Omar Kammouh
Assistant professor @TU Delft | Infrastructure resilience & optimization | Complex systems modeling | Matlab/Python enthusiast
Critical infrastructures are highly interconnected and their management can purposely contribute to the sustainable development and well-being of our communities. My research focuses on infrastructure resilience and optimization in a cross-sectoral and multi-actor setting. I especially enjoy modeling complex adaptive systems, like infrastructure. I also support the development of guidelines and solutions to apply responsive innovations. Reach out to me and I will be glad to chat with you!
---
I have received my B.Sc. (Honour, 2013) from Beirut Arab University and my M.Sc (Honors, 2015) from the University of Bologna in Civil and Environmental Engineering. I completed my Ph.D. in 2019 in the domain of community and infrastructure resilience at the Politecnico di Torino in Italy. Following my Ph.D., I was appointed as a research fellow at Delft University of Technology (2019-2021) and then at Eindhoven University of Technology (2021-2022) in the Netherlands. As of 2022, I have been appointed as an assistant professor at Delft University of Technology supporting both research and education.
Phone: 0031 6 39124445
Critical infrastructures are highly interconnected and their management can purposely contribute to the sustainable development and well-being of our communities. My research focuses on infrastructure resilience and optimization in a cross-sectoral and multi-actor setting. I especially enjoy modeling complex adaptive systems, like infrastructure. I also support the development of guidelines and solutions to apply responsive innovations. Reach out to me and I will be glad to chat with you!
---
I have received my B.Sc. (Honour, 2013) from Beirut Arab University and my M.Sc (Honors, 2015) from the University of Bologna in Civil and Environmental Engineering. I completed my Ph.D. in 2019 in the domain of community and infrastructure resilience at the Politecnico di Torino in Italy. Following my Ph.D., I was appointed as a research fellow at Delft University of Technology (2019-2021) and then at Eindhoven University of Technology (2021-2022) in the Netherlands. As of 2022, I have been appointed as an assistant professor at Delft University of Technology supporting both research and education.
Phone: 0031 6 39124445
less
InterestsView All (12)
Uploads
Finding the best grouping strategy is a known NP-hard problem, with several optimization strategies have been proposed, mainly based on nonlinear models which are computationally expensive and do not guarantee scalability. Furthermore, infrastructure intervention planning models mostly focus on grouping of interventions which are considered as given. In this paper, we propose a new efficient optimization model to optimize intervention grouping for interconnected infrastructure networks. We develop a scalable two-step optimization model where we first plan each individual intervention type based on a preventive maintenance policy accounting for the degradation behavior of objects, then group interventions to minimize the net costs, considering dependencies within and accross infrastructure networks.
We formulate the grouping problem as an Integer Linear Program, which can be solved exactly with standard solvers. The model accounts for interactions between infrastructure networks and considers the impact on all stakeholders. It also accommodates various intervention types like maintenance, removal, and upgrading.
Using a demonstrative application, we show that our model significantly reduces net costs and outperforms alternative nonlinear formulations and related heuristics in terms of both solution quality and computation performance. Additionally, the optimal intervention plan shows repetitive patterns, which suggests that a rolling horizon strategy could be used where the optimization problem is solved for shorter time horizons, leading to significant computational benefits.
Finally, different seismic scenarios have been applied to a large-scale virtual city model. The platform proved to be effective to analyze the emergency and could be used to implement countermeasures that improve community response and overall resilience.
Finding the best grouping strategy is a known NP-hard problem, with several optimization strategies have been proposed, mainly based on nonlinear models which are computationally expensive and do not guarantee scalability. Furthermore, infrastructure intervention planning models mostly focus on grouping of interventions which are considered as given. In this paper, we propose a new efficient optimization model to optimize intervention grouping for interconnected infrastructure networks. We develop a scalable two-step optimization model where we first plan each individual intervention type based on a preventive maintenance policy accounting for the degradation behavior of objects, then group interventions to minimize the net costs, considering dependencies within and accross infrastructure networks.
We formulate the grouping problem as an Integer Linear Program, which can be solved exactly with standard solvers. The model accounts for interactions between infrastructure networks and considers the impact on all stakeholders. It also accommodates various intervention types like maintenance, removal, and upgrading.
Using a demonstrative application, we show that our model significantly reduces net costs and outperforms alternative nonlinear formulations and related heuristics in terms of both solution quality and computation performance. Additionally, the optimal intervention plan shows repetitive patterns, which suggests that a rolling horizon strategy could be used where the optimization problem is solved for shorter time horizons, leading to significant computational benefits.
Finally, different seismic scenarios have been applied to a large-scale virtual city model. The platform proved to be effective to analyze the emergency and could be used to implement countermeasures that improve community response and overall resilience.
uncertainties in the restoration process; and Chapter 8 focuses on modeling infrastructure interdependencies using temporal networks at different spatial scales.
In this presentation, we will draw on the results of our latest research work on interventions planning optimization. We will present an integrative multi-system and multi-stakeholder optimization approach for managing infrastructure interventions. The proposed approach takes advantage of the benefits achieved by optimally grouping intervention activities. Intervention optimization leads to substantial savings on both direct intervention costs (e.g., cost of maintenance) and indirect unavailability costs (e.g., societal impact due to service unavailability) by reducing the number of system interruptions. The proposed optimization approach is formalized into a structured mathematical model that can account for the interactions between multiple infrastructure networks and the impact on multiple stakeholders (e.g., society and infrastructure operators), and it can accommodate different types of intervention, such as maintenance, removal, and upgrading. The different types of interdependencies, within and across infrastructures, are modeled using a proposed interaction matrix (IM). The IM allows integrating the interventions of different infrastructure networks, which are normally planned independently
To demonstrate the applicability of the proposed approach, an illustrative example of multi-system and multi-actor intervention planning is introduced. Finally, we illustrate how systems thinking approaches can be beneficial and we highlight the requisites of such approaches.
We propose an integrated model for infrastructure resilience enhancement that incorporates time-varying crew productivity in the recovery process. We develop the following integrated probabilistic formulation for time-varying human productivity:
η_q^'=ω_f (t)ω(q_κ⁄q_(κ,"min" ) )^(1-ε_κ ) η_q (1)
Where η_q and η_q^' are the base and the modified productivity values of a crew of type κ and size q_κ; q_(κ,"min" ) is the minimum required crew size; ω is a factor that captures specific conditions such as weather (Sharma et al., 2018); ε_κ is a small positive constant to adjust for the crew congestion in a team, and ω_f (t) is the time-varying reduction factor due to fatigue. We estimate the predictive distribution F[ω_f (t_i)|h_w,T_s,d_r ] where ω_f (t_i) is the fatigue reduction at any given time, based on the work hours per day h_w, daily shift start time T_s, and rest days per week d_r. We train the predictive model using data from Hursh et al. (2004) and Taoda et al. (2008). We then integrate the resulting model of time-varying productivity from Equation 1 with the formulation from Sharma et al. (2020) to develop a realistic model for the physical recovery of spatially distributed infrastructure. We also develop a computational approach that can provide stochastic optimization for physical recovery optimization and cost while also considering high-fidelity flow analysis.
We implemented the formulation for the resilience optimization of interdependent power and water infrastructure in Shelby County, Tennessee, USA, subject to a scenario earthquake. The infrastructure details and the earthquake scenario are available in Sharma et al. (2020). To study the impact of workforce productivity on resilience, we studied different work hours per day, h_w, rest days per week, d_r, and number of shifts in a day, s. Figures 1(a-b) are Pareto diagrams showing the results of the water infrastructure. Here, ρ_"physical" and ρ_"service" are resilience metrics (definitions available in Sharma et al., 2020) for physical recovery and recovery of functionality. Results indicate that work hours per day and rest days make a substantial difference to infrastructure resilience by affecting workforce productivity. We also observed that the effects worsened as the recovery times worsened. We found the proposed formulation effective in providing a monetary tradeoff of implementing rapid recovery using long working hours. Specifically for the case study, 16 work hours with two 8-hour shifts perform substantially better than the rest because it has twice the resources. Furthermore, shifts of 8, 10, and 12 work hours per day present a Pareto front at fixed resource levels.
Knowledge convergence, or consensus, refers to the alignment of diverse information, expertise, and viewpoints to form a common understanding or agreement on a specific topic, integrating potentially conflicting or inconsistent information (Ford, 2008). Conversely, knowledge divergence occurs when emerging perspectives or understandings clash with established knowledge, potentially causing fragmentation and misinformation, yet also fostering innovation and knowledge refinement (Ford, 2008).
Investigating the progression towards convergence or divergence necessitates an understanding of the scientific consensus process (Linkov et al. 2014; Cutter et al. 2013). The process of scientific consensus is depicted in Figure 1. Initially, as the topic gains interest, new scholars emerge, and their opinions align with those of established experts, leading to an increase in scientific consensus (i.e., premature consensus). This is followed by a period of debate and opinion divergence (Shwed and Bearman 2010), leading to a critical 'pivot point' where the consensus will either continue declining (indicating divergence) or start increasing again (indicating convergence). The scientific community's efforts in research, open communication, and standardized practices heavily influence this pivot. During these phases, the formation of distinct schools of thought may occur, leading to localized convergence within a global context of divergence.
To examine the state of scientific consensus in the resilience concept, this research integrates an in-depth literature review, insights from an academic expert workshop, and an analysis of significant historical events that have shaped the resilience narrative (Dedehayir and Steinert, 2016). A notable methodological tool used is the hype cycle model, which maps the evolution of resilience as a concept. This model correlates the prominence of resilience with key developmental milestones and impactful global occurrences, such as major natural disasters and technological shifts. The research also makes a distinction between the ontological (theoretical) and operational (practical) aspects of resilience. There is a broad consensus in the ontological perspective of resilience, indicating a shared theoretical understanding. However, operational definitions exhibit considerable variability, reflecting diverse applications and interpretations in practice.
Several impediments to achieving consensus in resilience research are identified. These include the lack of standardized terminology, ambiguity in resilience governance roles, challenges in defining the boundaries of complex systems, and inadequate data-sharing protocols. These barriers contribute to the conceptual disparity and affect the practical application of resilience principles across different domains.
The study also highlights the interplay between the exploration of innovative ideas and the consolidation of existing knowledge in resilience research. It emphasizes the risks associated with premature convergence, which could lead to an oversimplified view of resilience, and the complexities introduced by excessive divergence, particularly in educational contexts. Addressing these challenges necessitates a strategic balance, fostering innovation while maintaining consistency in resilience studies.
To tackle these challenges, the paper recommends developing a universally accepted resilience lexicon, conducting regular and comprehensive systematic reviews, and enhancing open dialogue among researchers and practitioners. Forming interdisciplinary groups and engaging in consensus-building activities, led by experienced researchers, is vital to integrate diverse perspectives and refine the understanding of resilience.
At the country level, a novel approach to assess the resilience of countries is presented. The approach is inspired by the classical risk analysis, in which risk is a function of vulnerability, hazard, and exposure. In the proposed analysis, the resilience-based risk is a function of resilience, hazard, and exposure. The methodology is applied to 37 countries for which the resilience is quantified. At the community level, the resilience of communities is tackled by proposing two indicator-based methodologies. The methodologies combine deterministic and probabilistic approaches within a single framework. The first method requires data on previous disasters as an input and returns a performance function for each indicator and a performance function for the whole community as an output. The second method exploits knowledge-based fuzzy modeling for its implementation. A matrix-based interdependency technique that serves as a weighting scheme for the different indicators is also introduced.
In the dissertation, we go in more details when tackling the resilience at the infrastructure level. Resilience can be defined using two main components, the damage incurred following a disastrous event and the restoration time of the system undergone the damage. Therefore, each of the two components is treated separately. For the damage component, two main lifelines, namely water and transportation networks, are tackled. For the water network, a simulation-oriented approach to evaluate the resilience of large-scale water distribution networks is proposed. The failure of the water system occurs when the water flow and the water pressure go below a certain threshold. The resilience of the network is evaluated using two indexes: (1) the number of users without water, (2) the drop in the total water supply. For the transportation network, a resilience evaluation methodology of large-scale transportation networks is presented. First, the road map of a virtual city is transformed into an undirected graph. Random removal of the roads is applied until the network’s failure point is reached. The network resilience is then calculated using the Destruction Spectrum (D-spectrum) approach. Multiple coding algorithms are presented in this chapter to solve several computational challenges. For the other component, restoration time, four main infrastructure systems (water, gas, power, and telecommunication) are considered. A large database that includes damage data for many earthquakes that took place in the last century has been collected from the literature. The database has been used to create restoration curves for the lifelines. The restoration curves have been presented in terms of probability of recovery and time; the longer is the time after the disaster, the higher is the probability of the infrastructure to recover its functions.
Finally, a generic resilience framework to assess the resilience of any engineering systems is presented. The temporal dimension is tackled using the Dynamic Bayesian Network (DBN). DBN extends the classical BN by adding the time dimension. This allows predicting the resilience state of a system given its initial condition. Two case studies are presented in the chapter to illustrate the applicability of the introduced framework. One case study evaluates the resilience of Brazil. The other one evaluates the resilience of a transportation system. The framework can be used to study systems that are not explicitly studied in the dissertation. Although it is probabilistic, the framework is practical and can be used by decision-makers in their day-to-day life.
The results of the dissertation provide valuable insights into the decision-making process regarding the resilience of communities. The solutions proposed by the models are improvements over previous work conducted and could benefit decision makers before, during, and after a disaster. In addition, not only does this research provide benefits to decision makers, but it steps beyond research and offers tools that are readily available for the emergency and infrastructure management communities.