Research in the Area of Disaster Response in the ISE Department at RPI
In the ISE department at RPI, we are creating new models, algorithms, and analytical tools to help better prepare for, respond to, and recover from significant disruptive events that impact the critical infrastructure and supply chain networks which society depends upon. Such events precipitate a wide range of impacts on the interconnected, complex systems that constitute our infrastructure for food, transportation, power, water, housing, and medical supplies. These technological systems are more vulnerable because they are interdependent; disruptions in one can spread to others, causing cascading and potentially catastrophic failures. This vulnerability is exacerbated by advances in communications and computing technologies that are now integral to the operations of our infrastructure systems. For example, efficient and effective global supply chains such could not function without both the logistical infrastructure to collect, store, and move goods and the information infrastructure to monitor and control the flow of those goods over the network.
Our research in this area has been funded by the National Science Foundation and the Department of Homeland Security. We are addressing problems in interdependent infrastructure resilience, disaster response in remote regions, and debris removal by applying tools from ISE. We welcome you to read more about each of these areas.
Interdependent Infrastructure Resilience
Civil and social infrastructure systems such as power, water, communications, and emergency services are crucial for the quality of life and safety of a community. Therefore, the resilience of such systems in the face of potentially disruptive events is extremely important for society. Resilience is often defined as the ability of a system (or set of systems) to withstand disruptive events and then rapidly recover from them. However, the resilience of civil and social infrastructure systems is complicated by their interdependencies, i.e., how one system needs services provided by other systems to properly function. Figure 1 demonstrates the interdependencies between the operations of infrastructure systems and how a disruption in one infrastructure can cause cascading disruptions in other systems. In this example, flooding has disrupted the operations of a power substation, thus disrupting "normal" power to the hospital. Backup generators at the hospital can help to ensure that it still has power; however, the power disruption also impacts the pump that provides potable water to the hospital. Therefore, the flooded power substation disrupts power to the water pump which causes a disruption to the water services to the hospital which in turns disrupts the operations of the hospital.
ISE researchers have created new optimization approaches for analyzing interdependencies in infrastructure systems and how they impact their resilience. For example, Professors Sharkey, Mitchell, and Wallace have analyzed new classes of integrated network design and scheduling problems that can help model the restoration efforts of these systems. Figure 2 provides an example of the output of such models - the green arcs currently have flow (modeling the services provided by the infrastructure network) on them, the white arcs do not have flow on them due to damage to the components of the network, and the purple arcs are the ones currently being repaired to restore the services provided by the network. The initial focus of this research was on modeling the restoration efforts of one infrastructure system and these researchers are now considering how to capture interests of multiple stakeholders in interdependent infrastructure restoration.
Interdependent infrastructure restoration is complicated by not only the interdependencies that exist between the operations of the systems but also their restoration interdependencies. For example, when trees come down onto power lines, there are tasks that need to be performed by multiple infrastructures and have constraints relating their completion times. In particular, the power company will need to send a crew out to inspect the damaged lines and ensure that the area is safe to enter. This task needs to be completed before the Department of Public Works can begin clearing the road of the downed trees. After the road is cleared, then the power company can send a work crew out to repair the downed lines. These precedence relationships are complicated by the fact that there are at least two decision-makers involved in the process - the managers of the power company and the decision-makers for the Department of Public Works. These relationships link the restoration decisions of multiple infrastructure systems and are thus referred to as restoration interdependencies. Figure 3 provides a comparison of the traditional scheduling environment (with a single decision-maker) and the environment faced during interdependent infrastructure restoration (with multiple schedulers and green restoration interdependencies linking their restoration efforts). Although the set of tasks is the same, the environment on the right is complicated by having both a red scheduler and a blue scheduler. Professors Sharkey, Mitchell, and Wallace are creating new optimization approaches to help to determine the value of coordination between these schedulers and, if full coordination is not possible, the value of information-sharing between them.
Disaster Response Capabilities in Remote Regions
In 2014, the National Academies of Science published a report led by ISE research professor Martha Grabowski on "Responding to Oil Spills in the U.S. Arctic Marine Environment" which highlighted the need for improved oil spill response (OSR) capabilities in the Alaskan Arctic especially given the potential interest in oil drilling in the area (see Figure 4). In response to this report, Dr. Grabowski, along with ISE faculty members Thomas Sharkey and Al Wallace, began collaborating with officials in District 17 of the United States Coast Guard and launched a research initiative for dynamic modeling of arctic research allocation (DMARA). The first phase of DMARA has created new optimization approaches to determine where and when to build OSR logistics infrastructure to increase OSR capabilities in the Arctic. Our focus was on creating new optimization models and algorithms in order to better understand long-term infrastructure investments, seasonal reallocations of response resources, and how the location of these response resources impact the capability to effectively respond to potential oil spill incidents. We are currently examining DMARA in the context of search and response for the Alaskan Arctic.