Resilient Infrastructure & Smart Engineering

Video on this page is an AI-generated visualization, not a record of a real event.

Research

Infrastructure Resilience

Every storm tests the systems a city runs on. This work studies how infrastructure withstands hazards, how failures spread between interdependent systems, and how communities plan and manage recovery.

Open the theme →

Infrastructure Management

Transportation systems and the construction industry are shaped by forces no agency controls: population, the economy, the environment and technology. This work turns their signals into composite indices that show what is changing, how fast, and which factors are behind each shift.

Open the theme →

Sustainable Built Environment

Building, demolition and disasters leave large amounts of material behind, and how much of it is recovered depends on local rules, markets and practice. This work studies the policies, practices and systems that decide where that material goes.

Open the theme →

Intelligent Construction

AI is moving into construction management, from estimating and scheduling to contracts and specifications. This work studies what makes construction professionals ready to adopt AI, and develops language-model tools for construction documents.

Open the theme →

Photographs on this page are AI-generated visualizations, not records of real events.

Publications

2026

2026
A Circular Disaster Management Framework for Resilient Cities
Parsaee, M., Abazari, T., Nickdoost, N., Goodarzi, M., Shahhosseini, M., Haghighat, F. · Sustainable Cities and Society, 138, 107189
View at Elsevier BibTeX RIS
2026
A Composite Index Framework for Quantifying External Influences on Built Environment Planning and Decision Making
Nickdoost, N., Movahedi, M., Sun, Y., Smith, D., Choi, J. · Journal of Management in Engineering (ASCE), 42(6), 04026052
Read the study View at ASCE BibTeX RIS
2026
A Dynamic Carbon Accounting Framework for On-Site Versus Off-Site Sorting in Construction and Demolition Waste Management
Konda, C., Nickdoost, N. · Construction Research Congress 2026 (ASCE), 408–418
Read the study View at ASCE BibTeX RIS
2026
Modeling AI Readiness and Adoption Intentions in Construction Management: A Technology Acceptance Model Approach
Nickdoost, N., Kim, J., Park, S., Song, K. · Proceedings of the Associated Schools of Construction 62nd Annual International Conference (EPiC Series in Built Environment, vol. 7), 284–293
Read the study View at EasyChair BibTeX RIS
2026
Modeling Interdependencies and Cascading Failures in Critical Infrastructure Using a Multi-System Simulation Framework
Nickdoost, N. · Construction Research Congress 2026 (ASCE), 397–407
Read the study View at ASCE BibTeX RIS

2024

2024
A Composite Index Framework for Data-Driven Decision-Making in the Construction Industry
Nickdoost, N., Choi, J. · Construction Research Congress 2024 (ASCE), 546–556
Read the study View at ASCE BibTeX RIS
2024
A Composite Index Framework for Quantitative Resilience Assessment of Road Infrastructure Systems
Nickdoost, N., Shooshtari, M. J., Choi, J., Smith, D., AbdelRazig, Y. · Transportation Research Part D: Transport and Environment, 131, 104180
Read the study View at Elsevier BibTeX RIS
2024
Automated Material Separation Activity Identification for Sustainable Demolition Operation
Jalloul, H., Alshami, A., Nickdoost, N., Choi, J., Moon, J. · Construction Research Congress 2024 (ASCE), 981–990
View at ASCE BibTeX RIS
2024
Identification and Prioritization of Multidimensional Resilience Factors for Incorporation in Coastal State Transportation Infrastructure Planning
Nickdoost, N., Jalloul, H., Choi, J., Smith, D. · Natural Hazards, 120, 1603–1663
Read the study View at Springer BibTeX RIS
2024
Optimization of a Temporary Debris Management Site on a Two-Way Roadway
Nickdoost, N., Alshami, A., Choi, J., Merrick, D. · Proceedings of the Canadian Society of Civil Engineering Annual Conference 2023 (Lecture Notes in Civil Engineering, Springer)
Read the study View at Springer BibTeX RIS

2023

2023
A Framework to Determine the Optimal Locations of Temporary Debris Management Sites in Response to a Hurricane Event
Nickdoost, N., Choi, J. · Canadian Society of Civil Engineering Annual Conference 2021 (Lecture Notes in Civil Engineering, Springer), 287–299
Read the study View at Springer BibTeX RIS

2022

2022
A Project Life-Cycle Approach to Managing Procrastination in Construction Projects: State-of-the-Art Review
Nickdoost, N., Choi, J., AbdelRazig, Y., Sobanjo, J. · Journal of Construction Engineering and Management (ASCE), 148(5), 03122001
Editor's Choice Reviewed Paper Honorable Mention, 2022
View at ASCE BibTeX RIS
2022
A System-of-Systems Framework of Data Analytics to Support Strategic Decision-Making in the Construction Industry
Nickdoost, N., Choi, J., Abdelrazig, Y. · Construction Research Congress 2022 (ASCE), 412–421
Read the study View at ASCE BibTeX RIS
2022
An Integrated Framework for Temporary Disaster Debris Management Sites Selection and Debris Collection Logistics Planning Using Geographic Information Systems and Agent-Based Modeling
Nickdoost, N., Jalloul, H., Choi, J. · International Journal of Disaster Risk Reduction, 80, 103215
Read the study View at Elsevier BibTeX RIS
2022
Challenges and Adaptive Measures for U.S. Municipal Solid Waste Management Systems During the COVID-19 Pandemic
Pinto, A. D., Jalloul, H., Nickdoost, N., Sanusi, F., Choi, J., Abichou, T. · Sustainability (MDPI), 14(8), 4834
View at MDPI BibTeX RIS

2021

2021
Evaluating Correlations of External Factors and Performance Measures of the Multimodal Transportation System in Florida
Sun, Y., Choi, J., Nickdoost, N., Kirtonia, S., VanDenBogaert, J. · Transportation Research Record: Journal of the Transportation Research Board, 2675(9), 1089–1105
View at SAGE BibTeX RIS

2018

2018
A New Method for Nonlinear Analysis of Base Isolated Structures
Moharrami, H., Nikdoost, N. · 16th European Conference on Earthquake Engineering, Thessaloniki, Greece
BibTeX RIS

Teaching

Fall 2026

BCN 4591CUG

Mechanical, Electrical and Plumbing Systems

The principles of mechanical, electrical and plumbing systems in buildings.

Spring 2025
Fall 2025
Spring 2026
Fall 2026

Fall 2026

BCN 6763Grad

Constructability

Integrating construction knowledge into planning, design and execution. The course was developed in full for the department and includes group presentations.

Fall 2024
Fall 2025
Fall 2026

Other courses

BCN 5036Grad
Research Methods
How research in construction management is designed and carried out: framing a question, reviewing the literature, choosing qualitative and quantitative methods, collecting and analyzing data, and reporting the findings.
- Spring 2025
- Spring 2026
BCN 4870CUG
Heavy Civil Construction
Methods, equipment and planning for heavy civil projects such as highways, bridges, earthwork and utilities, from equipment selection and productivity to cost.
- Summer 2025
BCN 6315Grad
Advanced Construction Technology
The latest technologies transforming construction. The course was developed in full for the department.
- Fall 2024

Gallery

Site visit to the Student Success Center under construction

Tour of the mechanical plant at UNF Physical Facilities

Mechanical, Electrical and Plumbing Systems lab session

Advanced Construction Technology class, Fall 2024

About

Navid Nickdoost, Ph.D., PMP

Assistant Professor, Construction Management Department, College of Computing, Engineering and Construction, University of North Florida, Jacksonville.

Assistant Professor of Construction Management at the University of North Florida since 2024, after a Ph.D. in civil engineering at Florida State University and seven years on construction projects as a project engineer, structural inspector and planner. The research, carried out in the RISE lab (Resilient Infrastructure & Smart Engineering), studies how infrastructure and the construction industry respond to the forces acting on them, from storms and economic shifts to material waste and new technology.

Research interests

Infrastructure resilienceHow infrastructure systems withstand hazards, how failures spread between interdependent systems, and how communities plan and manage recovery.
Infrastructure managementData-driven methods that help agencies plan, prioritize and invest in transportation and construction systems as the conditions around them change.
Sustainable built environmentReducing the waste and environmental footprint of construction, from material recovery and recycling to the policies that shape them.
Intelligent constructionHow artificial intelligence and data can change construction work, and what it takes for the industry to adopt them.

Experience

2024 – now
Assistant Professor
Construction Management Department, University of North Florida
2024
Postdoctoral Researcher
FAMU-FSU College of Engineering, June to August 2024
2012 – 2019
Project Engineer, Structural Inspector, Planner and PMO Member
Elmes Construction Company, Tehran

Education

2024
Ph.D., Civil Engineering
Florida State University
2021
M.S., Civil Engineering
Florida State University
2016
M.S., Civil Engineering
Tarbiat Modares University
2013
B.S., Civil Engineering
K. N. Toosi University of Technology

Infrastructure Resilience

Every storm tests the systems a city runs on. Roads flood, power fails and takes water and hospitals with it, and recovery can take months. This work studies how infrastructure systems withstand natural hazards, how failures spread between interdependent systems, and how communities plan and manage the recovery that follows.

Questions this work answers

AI-generated image: Flooded neighborhood at dusk with some blocks dark and a hospital lit in the distance

How do failures cascade between interdependent infrastructure systems?

Run a flood and watch a failure in one system pass to power, water, transport and health, then recover as repair crews work around blocked roads.

Open tool →

AI-generated image: Coastal highway bridge over open water at dusk with car light trails

How resilient are coastal roads to storms?

Score a road system's robustness and recovery speed, change the factor weights, and trace a weak score to its causes.

Open tool →

AI-generated image: Temporary debris site with long piles of storm debris and trucks on haul roads at dusk

How should debris sites and hauling be planned after a hurricane?

Follow the study from siting rules to choosing sites and sizing the truck fleet, shown on Liberty County roads.

Open tool →

AI-generated image: Storm debris stacked in a road median with a loader, grinder and a line of dump trucks

How should a roadside debris site be run?

Set loaders, grinders and the grinding schedule for a site in a road median and see operating cost and truck waiting time.

Open tool →

AI-generated image: Coastal city at dusk under an approaching storm, with some districts dark

PULSE: Perishable Utility & Lifeline Signals During Extremes

A monitor view of counties under watch during a storm and the lifeline status being tracked.

Example dataOpen view →

Publications in this theme

2026
Modeling Interdependencies and Cascading Failures in Critical Infrastructure Using a Multi-System Simulation Framework
Nickdoost, N. · Construction Research Congress 2026, 397–407
Read the study View at ASCE BibTeX RIS
2024
A Composite Index Framework for Quantitative Resilience Assessment of Road Infrastructure Systems
Nickdoost, N., Shooshtari, M. J., Choi, J., Smith, D., AbdelRazig, Y. · Transportation Research Part D, 131, 104180
Read the study View at Elsevier BibTeX RIS
2024
Identification and Prioritization of Multidimensional Resilience Factors for Incorporation in Coastal State Transportation Infrastructure Planning
Nickdoost, N., Jalloul, H., Choi, J., Smith, D. · Natural Hazards, 120, 1603–1663
Read the study View at Springer BibTeX RIS
2024
Optimization of a Temporary Debris Management Site on a Two-Way Roadway
Nickdoost, N., Alshami, A., Choi, J., Merrick, D. · CSCE Annual Conference 2023, Lecture Notes in Civil Engineering
Read the study View at Springer BibTeX RIS
2022
An Integrated Framework for Temporary Disaster Debris Management Sites Selection and Debris Collection Logistics Planning Using Geographic Information Systems and Agent-Based Modeling
Nickdoost, N., Jalloul, H., Choi, J. · International Journal of Disaster Risk Reduction, 80, 103215
Read the study View at Elsevier BibTeX RIS
2022
A Framework to Determine the Optimal Locations of Temporary Debris Management Sites in Response to a Hurricane Event
Nickdoost, N., Choi, J. · CSCE Annual Conference 2021, Lecture Notes in Civil Engineering, 287–299
Read the study View at Springer BibTeX RIS

Photographs and video on this page are AI-generated visualizations, not records of real events.

Intelligent Construction

AI is moving into construction management, from estimating and scheduling to contracts and specifications. Whether it helps depends on the people who use it and on tools built for construction's own documents and decisions. This work studies what makes construction professionals ready to adopt AI, and develops language-model tools for construction documents.

Questions this work answers

AI-generated image: construction project team reviewing a schedule on a large screen in a site office at dusk

What makes a construction professional ready to use AI?

Rate how aware you are of AI and how useful, easy and appealing it seems, and see the intention to adopt that a survey of 54 professionals predicts.

Open tool →

Publications in this theme

2026
Modeling AI Readiness and Adoption Intentions in Construction Management: A Technology Acceptance Model Approach
Nickdoost, N., Kim, J., Park, S., Song, K. · Proceedings of the ASC 62nd Annual International Conference, EPiC Series in Built Environment 7, 284–293
Read the study View at EasyChair BibTeX RIS

Photographs and video on this page are AI-generated visualizations, not records of real events.

Infrastructure Management

Transportation systems and the construction industry are shaped by forces no agency controls: population, the economy, the environment and technology. Their signals sit in many datasets, reported at different times and scales. This work turns them into composite indices that show what is changing, how fast, and which factors are behind each shift, so planners can respond before small changes add up.

Questions this work answers

AI-generated image: Florida highway interchange, airport and seaport at dusk

What is moving Florida's transportation system?

Follow how external conditions push on each mode, from cars to seaports, and trace any shift back to the factors behind it.

Open tool →

AI-generated image: busy construction sites across a growing city at dusk

How is the construction industry changing?

Watch one index of the many outside factors that move the construction industry, and open it up to see what drives each change.

Open tool →

Publications in this theme

2026
A Composite Index Framework for Quantifying External Influences on Built Environment Planning and Decision Making
Nickdoost, N., Movahedi, M., Sun, Y., Smith, D., Choi, J. · Journal of Management in Engineering, 42(6), 04026052
Read the study View at ASCE BibTeX RIS
2024
A Composite Index Framework for Data-Driven Decision-Making in the Construction Industry
Nickdoost, N., Choi, J. · Construction Research Congress 2024, 546–556
Read the study View at ASCE BibTeX RIS
2022
A System-of-Systems Framework of Data Analytics to Support Strategic Decision-Making in the Construction Industry
Nickdoost, N., Choi, J., AbdelRazig, Y. · Construction Research Congress 2022, 412–421
Read the study View at ASCE BibTeX RIS
2021
Evaluating Correlations of External Factors and Performance Measures of the Multimodal Transportation System in Florida
Sun, Y., Choi, J., Nickdoost, N., Kirtonia, S., VanDenBogaert, J. · Transportation Research Record, 2675(9), 1089–1105
View at SAGE BibTeX RIS

Photographs and video on this page are AI-generated visualizations, not records of real events.

Sustainable Built Environment

Building, demolition and disasters leave large amounts of material behind, and how much of it is recovered depends on local rules, markets and practice. Florida alone generates millions of tons of construction and demolition debris every year, and recycling performance varies widely from county to county. This work studies the policies, practices and systems that decide where that material goes.

Testing recycling policy before counties adopt it

AI-generated image: construction and demolition recycling yard with sorted piles of concrete, wood and metal at dusk

C&D recycling policy simulation for Florida counties

Most Florida counties operate without a construction and demolition debris diversion ordinance. County officials who want to improve recycling have no practical way to see how a policy would perform locally before committing public resources.

This project is building a simulation platform that lets county staff test policies under their own local conditions: diversion mandates, deposit-refund systems, differential tipping fees, facility certification, procurement mandates for recycled materials, or a combination.

Funded by the Hinkley Center for Solid and Hazardous Waste Management · 2026 to 2027

Visit the project site

Publications in this theme

2026
A Circular Disaster Management Framework for Resilient Cities
Parsaee, M., Abazari, T., Nickdoost, N., Goodarzi, M., Shahhosseini, M., Haghighat, F. · Sustainable Cities and Society, 138, 107189
View at Elsevier BibTeX RIS
2026
A Dynamic Carbon Accounting Framework for On-Site Versus Off-Site Sorting in Construction and Demolition Waste Management
Konda, C., Nickdoost, N. · Construction Research Congress 2026, 408–418
Read the study View at ASCE BibTeX RIS
2024
Automated Material Separation Activity Identification for Sustainable Demolition Operation
Jalloul, H., Alshami, A., Nickdoost, N., Choi, J., Moon, J. · Construction Research Congress 2024, 981–990
View at ASCE BibTeX RIS
2022
Challenges and Adaptive Measures for U.S. Municipal Solid Waste Management Systems During the COVID-19 Pandemic
Pinto, A. D., Jalloul, H., Nickdoost, N., Sanusi, F., Choi, J., Abichou, T. · Sustainability, 14(8), 4834
View at MDPI BibTeX RIS

Photographs and video on this page are AI-generated visualizations, not records of real events.

Modeling AI Readiness and Adoption Intentions in Construction Management: A Technology Acceptance Model Approach

Navid Nickdoost, Jonghoon Kim, Soomin Park, Kwonsik Song · Proceedings of the ASC 62nd Annual International Conference, EPiC Series in Built Environment 7, 284–293 · 2026

Read the paper Try the tool

AI tools are reaching construction management, but adoption is slow and uneven. This study surveyed construction professionals to find which perceptions drive their intention to use AI. Perceived usefulness and awareness of AI were the strongest predictors; attitude and ease of use mattered less.

Complete survey responses

Variance in intention explained

56 %

Average perceived usefulness

5.48 of 7

01 / The problem

Tools are maturing faster than adoption

AI is being applied to estimating, scheduling, quality, safety and contract administration, yet its use in everyday construction management remains limited and fragmented. Barriers include limited awareness, inconsistent training and uncertainty about AI's value.

Most research develops or benchmarks algorithms. Less is known about the perceptions that decide whether professionals accept AI tools, or how those perceptions differ across construction management functions.

02 / The method

A technology acceptance survey of professionals

The study uses the Technology Acceptance Model at the individual level, with five constructs: perceived usefulness, ease of use, attitude, intention to use, and awareness of AI. Items used a seven-point scale; awareness was scored from 0 to 1. Questions also covered seven management tasks, such as scheduling, estimating and quality control.

Professionals answered through email invitations and QR codes shared in person. 54 complete responses were analyzed. Most respondents had one to five years of experience, followed by those with more than twenty, and most were project engineers, project managers or executives.

Bar charts of respondents by work experience and by job title — **Figure 1.** Participants by (a) work experience and (b) job title.

03 / Perceptions

Contracts and cost estimating lead

Respondents were generally positive. Perceived usefulness scored highest (mean 5.48), then attitude (5.08) and intention (4.99); ease of use scored lower (4.34), and awareness was moderate (0.57 on a 0 to 1 scale).

By task, contract and legal compliance and cost estimating and control drew the highest intention, tasks with structured data and repeatable decisions. Safety monitoring and quality control drew lower intention despite being seen as useful.

Task	Intention	Usefulness
Contract and legal compliance	5.30	5.31
Cost estimating and control	5.26	5.41
Supply chain and logistics	5.09	5.17
Project planning and scheduling	4.96	5.41
Risk management	4.83	5.04
Safety monitoring	4.67	4.89
Quality control and inspection	4.59	5.07

Table 1, Panel B. Task-specific intention and usefulness, mean on a 1 to 7 scale.

Values from the paper.

04 / How the perceptions relate

Usefulness is the hub

Perceived usefulness correlated most strongly with attitude (r = 0.76) and intention (r = 0.69), and attitude with intention (r = 0.61). Ease of use related moderately to usefulness and attitude but only weakly to intention (r = 0.24).

Awareness related weakly to usefulness and attitude (r = 0.16) and modestly to intention (r = 0.36), suggesting many professionals know of AI without hands-on experience of its value. Most respondents fell in the moderate-to-high range of the study's AI Readiness Index.

Histogram of the AI Readiness Index and a correlation heatmap of the constructs — **Figure 2.** (a) AI Readiness Index, (b) correlations among the constructs.

05 / The model

Usefulness and awareness predict intention

A multiple regression of intention on the four perceptions was significant (F = 15.84, p < 0.001) and explained 56% of the variance (R² = 0.564). Residual and multicollinearity checks were satisfactory.

Perceived usefulness and awareness were significant. Attitude had a positive but non-significant effect; ease of use had a small negative, non-significant one, read cautiously because its scale had lower reliability (α = 0.61).

Try the model in the tool →

Predictor	Coefficient	p
Constant	−0.518	0.562
Awareness	1.657	0.008
Perceived usefulness	0.645	0.001
Ease of use	−0.149	0.354
Attitude	0.330	0.189

Table 3. Regression of intention to use AI. Marked rows are significant at p < 0.05.

Values from the paper, rounded to three decimals.

06 / For education

Teach AI as an applied skill

Because awareness and usefulness drive intention, the paper argues construction management programs should treat AI as an applied competency inside existing courses such as estimating, scheduling, quality control and risk management, with hands-on work, case studies, simulations and industry projects.

It also suggests building students' understanding of what AI can do and their ability to judge AI-generated insights, rather than focusing on simpler interfaces.

Nickdoost, N., Kim, J., Park, S., Song, K. (2026). Modeling AI Readiness and Adoption Intentions in Construction Management: A Technology Acceptance Model Approach. Proceedings of the Associated Schools of Construction 62nd Annual International Conference, EPiC Series in Built Environment 7, 284–293.

View at EasyChair BibTeX RIS

Figures reproduced from the paper, © the authors (EPiC Series, EasyChair); tables redrawn from its printed values.

← Intelligent Construction

A System-of-Systems Framework of Data Analytics to Support Strategic Decision-Making in the Construction Industry

Navid Nickdoost, Juyeong Choi, Yassir Abdelrazig · Construction Research Congress 2022, 412–421 · 2022

Read the paper

The construction industry changes slowly as new technologies spread, and sometimes in unexpected ways. This conceptual paper models the industry as a system of systems, from individual firms and agencies up to the national industry, and proposes a composite index and a dashboard so decision-makers can follow those changes and trace them to their causes.

Levels in the industry hierarchy

Aspects of industry health

01 / The problem

Gradual change and surprises are hard to catch

Technologies such as wearables, robotics and virtual reality spread through the industry gradually, and can bring unintended effects, such as automation reducing worker engagement. Catching these changes early lets managers act in time.

Earlier studies looked at one aspect at a time (prices, labor demand, productivity) and relied on statistical models that need long, frequent data, while most industry indicators are reported quarterly or yearly over a couple of decades at most.

02 / The industry as a system of systems

Firms, industry health and the national industry

The framework follows three phases: definition, abstraction and implementation. The industry is modeled at three levels: construction entities (firms, owners, suppliers, agencies), construction health, and the national industry.

Health has five aspects: economic, stability, social, development and quality (Naderpajouh et al. 2016). A lexicon of resources, operations, policies and economics marks the system's boundaries, and external factors come from the literature in macroeconomic, demographic and environmental groups.

Diagram of the construction system of systems with entity, state health and national levels — **Figure 2.** Construction system-of-systems hierarchy.

03 / Tracing a change

An index you can follow down to its cause

A composite index mirrors the hierarchy: external factors at the base, sub-dimensions from factor analysis, then the health aspects and one overall index.

The paper illustrates its use with the 2007 to 2009 market crisis. When the top-level trend drops, following it down would point to a housing-related sub-dimension that grew in importance, and a firm might respond by depending less on the housing market, for example by moving into non-residential work.

Diagram of the composite index hierarchy and a conceptual example tracing a drop during the 2007–2009 market crisis to housing-related factors — **Figure 3.** (a) Composite index hierarchy, (b) detecting changing behavior. Conceptual illustration.

04 / A dashboard for decisions

Set the weights, compare against your own performance

A mock dashboard lets users set weights for each aspect, move through the levels of the index, and compare it with their own performance.

The paper's example: a firm weighing a safety investment gives more weight to the social aspect. If the index rises, signaling a need for more safety measures, while the firm's safety performance falls, the firm may need to invest more in safety practices.

Mock dashboard with weight sliders for economic, social and stability aspects and a chart comparing the composite index with a firm's safety performance — **Figure 4.** Mock dashboard for the composite index. Illustrative values.

Nickdoost, N., Choi, J., Abdelrazig, Y. (2022). A System-of-Systems Framework of Data Analytics to Support Strategic Decision-Making in the Construction Industry. Construction Research Congress 2022, 412–421.

View at ASCE BibTeX RIS

A conceptual framework; the index and dashboard figures are illustrations, not computed results. Figures reproduced from the paper with permission from ASCE, © ASCE. This material may be downloaded for personal use only.

← Infrastructure Management

A Composite Index Framework for Data-Driven Decision-Making in the Construction Industry

Navid Nickdoost, Juyeong Choi · Construction Research Congress 2024, 546–556 · 2024

Read the paper

Construction firms make long-term plans under conditions they do not control: the economy, housing, material prices, population. Existing industry indices measure costs, activity or confidence, but none shows their combined effect. This study builds a composite index of the US construction industry that combines many external factors and can be traced from the top score down to the factor behind a change.

External factors identified

Industry health indicators

Factors kept per health aspect

01 / The problem

Strategy rests on a few factors and intuition

Inflation, interest rates, material costs and demand all shape a construction firm's prospects, yet strategic decisions are often made on intuition and limited information. Earlier data-driven studies looked at a few factors at a time, such as their effect on construction prices or labor demand, and could not track many factors continuously.

Industry indices such as ENR cost indices, the Architecture Billings Index and confidence indices are useful, but none shows the combined effect of the external factors driving the industry's health.

02 / The framework

From external factors to one industry index

Industry health is measured on five aspects (economic, stability, social, development and quality), following Naderpajouh et al. (2016). External factors for each aspect come from the literature.

Data are brought to a quarterly time step. Cross-correlation, allowing for lead and lag, selects the factors most associated with the health indicators. Dynamic factor analysis groups them into common trends, and weighted geometric aggregation builds the aspect indices and the overall index. Decision-makers can change the weights.

Diagram of external factors around construction industry health aspects, the four-step index process, and the index hierarchy — **Figure 1.** (a) External factors and industry health, (b) the index process, (c) the index hierarchy.

03 / The US construction index

36 factors, three common trends per aspect

The literature review identified 36 external factors across demographics, housing, the economy, material prices and the environment. Nine indicators measure health on the economy, stability and safety aspects, from sources including the US Census Bureau, the Bureau of Labor Statistics, Engineering News-Record and Associated Builders and Contractors.

Cross-correlation kept the top five factors for each aspect, and a three-trend model fit better than a two-trend one for every aspect. In the economy aspect, number of structures, real estate GDP and the rent price index form one trend the paper reads as the housing market. In the safety aspect, public residential and non-residential construction spending form one trend, and the fuel and metal price indices another.

04 / Results

The 2009 crisis and the pandemic show in the index

With equal weights at every level, the composite index drops sharply around 2009, reflecting the market crisis. The stability and economy aspects drop and recover during the COVID-19 period (2019 to 2021).

Tracing these changes down the hierarchy points to the industrial production index as one of the factors behind the drop, a signal of lower demand for projects or tighter material supply that a firm could respond to by reviewing forecasts, diversifying markets or finding other suppliers.

Charts of the composite index and the stability, economy and safety aspect indices from 2005 to 2021, with 2008 to 2010 and 2019 to 2021 highlighted — **Figure 2.** Composite index aggregation results.

05 / Limits

What the study does not cover

Cross-correlation shows association, not cause; the paper suggests causal methods such as Granger causality or PCMCI for selecting factors.
The 36 factors can be extended to a wider range of socioeconomic and environmental factors.
All weights were equal in this case; decision-makers can set their own.

Nickdoost, N., Choi, J. (2024). A Composite Index Framework for Data-Driven Decision-Making in the Construction Industry. Construction Research Congress 2024, 546–556.

View at ASCE BibTeX RIS

Figures reproduced from the paper with permission from ASCE, © ASCE. This material may be downloaded for personal use only.

← Infrastructure Management

A Dynamic Carbon Accounting Framework for On-Site Versus Off-Site Sorting in Construction and Demolition Waste Management

Chandana Konda, Navid Nickdoost · Construction Research Congress 2026, 408–418 · 2026

Read the paper

A contractor tearing down a building has to decide whether to sort the debris on site or haul it mixed to a recovery facility. This study builds a parameter-based model that puts the carbon and the cost of both choices side by side for each material, and shows that no single choice wins for every material.

Materials compared

Product systems modeled

Range tested per parameter

±20 %

Illustrative residential demolition case in Florida, per ton of mixed waste.

01 / The problem

Sort on site, or haul it mixed?

Construction and demolition waste makes up nearly 30% of all waste generated worldwide. Sorting it on site keeps materials cleaner and cuts hauling, but needs more labor, space and equipment. Sending it mixed to a material recovery facility uses the facility's machines and keeps the site simple, at the cost of longer trips.

Most studies look at either the carbon or the cost of these choices, not both. Life cycle models usually rely on fixed averages for distances, energy and recovery rates, and the combined carbon and cost models that exist work at the building design or policy scale, not at the contractor's decision.

02 / The framework

Carbon and cost from the same process model

The unit of comparison is one ton of mixed waste, followed from the moment it is generated until it is recycled or landfilled. Each of seven materials gets its own on-site and off-site process, 14 product systems in all, built in openLCA 2.5.0 and scored with the IPCC 2013 100-year global warming method.

On site, diesel runs the machines; off site, electricity runs the facility and the trucks travel farther. Cost flows for diesel, electricity, labor, equipment and landfill fees are written into the same processes, so every run returns kilograms of CO₂e and dollars per ton together.

Every input is a parameter a contractor can change to match a project.

Inputs

Environmental

Diesel use
Electricity use
Recovery rates
Transport distances
Emission factors

Economic

Fuel price
Electricity rate
Labor cost
Equipment cost
Landfill tipping fee

Model development

Build the life cycle inventory
Process chains: collection, sorting, transport, output
Link background databases (ecoinvent, openLCA)

Analysis

Scenario testing: on site vs. off site
Comparative metrics: GHG per ton, efficiency, costs
Hotspots: biggest contributors to emissions and costs

Total carbon emissionsTotal costs

Decision support

Figure 1. Carbon accounting framework for construction and demolition waste sorting.

03 / The case

A hypothetical house demolition in Florida

A residential job was chosen because it is a common source of mixed debris. Residential sites usually have room to stage bins, so on-site sorting is given higher recovery rates.

Diesel and electricity prices, the landfill tipping fee and labor wages come from Florida and national sources. Off-site hauls are longer: 40 to 70 km against 10 to 15 km.

Each parameter was then moved 20% up and down, one at a time, to see which ones drive the results.

Parameter	On site	Off site	Unit
Electricity (MRF)	–	7.80	kWh/ton
Transport distance	10–15	40–70	km
Labor	21	16	$/hr
Equipment cost	5	3	$/ton
Landfill fee	67.5	67.5	$/ton
Diesel price	3.6	3.6	$/gallon
Electricity price	0.12	0.12	$/kWh
Recovery rate: concrete	0.85	0.70	fraction
Recovery rate: asphalt	0.90	0.65	fraction
Recovery rate: wood	0.80	0.70	fraction
Recovery rate: metals	0.90	0.75	fraction
Recovery rate: gypsum	0.75	0.60	fraction
Recovery rate: roof shingles	0.70	0.55	fraction
Recovery rate: plastics/glass	0.65	0.50	fraction

Table 2. Parameters used in the waste sorting model.

04 / Results

On site is cleaner; off site is often cheaper

On-site sorting stayed near 5 to 6 kg CO₂e per ton for every material. Off site, emissions climbed to about 19 to 24 kg for concrete, wood, gypsum, shingles and plastics/glass; only asphalt and metals came out slightly lower off site.

Cost ran the other way for most materials. Off site was cheaper for concrete, metals, wood, gypsum and plastics/glass, while asphalt and shingles cost less to sort on site.

Emissions were most sensitive to recovery rates, and costs to labor and landfill fees. The ranking of on site against off site held across the ±20% range.

On-site sortingOff-site sorting

Emissions by material

Cost by material

Figure 3. On-site vs. off-site emissions and cost by material.

05 / The trade-off

The best choice depends on the material

Metals favor off site on both counts: about $35 and 4.6 kg per ton, against about $90 on site.

Asphalt is a true trade-off. Off site cuts emissions from about 5.6 to 4.5 kg but raises the cost from about $36 to $68 per ton.

Concrete runs the opposite way: off site saves money (about $31 against $50) but triples emissions (about 18.8 against 6.2 kg).

Wood, gypsum and shingles are cleaner on site, near 5 kg against 20 to 22 kg, for small cost differences. Plastics and glass fare worst: about $97 on site, or about 24 kg off site.

On-site sortingOff-site sorting

Figure 5. Trade-off between cost and emissions. Lines join the two options for each material.

06 / What it shows

Weigh carbon and cost together, material by material

Choosing on cost alone can bring much higher emissions, most of all for concrete and wood. The study suggests making on-site sorting the baseline, which also supports LEED waste diversion credits, and using off-site facilities for metals and asphalt, where central recovery lowers emissions.

Because recovery rates, distances, equipment use and prices are all inputs, a contractor can rerun the comparison for a specific job, and policymakers can see where incentives would tip the balance toward low-carbon sorting.

07 / Scope

One illustrative case

The results come from a single hypothetical residential demolition in Florida, not a recorded project.
Parameter values were checked against published benchmarks and reported industry ranges, and will differ by county, market and site.
Plastics and glass performed poorly either way, which points to the need for better recycling pathways for them.

Konda, C., and Nickdoost, N. (2026). A Dynamic Carbon Accounting Framework for On-Site Versus Off-Site Sorting in Construction and Demolition Waste Management. Construction Research Congress 2026, 408–418.

View at ASCE BibTeX RIS

Text and figures from the authors' camera-ready manuscript.

← Sustainable Built Environment

Modeling Interdependencies and Cascading Failures in Critical Infrastructure Using a Multi-System Simulation Framework

Navid Nickdoost · Construction Research Congress 2026, 397–407 · 2026

Read the paper Try the tool

In a flood, power, water, roads and hospitals fail together, and each failure feeds the next. This study builds an agent-based simulation in which the hazard, the damage, the repair crews and the patients all move through time on the same road network, so recovery depends on whether crews and ambulances can get through. It is demonstrated on a hypothetical urban flood.

Infrastructure sectors linked

Drop in energy supply at onset

~50 %

Lowest road connectivity

~80 %

Results of the hypothetical flood scenario in the paper.

01 / The problem

Models treat failure and recovery as separate, static steps

Floods and hurricanes disrupt several infrastructure systems at once, and a failure in one sector spreads to others. Most models describe these links in a static or pairwise way and rarely show how disruption and recovery unfold over time.

Recovery is often assumed to happen after fixed delays, even though repair crews and ambulances must travel damaged roads. Hospitals are usually treated as simply open or closed, and transportation is rarely modeled as something that amplifies the cascade.

02 / The framework

Hazard, sectors, crews and people in one simulation

A hazard field, such as flood depth, changes over time and degrades the assets it reaches. Each asset is an agent whose function can be partly lost, so a hospital can keep running at reduced capacity and a road can stay open but slower.

Repair crews and ambulances travel a road graph whose travel times change with the flood. Population agents turn lost service into consequences: health declines faster when access to care is cut.

Failures and repairs spread through the same links, so clearing a road or restoring a substation sets off a recovery cascade, just as a failure sets off a failure cascade.

Framework diagram: hazard dynamics, infrastructure sectors, dependency mechanisms, restoration agents, resilience metrics and community impacts — **Figure 1.** Architecture of the simulation framework.

03 / The scenario

A hypothetical urban flood

A grid city with a road network, substations, pumping stations and hospitals, written in Python. Floodwater spreads fast during a shock phase, then drains. When water on a road passes a threshold, debris builds up until the road is blocked.

Pumps stop when their energy falls below 0.5. Hospitals switch to generator fuel when power drops and lose treatment capacity when power or water run short. Patients lose health over time; a death while water or energy is below target is recorded as avoidable.

The threshold values are hypothetical; the paper chose them to reflect typical operating limits in the resilience literature.

Grid city showing flood severity, open roads, roads with debris, hospitals, pumps and substations — **Figure 2.** The simulated city: flood severity, roads, hospitals, pumps and substations.

04 / Results

Recovery came in steps, as crews reached each site

Energy supply fell by nearly 50% in the first hours and recovered in steps, one per repaired substation, staying above 90% by about 70 hours. Pumps stopped when their substations failed, water supply was close to zero by hour 25, and the sector did not fully recover until about 80 hours.

Nearly all hospitals ran on generators in the first 20 hours. Over 120 patients were treated within the first 10 hours, but a long tail of delays ran past 100 hours.

Road connectivity fell to about 80% and returned by about 90 hours. For the first 80 hours, crews spent most of their time traveling rather than repairing.

Six charts: energy availability, water supply, hospitals on backup power, time to first treatment, crew time split and road connectivity over time — **Figure 3.** Simulation results: (a) energy, (b) water, (c) hospitals on backup power, (d) time to first treatment, (e) crew time split, (f) road connectivity.

05 / What it shows

Roads decide how fast a city recovers

Blocked roads delayed repair crews, stretched outages and kept patients from hospitals; clearing debris brought quick improvements across sectors. Even as floodwater receded, debris kept the network split until a few key roads were cleared.

Hospitals ran at partial capacity instead of simply open or closed, and generator backup bridged only a short time. Targeted repairs, such as restoring the substations that power pumps or clearing key roads, set off recovery across the system.

06 / Limits

From a hypothetical city to a real one

The scenario and the agent thresholds are hypothetical. The model was verified through tests of each agent and its routing, and its structure and early outputs were reviewed with domain experts; it was not calibrated to a real event.
Applying it to a real city needs data on physical links, such as which substation feeds which hospital, that is often proprietary or undocumented.
Calibrating decision thresholds needs detailed records from past disasters, which are often sparse or spread across agencies.

Nickdoost, N. (2026). Modeling Interdependencies and Cascading Failures in Critical Infrastructure Using a Multi-System Simulation Framework. Construction Research Congress 2026, 397–407.

View at ASCE BibTeX RIS

Text and figures from the author's camera-ready manuscript. Figures reproduced from the paper.

← Infrastructure Resilience

A Framework to Determine the Optimal Locations of Temporary Debris Management Sites in Response to a Hurricane Event

Navid Nickdoost, Juyeong Choi · CSCE Annual Conference 2021, Lecture Notes in Civil Engineering, 287–299 · 2023

Read the paper

After a hurricane, debris goes to temporary sites where it is sorted and stored, and where those sites sit decides how fast a county is cleared. This study picks the sites with both the map and the cleanup in view: GIS screens out unsuitable land, an agent-based simulation runs the trucks, and an optimizer chooses the sites and routes that finish the job fastest.

Debris in the scenario

160,000 CY

Sites chosen out of 45

Trucks in 10 crews

Hypothetical hurricane scenario for Liberty County, Florida.

01 / The problem

Site selection usually ignores the cleanup itself

Hurricane debris can block emergency vehicles, and hazardous and organic waste threatens health. Temporary debris management sites (TDMS) sort, store and recycle it, and their number and location control how the trucks work.

Earlier studies chose sites from geographic rules alone, such as distance from wetlands, floodplains and homes. Studies of truck routing, in turn, optimized one part of the system but gave planners no way to watch the whole operation run and spot its bottlenecks.

02 / The framework

Screen the land, simulate the trucks, search for the best mix

Suitability analysis. A GIS restriction model removes land near wetlands, industry, roads and utilities, plus farms, forests, urban land and the 100-year floodplain. The remaining sites larger than 2 hectares are ranked on slope and distance to buildings, wetlands, green land and roads.

Operation simulation. An AnyLogic agent-based model puts trucks, debris piles and sites on the real road network. Each truck loads debris until full, then drives to the closest site, unloads in 10 minutes and returns to its task list.

Optimization. The OptQuest engine tries combinations of sites and truck task lists to minimize total collection time, with penalties when hazardous debris, or debris inside communities, is delivered late.

Framework diagram: GIS analysis module, debris estimation and resources feed a simulation module and an optimization module, producing optimal sites and truck schedules — **Figure 1.** The framework for finding the optimal TDMS locations.

03 / The case

A hypothetical hurricane in Liberty County, Florida

HAZUS-MH estimated about 160,000 cubic yards of mixed debris for a major hurricane, based on the Hurricane Michael scenario. The debris was spread over 300 locations along the roads; half of it was hazardous or perishable, and half lay inside communities.

Fifty trucks of 18 cubic yards each worked in ten crews of five, and five sites had to be chosen from the 45 suitable ones. Most suitable land ranked average; a few sites in the far south ranked best, but sat far from where people live.

Map of Liberty County showing suitable site locations ranked low, average and high, debris locations, state highways and roads, with an inset of the main town — **Figure 4.** Suitable TDMS locations and debris locations.

04 / Results

Priority debris cleared in half the time

The search was run for 10,000 iterations, beyond which the objective barely improved. It placed three sites near the towns, all ranked average, and took the two best-ranked sites in the south to serve the southern roads.

Hazardous and perishable debris was collected in about half the time of the other debris, although the two volumes were the same. Debris inside communities was likewise all collected in half the time spent on debris outside them.

Map of the five chosen sites with truck crews and debris, and three charts: objective function over iterations, and cumulative debris delivered by type and by location over time — **Figure 5.** (a) Objective function over iterations, (b) the optimal TDMS locations, (c) hazardous vs. nonhazardous debris delivered over time, (d) debris inside vs. outside communities delivered over time.

05 / What it shows

The best site is not always the best-ranked one

Average-ranked land close to communities beat better-ranked land far away, because the trucks finish faster from there. Choosing sites with the cleanup in the loop gives both the site locations and a truck schedule that clears the most harmful debris first.

The animated simulation lets planners watch the operation, find its weak points and test other scenarios. The suitability model works for any county, and trucks, debris estimates and agent behavior such as capacity, speed and unloading time can all be changed.

06 / Scope

A first version of the model

The hurricane is hypothetical, and the type and location of each debris pile along the roads were generated at random.
Each site is assumed to take unlimited debris, and the work inside a site is not modeled; trucks simply spend 10 minutes unloading.
The optimizer returns near-optimal solutions, not a proven optimum.

Nickdoost, N., and Choi, J. (2023). A Framework to Determine the Optimal Locations of Temporary Debris Management Sites in Response to a Hurricane Event. CSCE Annual Conference 2021, Lecture Notes in Civil Engineering, 287–299.

View at Springer BibTeX RIS

Figures reproduced from the paper, © Canadian Society for Civil Engineering 2023.

← Infrastructure Resilience

Optimization of a Temporary Debris Management Site on a Two-Way Roadway

Navid Nickdoost, Ahmad Alshami, Juyeong Choi, David Merrick · Proceedings of the CSCE Annual Conference 2023, Lecture Notes in Civil Engineering, Springer · 2024

Read the paper Try the tool

When there is not enough open land after a hurricane, a low-traffic road is often turned into a temporary debris management site. Its narrow layout changes how trucks unload and how debris is sorted, stored and ground. This study simulates the work inside such a site and searches for the mix of loaders and grinders, and the grinding schedule, that keeps cost and truck waiting time low. The layout comes from a real site in North Port, Florida; the debris arrivals are a hypothetical hurricane scenario.

Resource cost

~60 % lower

Truck waiting time

57 % less

Site cleared after the roads

25 days

Optimized plan compared with a randomly generated plan, in the paper's hypothetical scenario.

01 / The problem

There is often not enough open land for a debris site

Disasters can produce more debris than local waste systems can handle; after Hurricane Harvey in 2017, more than 200 million cubic yards had to be removed. Temporary debris management sites sort, store and process this debris before it goes to final disposal, so how well they run affects how long recovery takes.

Environmental and legal rules often leave too little land for these sites, and a low-traffic roadway is used instead. Earlier productivity studies looked at a single site on vacant land with a flexible layout. None studied how to allocate equipment given a site's layout, the resources available and the rate debris arrives, and none offered a way to see the operations inside the site.

02 / The framework

Simulate the site, then search for the cheapest plan

Simulation. A discrete-event model in AnyLogic follows debris through the site: measuring at the scale, unloading, sorting, storing, grinding and loading out. Trucks, loaders and grinders are modeled as resources. Grinding halves the volume of vegetative debris, and for safety the site stops all other work while grinders run.

Optimization. The AnyLogic optimization engine tries different numbers of loaders (1 to 6) and grinders (1 to 2), grinding days, start hours and durations, out of 73,728 possible plans. It keeps the plan with the lowest cost of trucks, loaders and grinders, where truck cost is the time trucks spend waiting inside the site, plus a penalty when debris is not removed by a deadline.

Diagram of the framework: site layout, debris entry and resources feed a simulation module and an optimization module — **Figure 1.** The simulation-based optimization framework for allocating resources at a debris site.

03 / The case

A debris site in the median of a two-way road

The site is a narrow strip in the middle of a two-way road, taken from a real site in North Port, Florida. One area holds construction and demolition (C&D) debris, and two hold raw and ground vegetative debris. The storage area is 800 m long and 20 m wide; with 5 m piles it holds 80,000 m³. Only one truck can park to unload at a time, and trucks cannot pass a parked truck.

Debris arrivals are hypothetical. In total, 95,000 CY of vegetative debris and 6,700 CY of C&D debris arrive over 96 days, hauled by 50 trucks of 30 CY. Equipment costs $120 per hour for a truck, $48 for a loader and $60 for a grinder.

Layout of the roadway debris site: a C&D area and two vegetative debris areas along an 800 m strip with a scale tower — **Figure 5.** Two-way roadway site layout.

Chart of cumulative vegetative and C&D debris arriving at the site over time — **Figure 6.** Vegetative and C&D debris entering the site (hypothetical schedule).

04 / Results

More loaders, lower cost

The best plan found uses six loaders and two grinders, grinding seven times a week for one hour each. It clears the site within 25 days after all debris has been removed from the roads.

Compared with a randomly generated plan of two loaders and one grinder, grinding five times a week for three hours, the best plan cost about 60% less even though it uses more equipment. The saving comes from trucks: going from two to six loaders cut truck waiting time by 57%. Both plans grind for a similar time each week (14 versus 15 hours).

Four charts: objective function over iterations, and resource cost, truck waiting time and site capacity over time for the optimal and random plans — **Figure 7.** Optimization results: objective function, and cost, truck waiting time and site capacity over time for the optimal and the random plan.

05 / Scope

What the results rest on

The debris arrival schedule is hypothetical; the site layout is real.
The savings are measured against one randomly generated plan.
The paper notes that a road with little traffic can be set up as a debris site quickly because the infrastructure is already there, and that the framework can be adapted to other site layouts.

Nickdoost, N., Alshami, A., Choi, J., Merrick, D. (2024). Optimization of a temporary debris management site on a two-way roadway. Proceedings of the Canadian Society of Civil Engineering Annual Conference 2023, Lecture Notes in Civil Engineering, Springer, 357–370.

View at Springer BibTeX RIS

Text and figures from the authors' manuscript (2023). Figures reproduced from the paper.

← Infrastructure Resilience

An Integrated Framework for Temporary Disaster Debris Management Sites Selection and Debris Collection Logistics Planning Using Geographic Information Systems and Agent-Based Modeling

Navid Nickdoost, Hiba Jalloul, Juyeong Choi · International Journal of Disaster Risk Reduction, 80, 103215 · 2022

Read the paper Try the tool

After a hurricane, debris is hauled to temporary debris management sites before it is processed. Where those sites sit shapes how long collection takes. This study links site selection to the collection itself: GIS finds sites that meet the siting rules, an agent-based model simulates trucks collecting debris to those sites, and an optimizer picks the sites and truck routes that shorten collection while clearing priority debris first. It was tested on Liberty County, Florida, after Hurricane Michael.

Collection time, real-world conditions

~156 days

With the optimized sites

~112 days

Vegetative debris collected within

60 days

01 / The problem

Debris sites are usually chosen on geography alone

Disasters leave debris that can exceed a year of normal waste, block roads needed for rescue, and carry health and environmental risks as contaminated and organic material decays. Managing it takes a large share of the response budget: more than 25% of disaster response expenses, according to FEMA figures cited in the paper.

Most earlier studies pick temporary debris sites from geographic criteria and leave out how the choice affects the collection that follows. The few that address both do not give decision-makers a way to set collection priorities or to watch the operation and find its bottlenecks.

02 / The framework

Three modules: suitability, simulation, optimization

Suitability analysis. GIS removes places that break the siting rules (too close to wells, rivers, lakes, floodplains, wetlands or sensitive areas; inside residential, institutional, commercial or industrial land) and ranks the rest by slope and distance to buildings, wetlands, green land and roads, using a weighted overlay. Sites must be larger than 2 ha.

Operation simulation. An agent-based model in AnyLogic places debris, trucks and debris sites on the real road network. Trucks collect their assigned debris and haul it to a site; a site's capacity is its floor area times a 5 m pile height times a 60% storage ratio, and a full site makes trucks wait.

Optimization. The AnyLogic optimization engine (OptQuest metaheuristics) changes the site locations and each truck's task list, runs the simulation, and keeps the best plan. The objective is total collection time plus geographic terms, plus a penalty that grows when priority debris waits too long.

Diagram of the framework: GIS analysis module, simulation module and optimization module feeding optimal debris site locations and truck scheduling — **Figure 1.** The framework for identifying optimal locations for temporary debris management sites.

03 / The case

Liberty County, Florida, after Hurricane Michael

Hurricane Michael generated over 5 million cubic yards of debris. In Liberty County, 47,400 debris tickets were issued to collect about 525,700 CY of vegetative, construction and demolition, and ash debris, plus 34,000 leaners and hangers, from about 330 miles of roadway. 150 trucks with an average capacity of 50 CY did the work in about five months, using four temporary debris sites.

Debris and hauling records came from the primary debris hauling contractors; land use, roads, elevation and flood zones came from FDEP, FDOT, USGS and FEMA.

The suitability analysis found 45 locations able to host a debris site. Most rated average; a few in the south rated high.

Maps of debris source locations and debris volume along Liberty County roads — **Figure 6.** Debris source locations and volume distribution in Liberty County after Hurricane Michael.

Map of Liberty County with suitable debris site locations rated low, average and high — **Figure 7.** Suitable locations for temporary debris sites in Liberty County.

04 / Does the model match reality?

Calibrated against the actual cleanup

Truck speed, loading time and unloading time were calibrated to 25 mph, 6 minutes and 9 minutes. With these values the simulation tracked the cumulative debris actually collected over time.

Changing truck speed by 3 mph changed the total duration by 3.7%; changing truck waiting time by 1 minute changed it by 4%.

Charts comparing simulated and real cumulative debris collected, and uncertainty for truck speed and waiting time — **Figure 8.** Simulated versus real debris removal, and uncertainty in truck speed and waiting time.

05 / Results

Collection time fell from about 156 to about 112 days

The optimizer moved the four sites slightly from the ones used in practice. All four rated average for suitability; three sit closer to urban areas to speed collection there, and the fourth serves the county's southern half.

Vegetative debris, a fire hazard that also loses recycling value as it decays, and debris inside urban areas were set as priorities. Vegetative debris was collected within 60 days, about half of the ~112 days, and debris within the community within ~70 days.

Simulation snapshot with original and optimal debris sites, and curves of debris delivered over time — **Figure 9.** Optimization results: optimal site locations and debris delivered over time, overall, by debris type and by location.

06 / Planning trade-offs

More trucks pay off; more than four sites does not

The study ran 27 scenarios: 3, 4 or 5 debris sites with 50 to 210 trucks. Going from 50 to 210 trucks cut total collection time by about 60%, and the time to collect vegetative and urban debris by almost 80%.

More sites shortened the average collection time but raised total cost, mainly from operating the sites. Cost-effectiveness improved 63% from 50 to 130 trucks but only 13% from 130 to 210, and more than four sites reduced it.

Compare the 27 plans in the tool →

Line charts of collection duration against number of trucks for 3, 4 and 5 sites — **Figure 10.** Effect of trucks and sites on total duration and on collecting vegetative and urban debris.

Table and chart of normalized cost-per-benefit for the 27 scenarios — **Figure 12.** Cost-per-benefit across the 27 scenarios.

07 / Limits

What the study does not cover

All debris is assumed known and fixed at the start of recovery; in practice some is moved to the curb later.
Work inside the debris sites is simplified to a handling rate, and trucks spend a fixed time unloading, so congestion at busy sites is not modeled.
How fast debris decays, which would set collection deadlines, is not yet evaluated.
Social effects beyond recovery time and health risk, such as neighborhood disturbance and disparities, are left for future work.

Nickdoost, N., Jalloul, H., Choi, J. (2022). An integrated framework for temporary disaster debris management sites selection and debris collection logistics planning using geographic information systems and agent-based modeling. International Journal of Disaster Risk Reduction, 80, 103215.

View at Elsevier BibTeX RIS

← Infrastructure Resilience

A Composite Index Framework for Quantifying External Influences on Built Environment Planning and Decision Making

Navid Nickdoost, Mohammad Movahedi, Yanshuo Sun, Dennis Smith, Juyeong Choi · Journal of Management in Engineering, 42(6), 04026052 · 2026

Read the paper Try the tool

Population, the economy, the environment and technology shape how much is asked of a transportation system, but they sit outside any agency's control and are reported in scattered datasets. This study selects the external factors that actually move each mode's performance, combines them into a composite index built along the same levels agencies plan at (mode, system and state), and sets it beside a performance index so planners can see where performance is not keeping pace. It was applied to Florida's transportation system.

External factors with data

78 of 98

Performance measures with data

51 of 56

Factors kept per mode

01 / The problem

Outside forces are tracked, but not tied to decisions

External factors change demand, operating conditions and how fast infrastructure wears out, often through delayed and interacting paths. Small shifts can build up and appear suddenly at higher planning levels. They are hard to track because the data are spread across many sources and reported at different time and space scales.

Agency dashboards show these trends but stay descriptive. Most composite indices are built for benchmarking, do not follow how infrastructure is governed, and cannot trace a system-level change back to its drivers. Scenario planning depends on the scenarios assumed.

02 / The framework

Screen for drivers, find common trends, aggregate by level

External factors and performance measures come from literature and agency sources and are brought to a quarterly time step. For each mode, causal screening keeps the factors linked to performance: PCMCI for monthly and quarterly data, and Granger causality with cross-correlation for annual data. Ten factors are kept per mode.

Dynamic factor analysis groups the factors into a few common trends. Weights come from factor loadings and explained variance, with optional weights from decision-makers, and geometric aggregation builds the mode, system and state indices.

A parallel performance index, built on the same levels from mobility and safety measures, shows the supply side.

Figure 1. Composite index development process.

03 / Results

Two downturns: 2008 and 2020

From 2005 to 2021 the state-level index rose overall, with two clear declines, in 2008 to 2010 and 2020 to 2021. Recovery from the financial crisis was slow; the pandemic drop was sharp but rebounded within about 18 months and was back to its prepandemic level by the end of 2021. Ground, air and water indices follow the same two downturns.

Charts of the state-level composite index from 2005 to 2021 with its 95% confidence band, and the ground, air and sea transportation and six mode indices — **Figure 5.** Composite index at the state, system and mode levels.

04 / Tracing a drop

From the mode index down to the factors

Because the index is built in levels, a change at the top can be followed down to its causes. All four dimensions of the auto index fell during the pandemic: economic conditions, state employment, national employment and mobility.

Following them to the base level points to GDP, the economic condition index, employment at state and national level, state unemployment, and vehicle miles traveled, several of which dropped sharply in 2020. For aviation, the drop in tourism drove the pandemic decline, and seasonal precipitation explains its oscillating pattern.

Auto mode index from 2005 to 2021 with its four dimensions (economic conditions, state employment, national employment, mobility) and the ten external factors below them, the COVID-19 period shaded — **Figure 6.** Trend analysis for auto using its dimensions (latent variables).

05 / Demand against performance

Drivers rose faster than performance

The composite index (external pressure) and the performance index rose together from 2011 to 2019, dropped together in 2020 and recovered together afterwards, as expected since the factors were chosen for their link to performance.

Their slopes differ. From 2012 to 2019 the composite index rose faster, meaning the drivers were signaling more planning need than performance was meeting. After the 2020 drop it also rebounded faster, pointing planners to operational, behavioral and policy factors that may have slowed the performance recovery.

Line chart comparing the composite index and the performance index from 2011 to 2021 — **Figure 7.** Comparison of the performance index (PI) and composite index (CI).

06 / Checks

Stable under other choices, readable to planners

Rebuilding the index under 32 combinations of eight interpolation, two disaggregation and two normalization methods kept the 2008 to 2009 decline, the 2020 drop and the post-2020 recovery. The curves form two clusters by normalization method, which changed only the level of the index, not its path; interpolation and disaggregation made almost no difference.

Six FDOT planners and managers from the Trends and Emerging Transportation Office, the State Seaport Program and the Transit Office reviewed the framework in two virtual sessions and found the indices interpretable and relevant to their planning.

Line chart of 32 state-level composite index trajectories from 2005 to 2021, forming two clusters by normalization method with the same turning points — **Figure 8.** Robustness analysis results: 32 combinations of normalization, disaggregation and interpolation.

07 / Limits

What the study does not cover

The index is for monitoring and diagnosis, not forecasting.
It does not model interactions among parts of the transportation system itself, such as cross-modal feedbacks.
Results depend on the factors included and on the quality of public data; regions with weaker data would see more uncertainty.

Nickdoost, N., Movahedi, M., Sun, Y., Smith, D., Choi, J. (2026). A Composite Index Framework for Quantifying External Influences on Built Environment Planning and Decision Making. Journal of Management in Engineering, 42(6), 04026052.

View at ASCE BibTeX RIS Supplemental materials

← Infrastructure Management

Identification and Prioritization of Multidimensional Resilience Factors for Incorporation in Coastal State Transportation Infrastructure Planning

Navid Nickdoost, Hiba Jalloul, Juyeong Choi, Dennis Smith · Natural Hazards, 120, 1603–1663 · 2024 (online 2023)

Read the paper Try the tool

Transportation agencies cannot track every factor that shapes resilience, so they need to know which ones matter most. This study gathered resilience factors for wind and water hazards from research and state plans, asked Florida transportation experts to rate them, and ranked them into priority tiers. A workshop with practitioners then named the barriers to using these factors in planning and how to overcome them.

Factors found in research and plans

200

Consolidated into

15 factors

Expert survey responses

44 of 283

01 / The problem

Agencies have limited resources to track resilience

Coastal transportation systems often sit in low-lying areas exposed to flooding and hurricanes, and Florida has been hit by almost 40% of the hurricanes making landfall in the United States. Disruptions slow emergency services, cut access to critical amenities and hold back recovery.

Measuring resilience means tracking technical, socioeconomic and environmental factors, but agencies have finite resources. Existing studies did not say which factors to prioritize, or address practical questions such as whether measurable data exist for them.

02 / The method

Identify the factors, then let experts rank them

Identify. A systematic review selected 59 academic articles, and the plans of ten states recognized as leaders in transportation resilience planning, plus Florida, were reviewed. This produced 200 factors, consolidated into 15 measurable ones; through expert input, 13 apply to wind hazards and 10 to water hazards.

Rank. Experts rated each factor on a 1 to 5 scale for significance, relevance and comparability. Rather than splitting the ranked list into equal groups, statistical tests placed factors in the same tier when their ratings did not differ significantly.

Two-step method: identify resilience factors from literature and DOT plans, then prioritize them through a survey of transportation experts — **Figure 1.** Two-step research method.

03 / The survey

44 Florida transportation professionals

The survey went to 283 professionals from the Florida Transportation Plan's Resilience Sub-Committee, MPOs, FDOT district offices, rail, port and airport agencies, private firms, transit agencies, non-profits and academia. 44 responded (15.5%). All work in transportation planning with resilience expertise.

Using the Cochran formula, the paper estimates that 44 responses keep the standard error between 8% and 10% on average.

Pie chart of survey respondents: local 30%, state 21%, regional 18%, private sector 14%, special district 6%, other 6%, university 5% — **Figure 3.** Survey respondents by type of organization.

04 / Wind hazards

Emergency response, age of infrastructure and exposure come first

For wind-related hazards such as hurricanes, the top tier mixes socioeconomic, technical and environmental factors. Emergency response, a rapidity factor, ranked first on all three criteria.

Economic growth ranked lowest, which the paper attributes to a stronger emphasis on physical infrastructure when assessing transportation resilience.

Factor	Aspect	Dimension	Mean	Tier
Emergency response	Socioeconomic	Rapidity	3.89	Tier 1
Age of infrastructure	Technical	Robustness	3.77	Tier 1
Exposure	Environmental	Robustness	3.66	Tier 1
Maintenance level	Technical	Robustness	3.62	Tier 2
Proximity	Environmental	Robustness	3.58	Tier 2
Travel safety	Socioeconomic	Robustness	3.49	Tier 2
Recoverability	Technical	Rapidity	3.36	Tier 2
Network demand	Socioeconomic	Robustness	3.29	Tier 2
Utilities and drainage	Technical	Robustness	3.28	Tier 2
Social vulnerability / equity	Socioeconomic	Robustness	3.17	Tier 2
Network accessibility	Technical	Rapidity	3.02	Tier 2
Link capacity	Technical	Robustness	2.98	Tier 3
Economic growth	Socioeconomic	Both	2.89	Tier 3

Figure 5. Ranking of factors for wind-related hazards.

Overall mean of the significance, relevance and comparability ratings on a 1 to 5 scale. Values from the paper.

05 / Water hazards

Physical infrastructure dominates the top tier

For water-related hazards such as inland flooding, five factors reached the top tier. Besides emergency response and age of infrastructure, utilities and drainage, recoverability and maintenance level rated high, underscoring the role of drainage and physical condition during floods.

The three lowest, social vulnerability and equity, economic growth and mobility, are not directly tied to physical infrastructure.

Factor	Aspect	Dimension	Mean	Tier
Utilities and drainage	Technical	Robustness	4.14	Tier 1
Emergency response	Socioeconomic	Rapidity	4.13	Tier 1
Age of infrastructure	Technical	Robustness	4.08	Tier 1
Recoverability	Technical	Rapidity	4.01	Tier 1
Maintenance level	Technical	Robustness	3.79	Tier 1
Network topology	Technical	Robustness	3.70	Tier 2
Travel safety	Socioeconomic	Robustness	3.60	Tier 2
Social vulnerability / equity	Socioeconomic	Robustness	3.49	Tier 2
Economic growth	Socioeconomic	Both	3.45	Tier 2
Mobility	Socioeconomic	Both	3.01	Tier 3

Figure 6. Ranking of factors for water-related hazards.

Overall mean of the significance, relevance and comparability ratings on a 1 to 5 scale. Values from the paper.

06 / From research to practice

Five recommendations from practitioners

Sixteen transportation professionals from FDOT district offices, MPOs and private firms reviewed the findings at a workshop. They agreed the factors cover the multidimensional nature of resilience and agreed with the prioritization.

A review of plans from seven coastal states found that six of them recognize the age of infrastructure as a resilience measure.

Proper communication of what resilience means across planners and state priorities.
Funding availability and public engagement to support new metrics and data.
Plans alignment of resilience goals across state, regional and local plans.
Data collection through reliable sources and standard protocols.
Comprehensive resilience assessment frameworks to measure resilience and its change over time.

Nickdoost, N., Jalloul, H., Choi, J., Smith, D. (2024). Identification and prioritization of multidimensional resilience factors for incorporation in coastal state transportation infrastructure planning. Natural Hazards, 120, 1603–1663.

View at Springer BibTeX RIS

Figures 1 and 3 reproduced from the paper, © The Author(s), under exclusive licence to Springer Nature B.V. 2023 Figures 5 and 6 redrawn from the values printed in the paper.

← Infrastructure Resilience

A Composite Index Framework for Quantitative Resilience Assessment of Road Infrastructure Systems

Navid Nickdoost, Mohammad Javad Shooshtari, Juyeong Choi, Dennis Smith, Yassir AbdelRazig · Transportation Research Part D, 131, 104180 · 2024

Read the paper Try the tool

Road resilience has many parts: how well roads withstand a storm, how fast they recover, and the technical, social, economic and geographic conditions behind both. This study builds a Resilience Index that combines these factors in one hierarchy, so planners can see whether resilience is rising or falling and trace any change to the factors that caused it. It was applied to the roads of FDOT District 5 in Central Florida.

Resilience factors identified

Counties in the case study

Decision-makers at the workshop

01 / The problem

Resilience studies look at one aspect at a time

Technical factors (the condition of roads and the shape of the network), socioeconomic factors (social vulnerability, economic growth) and geographic factors (exposure to hazards) all shape how a road system holds up in a disaster. Overlooking some of them can lead to misleading conclusions.

Existing research mostly measures single aspects, and earlier studies did not give planners a quantitative tool to monitor resilience continuously and catch changes as they emerge.

02 / The index

Five levels, from indicators to one score

Measurable indicators roll up into resilience factors. Dynamic factor analysis groups the factors into a smaller set of components. Components combine into robustness (keeping the system working after a disruption) and rapidity (how fast it recovers), for wind and for water hazards, and those combine into one system score.

Weights come from how much each factor explains in the factor analysis, and decision-makers can add their own weights for components, dimensions and hazards. Factors come from the literature and agency plans and were grouped through expert interviews.

Because the levels are linked, a rise or drop in the top score can be followed down to the factors behind it.

Hierarchy of the Resilience Index from indicators and factors up through components, robustness and rapidity, wind and water hazards, to one index — **Figure 2.** The Resilience Index hierarchy.

03 / The case

FDOT District 5, Central Florida

The index was built for the road system of FDOT District 5: Brevard, Flagler, Lake, Marion, Orange, Osceola, Seminole, Sumter and Volusia counties, with an estimated 74 million vehicle miles traveled each day.

Most indicators came directly from agencies such as FDOT, the Census Bureau and the U.S. Bureau of Economics; network topology indicators were calculated from historical network data. Data were brought to a quarterly frequency. Emergency response data were not publicly available for the district, so that factor was left out.

Maps of Florida highlighting the District 5 counties and the district's road network — **Figure 3.** FDOT District 5 counties and roadways.

04 / Results

Resilience rose from 2015 to 2021, with a drop in 2021

The district's road resilience increased over 2015 to 2021, and so did its resilience to both wind and water hazards. Rapidity improved more visibly, while robustness fluctuated and slightly decreased.

The most recent drop, in early 2021, traces down to the water-hazard index, then to both its robustness and rapidity, and finally to falling values in three factors: average age of infrastructure, social vulnerability and mobility.

Time series of the overall index, the wind and water hazard indexes, and robustness and rapidity indexes — **Figure 4.** Index results at the system, hazard and dimension levels.

05 / Why every factor matters

One factor alone can point the wrong way

Within wind robustness, economic growth rose while the average age of infrastructure moved the other way. An agency watching only the physical condition of its roads could conclude resilience was falling, when other factors were pulling it up.

Robustness and rapidity also moved differently: rapidity for water hazards improved while robustness fluctuated. For this district, the paper concludes that investing in robustness would be more effective than investing in rapidity.

Nine small charts of resilience factor trends, including network recoverability, safety, average age of infrastructure, economic growth and mobility — **Figure 5.** Resilience factors behind the water robustness and water rapidity indexes.

06 / Is the result stable?

The upward trend holds under other methods

A Monte Carlo analysis rebuilt the index with different choices for filling missing data (linear or cubic spline interpolation; division or Denton disaggregation) and for scaling it (min-max or standardization). Every combination still showed resilience increasing over the period.

The scaling method changed the range of the results most; the interpolation and disaggregation choices caused only small variations.

At a workshop with 16 decision-makers from district and state transportation agencies, the index was discussed for scaling to other districts, setting goals and tracking progress, comparing regions, and aligning resilience goals across state and local plans.

Index results under eight combinations of imputation, disaggregation and normalization methods, all trending upward — **Figure 6.** Uncertainty analysis for the top level of the index.

07 / Limits

What the study does not cover

Good, quantifiable resilience data were not always available, and short data records limited the time frame.
Resilience planning is still new for state DOTs, which need to define and report high-frequency resilience indicators.
The study covers roads in a coastal region; other hazards, other regions and other infrastructure systems are left for future work.

Nickdoost, N., Shooshtari, M. J., Choi, J., Smith, D., AbdelRazig, Y. (2024). A composite index framework for quantitative resilience assessment of road infrastructure systems. Transportation Research Part D: Transport and Environment, 131, 104180.

View at Elsevier BibTeX RIS

← Infrastructure Resilience

Resilient Infrastructure & Smart Engineering

Infrastructure Resilience

Infrastructure Management

Sustainable Built Environment

Intelligent Construction

2026

2024

2023

2022

2021

2018

Mechanical, Electrical and Plumbing Systems

Constructability

Research Methods

Heavy Civil Construction

Advanced Construction Technology

Research interests

Experience

Assistant Professor

Postdoctoral Researcher

Project Engineer, Structural Inspector, Planner and PMO Member

Education

Ph.D., Civil Engineering

M.S., Civil Engineering

M.S., Civil Engineering

B.S., Civil Engineering

Infrastructure Resilience

How do failures cascade between interdependent infrastructure systems?

How resilient are coastal roads to storms?

How should debris sites and hauling be planned after a hurricane?

How should a roadside debris site be run?

PULSE: Perishable Utility & Lifeline Signals During Extremes

Intelligent Construction

What makes a construction professional ready to use AI?

Infrastructure Management

What is moving Florida's transportation system?

How is the construction industry changing?

Sustainable Built Environment

C&D recycling policy simulation for Florida counties

How this works