The research objective of this GOALI award is to develop advanced optimization methods for product design. These methods are based on the mathematics of biogeography, which is the study of the distribution, migration, speciation, and extinction of biological species. Biogeography is nature?s way of distributing species. This research will explore the feasibility of using the concepts of biogeography to aid in general problem solving. The research includes the investigation of biogeography theory, the development of biogeography-based problem-solving methods, the application of the methods to general product design, specifically automotive design, and benchmark testing of these methods. Deliverables include Internet-based biogeography optimization software for general-purpose product design optimization.
If successful, the results of this research will provide a framework for optimization based on biogeography theory that has the potential to greatly improve our ability to design products which possess a high number of variable parameters. These methods will be used early in the design process, before the detailed product design has been developed, to support the designer in exploring a wide range of design options. The software developed as a result of this research will be available for solving a variety of optimization problems. Close collaboration with the industrial partner, General Motors, will provide the opportunity for the research team to test these methods on industrial problems. Graduate students will gain industry experience through summer internships at GM. The minority population of Cleveland State University provides a rich environment from which to recruit minority students for involvement in this research. Dissemination will occur through journal publications, industrial seminars and presentations at technical conferences.
This grant enabled the development of biogeography-based optimization (BBO), which is a new method for finding optimal solutions to a wide variety of engineering problems. BBO, as its name implies, is motivated by biogeography. Biogeography is the study of the migration, speciation, and extinction of biological species. This research has generalized the mathematical models of biogeography and applied them to the solution of engineering problems. BBO is an algorithm that is implemented on a computer, and falls under the broad category known as "artificial intelligence." Some of the important issues that are related to computer algorithms, and especially optimization algorithms such as BBO, include the following. (1) Is the algorithm stable? That is, can the algorithm be guaranteed to output reasonable results? (2) Does the algorithm converge? That is, can the algorithm be guaranteed to eventually find the correct answer to the problem it is trying to solve? (3) Are there good mathematical models for the algorithm? If this question can be answered in the affirmative, then the mathematical properties of the algorithm (stability, convergence, dynamic behavior, etc.) can be proven without relying on time-consuming computer simulations. (4) Can the details of the algorithm’s motivating paradigm (bioegeography, in this research) be generalized to improve the performance of the algorithm? This research answered all of the above questions in the affirmative. One of the important BBO applications in this research was the optimization of complex systems. A complex system is defined as multiple related systems, where each system includes multiple objectives and constraints. Complex systems are becoming more prevalent in industry due to the increasingly multidiscplinary aspects of engineering practice and research. For example, an aircraft design problem might include a body frame design, flight controller design, wing design, landing system design, electrical system design, hydraulic system design, and several other related problems. Each of these problems is independent to a certain extent, but they all interact and affect each other. BBO is well-suited to complex system optimization due to its representation of candidate solutions as islands, which stems from its motivating framework of biogeography. Groups of islands (archipelagos) are used in BBO to represent candidate solutions to individual problems, and groups of archipelagos represent candidate solutions to multiple related problems. The establishment of BBO as a viable optimization tool has made an impact on several areas of engineering. Applications of this research included the following. (1) Mobile robot control optimization. The performance of mobile robots is determined in large part by their control algorithms, and BBO has found effective control algorithms for mobile robots. (2) Robot manipulator control optimization. The performance of robot arms in manufacturing facilities is determined by their control algorithms. BBO has found effective control algorithms for robot arms, and these results have been implemented in industry. (3) Cardiac disease diagnostic algorithms. Electrocardiograms (ECGs) are often examined by medical experts to determine if an individual has a cardiac abnormality. Automated diagnostics can be effective tools to assist physicians in this task, and BBO has been used to find cardiac diagnostic algorithms. (4) Prosthetic leg control optimization. Poor prosthetic leg performance requires compensation by the amputee, which leads to negative side effects such as arthritis and back pain. Therefore, a prosthetic leg needs to be effectively controlled so that the prosthesis closely emulates natural human gait. BBO has been used to find effective prosthesis designs and control algorithms. The theoretical support and successful applications of BBO produced by this research have led to its acceptance by many other researchers around the world, and its use in many applications. According to Google Scholar, in 2008, which was the first year of this grant, there were 9 publications about BBO. There were 40 publications in 2009, 99 in 2010, 169 in 2011, 304 in 2012, and 433 in 2013. According to Google Patents, there have been 27 patents related to BBO. Cleveland State University’s position as a leader in minority education provided several opportunities to recruit minority students for this research. Four students from under-represented groups participated in this research. In total, this project was instrumental in funding the education of two doctoral degree graduates, eight master’s degree graduates, and six bachelor’s degree graduates. The results of this research have been disseminated in several ways, including the availability on the Internet of BBO software, journal and conference publications, a textbook publication, and tours of the principal investigator’s research lab to about 200 high school students each year.
