By: Lisa R. Weidman
It sounds like something from a science fiction movie: Researchers are studying immune systems–more specifically clonal selection algorithms–and using the technology to increase efficiency in facilities like hospital emergency rooms, manufacturing plants, and grocery stores.
At Penn State Berks, Dr. Sadan Kulturel-Konak, Associate Professor of Management Information Systems, and Dr. Berna Ulutas, a post-doc fellow who visited the college from May 2009 to May 2010, worked together to study how artificial immune systems (AIS) can be used as a tool to solve facility layout problems.
Generally defined as arranging machines and departments in a facility so that the material flows among those departments are minimized, facility layout problems are known to be complex and hard to solve with exponentially increasing difficulty with the growing number of departments.
Kulturel-Konak and Ulutas explain that the immune systems of living organisms are very complex in their nature, and they are comprised of many efficient mechanisms, such as clonal selection, learning ability, memory, and flexibility. These systems have attracted the attention of researchers who believe they may be put to use to solve facility efficiency challenges, such as pattern recognition, job shop scheduling, and computer security, to name a few.
Artificial immune systems (AIS), which are inspired from immune system principles and models, are basically classified as population-based (clonal selection, negative selection, and bone marrow) and network-based models (continuous and discrete). As a solution technique, the researchers have applied the clonal selection algorithm (CSA).
While application of the CSA on facility layout is a fairly new concept, the researchers have achieved impressive results for their project titled “Assessing Alternative Layouts in Uncertain Environments,” and they have submitted a number of papers on the subject to well known journals in the fields of industrial engineering and operations research.
Kulturel-Konak and Ulutas explain that in the clonal selection theory, evolution occurs based on the two basic principles of the vertebrate immune system: affinity maturation and receptor editing. During proliferation, the antibodies undergo a hypermutation, which diversifies the repertoire of the white blood cells. After the hypermutation processes, the receptor-editing mechanism is applied and the worst percentage of antibodies in the population is eliminated and replaced by randomly created antibodies.
“The clonal selection algorithm is a population-based algorithm that has some similarities with well known heuristics, such as genetic algorithms,” explains Kulturel-Konak. “These algorithms create small changes that mutate the layouts and generate new ones, changing the entire scenario.”
“Since parallel search methods perform better than strictly serial and randomized search methods, we used clonal selection algorithms that can exploit and explore the solution space effectively.”
Ulutas adds, “In our algorithm, each antibody string, which is formed of the department sequence, corresponds to a layout or solution, and population is defined as the set of these layouts. The layouts are assessed based on the material handling costs among departments. The goal is to assign departments to locations so that the material handling costs are minimized.”
Since material-handling is not a value-added part of any business process, the most efficient layouts will have the lowest material handling costs.
Ulutas is an Assistant Professor of Industrial Engineering at Eskisehir Osmangazi University in Turkey, where she earned her Ph.D. in Industrial Engineering. She met Kulturel-Konak in the summer of 2008 when they were both attending the Women Industrial Engineers Workshop in Turkey. Ulutas applied for funding from the Scientific and Technological Research Council of Turkey International Postdoctoral Research Scholarship Program to work with Kulturel-Konak as a post-doc fellow at Penn State Berks.
Despite the fact that an ocean now separates them, Kulturel-Konak and Ulutas continue to collaborate on their research.
When asked why she wanted to travel to the United States to do research, Ulutas explains, “Penn State is one of the most respected universities for engineering, but my main reason for wanting to come to Berks was the opportunity to study with Dr. Kulturel-Konak.”