Researchers in the Department of Industrial and Systems Engineering at Texas A&M University are working to better the world with their latest projects in energy, information and biological systems. They are dedicated to tackling these large-scale problems, as well as the variability and uncertainty that may arise.
![Operation Research Article](../_news-images/thinkstockphotos-476034794.jpg)
Advancing Optimization
Dr. Sergiy Butenko, Donna and Jim Furber ’64 Faculty Fellow in industrial and systems engineering, is working to develop mathematical global optimization techniques, while taking advantage of other network representations of complex systems.
“By exploring structural properties, we aim to characterize and optimize the corresponding network representations of the complex system,” Butenko said.
Currently supported by the National Science Foundation and the Office of Naval Research, Butenko’s team works to solve these problems in social networks, biological systems and financial networks.
“Our goals include solving several important classes of hard discrete optimization problems in networks,” Butenko said, “as well as applying the developed methods to various areas ranging from social network analysis to materials.”
Operating Efficiently
Dr. Natarajan Gautam, associate department head for undergraduate affairs, is conducting research on efficiently adding renewable energy to the power grid. In an effort to maintain great performance with minimal waste and low cost, his team is looking for ways to generate and transmit power more efficiently.
“This is particularly challenging when the environment has significant uncertainty and is constantly varying over time,” said Gautam. “The ultimate objective is to improve the quality of life.”
The amount of solar power generated has a pattern of variability that usually occurs on a daily or weekly basis. When researchers look for this pattern, often described as riding the wave, the variability and uncertainty becomes a bit easier to handle. In turn, decisions are made effectively when looking at the design and control of the system.
“We usually have some historic information with which we build a model to simplify a randomly evolving process,” Gautam said. “Then, at each time instant when the environment is revealed, we need to make a decision that is best for the long run. Some challenges are simplifying the large amount of information collected to develop algorithms that run as fast as possible on a computer.”
Gautam and his team work in sensor networks, healthcare and smart manufacturing researching reliability and uncertainty.
Analyzing Efficiently
As director of the Mathematical Optimization and Data Analysis Lab (MODAL), Dr. Kiavash Kianfar leads a team in analyzing large bodies of data and optimizing large systems. The goal of this research is to develop and implement faster and more efficient computational methods.
“The research done in MODAL is funded by the National Science Foundation, Qatar National Research Fund and the Texas A&M College of Engineering,” Kianfar said.
The research in MODAL can be categorized under three initiatives. The first looks for ways to develop faster and efficient algorithms to solve problems from a set of feasible solutions.
“We have a project working on improving the efficiency of a system’s production by putting together modules of various sizes and capacities,” Kianfar said. “This kind of planning arises in power generation, data storage and manufacturing systems.”
The second initiative studies the reliability of large systems of sensors. These systems are used to control the operation of manufacturing and service settings.
“Our underlying question is, ‘What are the best methods to determine the robustness of a particular system design against failure of its individual sensors?’,” he said.
The final initiative develops fast and resource-efficient algorithms to process biological data. The data is generated by genome sequencing technologies in research and clinical settings.
“Next-generation sequencing has revolutionized the research and treatment of complex diseases,” Kianfar said. “The research has resulted in enormous volumes of data which need to be processed and analyzed to answer various biological and medical questions. Our research aims at developing algorithms for this purpose.”
Achieving Accuracy
Dr. Erick Moreno-Centeno and his team are looking to more accurately calculate answers to algorithms.
When calculations are made, they are often rounded up or down after any number of decimal places. Proceeding with calculations and using the rounded off answers creates errors. These errors may accumulate and create a snowball effect. Therefore, the final results obtained may be dramatically incorrect. This is troublesome in applications that require completely accurate solutions.
“Our goal is to eliminate the inaccurate solutions and invalid solver output in an efficient manner,” said Moreno-Centeno.
Adolfo Escobedo, Ph.D. student, has supported Moreno’s efforts to develop a process that guarantees completely accurate solutions in algorithms.
“We have eliminated round-off errors in two of the core subroutines, the renowned LU and Choleksy factorizations, of linear optimization algorithms,” Escobedo said. “Now, we are trying to collaborate with other researchers who can implement our contributions into a commercial-grade optimization solver.”
This research will greatly impact any industry where inaccurate results can have dire consequences.
“One example is the $300 billion energy industry where inaccurate solutions can cost millions and even lead to disastrous blackouts,” Moreno-Centeno said.
Another example is the healthcare industry, where practitioners must be certain that the recommendations given by the software are accurate because inaccurate results can have serious health and, consequently, financial liabilities.
Sensing the Breeze
Dr. Lewis Ntaimo, Marilyn and L. David Black Faculty Fellow in industrial and systems engineering, is using sensors installed in wind turbines, to find the best way to gain preventive maintenance policy for turbine systems under random weather conditions. This policy decreases the expected average cost of wind turbines.
“This work is important because wind farm managers are faced with a very difficult problem,” said Ntaimo. “They must decide when to perform maintenance on wind turbines and before the components fail to minimize downtime and maintain production.”
If the components fail, they must be replaced as soon as possible to keep production losses low.
“Performing maintenance requires a maintenance crew and very expensive equipment such as a crane to allow the crew to reach the wind turbine,” Ntaimo said. “These can be located hundreds of feet above the ground.”
Ntaimo and his team have begun applying policies, such as the one used for wind turbines, to the healthcare, energy and space industries.
Impacting the World
The research these faculty members are conducting will make the world we live in a better place, while operating efficiently and maintaining the quality of the products and services we deliver. The Department of Industrial and Systems Engineering continues to focus on positively impacting the world for future generations.
“The main impact of our research is to make the goods and services faster, better and cheaper in terms of quality and environmental impact,” Gautam said.