY18, and FY19 the focus was verification of the math programming optimization approach (“proof-of-concept”), development of a MATLAB-based software prototype, and extension of the underlying math programs. In FY20, the model was ported to a much more flexible, powerful computational environment (Jupyter Hub, multi-core solver licenses) for use by NIST researchers, and in FY21 a publicly released beta version of ARC was available for download at https://www.nist.gov/services-resources/software/nist-arc-nist-alternatives-resilient-communities-tool.

ARC has since been continually extended. In FY22, the underlying math programs were extended to include additional parts of the community resilience system (e.g., distributed infrastructure networks and service areas) and to make initial steps toward coordination with the AEO-developed EDGe$ software.  In FY23, ARC was extended beyond the initial riverine flooding hazard to seismic hazard. Also in FY23, fully open-source or free-for-noncommercial-use solvers (e.g., HIGHS, SCIP, COIN-OR CBC) were tested, representing a free option for communities.  In FY24, ARC was adapted and demonstrated for tornado and wildfire hazards, and tools to better address risk in resilience decision-making, and the model was tested on publicly accessible cloud platforms for ease of use.

In FY25, resource-permitting, the hazards to which ARC can be applied will extend with demonstration to communities impacted by coastal flooding and by multi-hazards. In addition, insights will be gained from an agent-based model developed to better understand the impacts of sea-level rise to communities.