The three-year $347,995 grant from the Department of Defense Air Force Office of Scientific Research aims to identify and optimize counter measures to protect against potential adversarial and naturally occurring disruptions to systems management.
Assistant Professor Ningji Wei with the Department of Industrial, Manufacturing, and Systems Engineering (IMSE) in the Edward E. Whitacre Jr. College of Engineering (WCOE) will apply his expertise in decision making under uncertainty in large systems to improve resilience.
“The first step requires evaluating the previous methods that have been employed to achieve the goal,” said Wei. “What does prior experience show about pre-incident planning and about post-disruption counteractions?”
Both approaches to developing resilience in the organization have disadvantages. On the planning side, it is not possible to predict the future. Often organizations may become too conservative because they are seeking to avoid a wide array of scenarios, some of which may never occur. After an incident occurs, the timing of the organization’s accurate response is crucial to a positive outcome, but there is often a very short window of time available to determine and activate the solution.
For any organization, planning for a potential disruption to services—whether that involves transportation, communication, or some other critical infrastructure—requires evaluating the potential impact from a wide array of scenarios that may occur at any given location, or even multiple locations, within the system.
Wei’s project will focus on addressing the challenges of finite adaptability and scalability. Finite adaptability is a process that seeks to optimize decision making under uncertainty. Wei will identify a small number of potential solutions that can be applied effectively across a range of scenarios.
“This approach makes effective counteraction more manageable and streamlines the decision-making process,” Wei explained. “It is not necessary to plan for every possible disruption, and in fact, it is detrimental to the system. As the defender, the best strategy is to develop a small number of options that optimize the system’s response to a wide range of disruptions.”
Once these counteractive measures are identified, then the second challenge must be solved for. The scale of the disruption matters. As the problem becomes larger, the computational power needed to evaluate and solve also increases exponentially. That increase often expresses itself in time, more time is needed to process and present the best solution.
Often, time is a constraint that doesn’t allow for a lengthy evaluation process.
Wei’s project will determine the opportunities where compromises can be made in computational efficiency to produce solutions that are good enough to accomplish the desired outcome based on the time available for the decision to be made.
“Through machine learning, one can identify patterns that indicate the solutions that are sufficiently good that can be computed more efficiently,” said Wei. “This gives decision makers the ability to more efficiently determine which potential counter measures to activate within specific constraints.”
For example, if a disruption occurs to a transportation system, the key constraint is how much time is available to make the decision about the best alternative route or series of routes? If there is sufficient time, the organization can run a complete analysis to arrive at the optimal alternative route and activate that measure.
But what if a decision has to be made in 10 minutes and it takes an hour to run the full analysis for the best solution? Wei’s tool will provide decision makers the ability to adjust the computational power to yield counter-measure options that are optimized for the constraints.
Wei’s findings from this project will be incorporated into future undergraduate IMSE coursework as well as in a computational foundations camp for incoming doctoral students.
“Dr. Wei’s work addresses an important challenge—how to make good decisions quickly when systems are disrupted—and I am particularly pleased that he plans to bring what he learns through this research into our classrooms, directly benefiting our students,” said Burak Eksioglu, chair for the IMSE Department.