Improving Human-Computer Interaction in Locational Decision Making
Paul Densham Principal Investigator
MetadataShow full item record
The selection of combinations of sites at which to locate facilities has long been an important issue within the theoretical domain of geography and related scientific disciplines; site selection also has an immense range of practical applications for businesses, governmental agencies, and other organizations. Critical limitation on the development of more sophisticated location decision-making procedures have been capacity constraints on computational systems. With large numbers of possible combinations to be explored, the storage and processing capacities of even the largest computers quickly are exceeded. Other limitations on the development of practical locational decision-support systems have been difficulties in developing clear, effective, and accurate displays of the results of computational procedures. Even when procedures succeed, users frequently must wade through enormous tables in order to obtain the information on which a decision ultimately is based. This project will try to diminish both sets of limitations. One line of inquiry will consist of the development and testing of new algorithms that will allow use of parallel computing systems for the solution of locational decision problems. The second line of inquiry will focus on the development of more effective visual displays of data and solutions through the use of object-oriented programming techniques. This project will test the use of parallel computing architectures for the solution of complex locational decision problems. Although concentrating on combinatorial site-selection problems, the results of this research will provide insights into the utility of parallel structures for solution of a much wider range of computational problems. The project also will significantly enhance understandings of the efficacy of object-oriented programming methods for more effective visualization of the results of locational decision-making processes. In doing so, broader issues of visualization of computer- generated data and solutions will be addressed.