Rainwater harvesting is the process of collecting, storing, and distributing rainwater for reuse, rather than allowing it to run off. Greywater recycling consists of purifying processes that remove contaminants present in wastewater coming from faucets, showers, baths, clothes washing, and dishwashing, so that it can be reused, instead of becoming inlet wastewater for municipal wastewater treatment plants. In urban areas, rainwater harvesting and greywater recycling systems provide additional water supplies for houses, buildings, and industries, and also help to mitigate flooding events and pollution. Construction companies are looking for eco-friendly systems that benefit the environment while being profitable. This paper proposes a decision support system (DSS) for the design of rainwater harvesting and greywater recycling system (RHGRS) in residential construction projects. The DSS uses optimization to support construction companies in the design of a RHGRS that minimizes construction and operational costs. With the underlying optimization routine, the DSS addresses both the challenges of deciding whether to construct or not a RHGRS for a residential unit and its minimum-cost design. It is expected that the DSS will help construction companies to make decisions toward sustainable and eco-friendly cities.

Contact: Andres Medaglia

Increasing urbanization is an undeniable feature of this century. Addressing urban growth under severely constrained resources is a task that demands comprehensive efforts in planning towards efficient development: optimizing the treatment of resources (natural, economic) and residues (emissions, waste, idle time), for which an accurate understanding of urban processes is necessary. While network analysis and systems dynamics provide tools to describe and understand relationships within and between stakeholders and the physical world, statistical models and simulation allow to estimate and predict behavior patterns that enable better decision making. Qualitative and quantitative descriptions of urban problems are translated to optimization models that provide support in the design of public policy and strategic planning in problems such as: land-use regulation; incentives and planning for renewable energy adoption; configuration and planning for transportation systems; resource allocation for disaster preparedness.

Contact: Camilo Gomez

Effective resource allocation and quantification of its benefits are essential in practical engineering problems. However, determining the allocation and location of resources is computationally challenging when dealing with large, interconnected systems, as is the case of infrastructure networks. We address the allocation and location of different types of resources to a set of potential sites in an infrastructure system, aiming to satisfy service demands at a minimum cost (e.g., in the context of disaster preparedness). Exact optimization methods become impractical for realistic network sizes, whereas fully heuristic approaches lack optimality guarantees. We use exact optimization to solve a heuristically pre-processed version of the resource location/allocation problem, which incorporates constraints that capture topological network properties based on community detection algorithms. Adding topological constraints leads to significant reductions in computation time without considerable deviation from optimal solutions, as community detection mostly forbids solutions that are unattractive in terms of network topology. The proposed approach to resource allocation incorporates a novel way to account for infrastructure network configurations within linear programming, leading to computational efficiency while producing solutions that capture the inherent topological properties of the network.

Contact: Camilo Gomez

In the context of infrastructure resilience, key decisions occur in the aftermath of adverse events, but most importantly, preemptive decisions must be made without knowledge about the characteristics of a potential future event. An optimization framework is proposed to address problems in which preemptive decisions are coupled with those eventually required to respond to an uncertain adverse event. Strategic decisions are pursued regarding whether to proactively retrofit or reactively repair bridges in a transportation network under seismic hazards, aiming to minimize the cost of maintaining travel times under a set of adverse scenarios. A two-stage stochastic programming approach is proposed, which relates pre- and post-event decisions, using a decomposition of the optimization problem which is advantageous when dealing with complex networks. We perform an analysis of the San Francisco Bay Area transportation network, as an instance of a realistic, complex infrastructure network. Results evidence the potential of the approach to provide risk-informed decision support, while being able to deal with large sets of components and scenarios under an exact optimization approach, and solving problems with millions of variables and constraints.

Contact: Camilo Gomez

Resilience of infrastructure networks is key for urban development, mostly when understood as the way in which physical systems help societies prosper. In this sense, infrastructure engineering must incorporate the societal aspects related to infrastructure operation, namely: the organizations involved in infrastructure operation (government, agencies, contractors, regulators), and the general public who uses and benefits from these systems. The explicit consideration of these parties impacts the modeling and analysis of infrastructure systems with regards to how decisions are made (depending on the type of stakeholder, and its organizational constraints) and how the impact is measured (depending who the constituents are). For example, in the context of risk assessment and management, organizational constraints may imply stakeholders with divergent interests, or centralized vs. decentralized resource availability, decision processes, and operations, etc.; regarding user-related aspects, socio-economic vulnerability may have an impact on how different segments of the population experience the (lack of) service/benefits of an infrastructure system. Incorporating these aspects may affect the definition of performance metrics of interest and, consequently, influence risk management policies. In this project, we propose recovery strategies for interdependent networks of power, gas, and water in Shelby County, TN (USA), considering operators’ as well as societal objectives within an optimization framework.

Contact: Camilo Gomez

Although technological developments have provided momentum to extend the frontier of commercially feasible network deployments, the latest Millennium Development Goals Report from the United Nations shows that the digital divide between the rich and poor is increasing due to the lack of affordable services. Therefore, an economic framework is needed to create conditions for affordable network services. In this paper, we introduce a set-aside mechanism that can satisfy this need by reserving resources for targeted groups and resolving the practical problem of having greedy users that rationally compete for cheaper resources. In this mechanism, prices are tailored to users’ budget capacities and quality needs. The simulation results indicate that it is possible to increase the resource allocation for delivering services to the poorest while simultaneously maintaining operators’ revenues by inducing users to discriminate among themselves.

Contact: Sergio Cabrales

This paper explores the properties of using a generalized additive model with embedded variable selection for the prediction of bankruptcy. During the last two decades, many machine learning and artificial intelligence algorithms have been implemented on this problem with the main purpose of provide higher prediction accuracy. However, there exists a duality between interpretation and prediction that have prevented the massive use of these novel methods. Using an additive model allows to incorporate nonlinear effects for each predictor enhancing the predictive power over classical linear models, but at the same time, keeping separated the marginal effects for interpretation. Also, an innovative penalization likelihood approach (GAMSEL, Chouldechova and Hastie (2015)) automatically selects important financial ratios and classifies them between linear and nonlinear effects, improving the interpretation of the estimations.

Contact: Carlos Valencia, Sergio Cabrales

The COPA group conducts research on a wide range of energy issues. One of the topics is the oil economy, commodity price review and volatility estimation, energy business strategies, risk management, hedging strategy, and regulation of energy industries. Another new area of research focuses on the economics of renewable energy, especially wind, solar and water.

We analyze that International Oil Companies generally maximize the expected net present value of their profits, whereas for National Oil Companies, profits are not necessarily the only objective. Instead, National Oil Companies’ objectives might have a wide range of objectives such as wealth re-distribution, economic development, domestic employment, subsidized domestic fuel price, energy security, royalties and dividends (Shleifer & Vishny, 1994; Hartley & Medlock, 2008; Eller, Hartley, & Medlock, 2011; Hartley & Medlock III, 2013). We develop a production, reinvestment (exploration and recovery) to study the effects of overemployment and domestic fuel subsidies on operations and finance of National Oil Companies. I find that overemployment and domestic fuel subsidies dramatically reduce market value, production, and reinvestment of National Oil Companies.

Contact: Sergio Cabrales

This work presents new evidence on the determinants of privatization of National Oil Companies (NOC) from a panel data of 49 oil-producing countries for the period comprised between 1996 and 2014. We find – ceteris paribus – that countries with greater control of corruption, a perception of high government effectiveness, a more participatory political system, greater regulatory quality, and high efficiency in the execution of contracts and property rights, are more likely to privatize their NOCs, compared with countries that have a lower level in these governance indicators. Likewise, in terms of the control variables, the results of previous studies were ratified regarding the way in which high oil prices negatively influence the probability of privatization of oil companies, while its deviation from the trend and population have a positive effect on these.

Contact: Sergio Cabrales