Elias N. Houstis
Professor Emeritus of Computer Science
Joined department: 1984
Elias Houstis has served as acting and associate head of the Department of Computer Science. He is on the editorial board of Neural, Parallel and Scientific Computational, Computational Engineering Science, and HPC Users Web-Based journals and a member of the IFIP WG 2.5 Working Group in Numerical Software. Houstis's current research interests are in the areas of problem solving environments (PSEs), parallel computation, performance evaluation and modeling, computational intelligence, computational finance, and on-line learning. He is one of the principal designers of several domain specific PSEs (i.e., Parallel ELLPACK, PDELab) and numerous performance evaluation studies of PDE software and parallel architectures. He is leading the Parallel ELLPACK group, which is developing infrastructure and implementing methodologies for reusing "legacy" PDE software on a variety of physical and virtual parallel machines and designing a Web Parallel ELLPACK server. Houstis has been involved in the designing of a knowledge based framework (known as PYTHIA) to support the selection of algorithm and machine pairs for a given class of PDE problems based on performance knowledge. This framework has been applied to a simulation system for designing HPC systems (POEMS project), a virtual laboratory environment, and recommender system for mathematical software. He has published several books and over 120 technical articles. He has supervised 14 PhD students and several MS students. His research has been supported by the Air Force Office of Scientific Research, the Army Research Office, DARPA, DOE, ESPRIT, INTEL, IBM, AT&T, Kozo-Japan, Purdue University, National Science Foundation, and the Greek Research Foundation.