Biography

I have worked for nearly 20-years in industry while also affiliated with a number of universities in Canada, North Carolina, Germany, Slovenia, Saudi Arabia, and Brazil. In 1986 I initiated, then continued as a General Chair until 1992, the International ACM Workshop Series on Layout and Logic Synthesis, held yearly at MCNC, Research Triangle Park, NC (with about 150 attendees from Americas, Europe, and Asia). This trend-setting series of workshops was a follow-up on the innovative organization of IEEE special sessions at ISCAS85 (Kyoto, Japan) and at ISCAS89 (Seattle, WA); both sessions were associated with distributions of challenging test generation benchmarks, now known under acronym of ISCAS85 and ISCAS89. As of October 2021, there are O(807,000) and O(1,110,000) search engine hits for each of these two acronyms alone. A world-wide following from other special interest groups, including the NSF-sponsored DIMACS series of workshops, ensued in 1990's.

Since 2000, I have been actively engaged in promoting a more disciplined approach to performance testing of combinatorial algorithms. Particularly, the use of graph isomorphs did expose large performance runtime variability of competing algorithms and thereby did enable rigorous statistical runtime performance testing of the underlying solvers. Our experiments with hard combinatorial problems in different domains continue to demonstrate the intrinsic merits of a statistical model that reliably predicts the asymptotic runtime complexity of any solver -- provided that the solver is tested, without censoring, on well-defined instance families of increasing size.

For a number of years I  coordinated hosting of departmental seminars as well as the Triangle Computer Series Distingushed Lecturer Seminar Series (TCDLS). This series transitioned from DVD recordings to video posting  on YouTube:  a total of 49 talks from 2007 to 2012 have been posted. As of March 2023,  the number of views for each1-hour talk ranges from 367 to 9400.

My web site extension has been moved to a new location. For updates, see www.opusFB.org

Research Areas

  • Algorithms and Theory of Computation
  • Scientific and High Performance Computing

Education

Ph.D. 1970 Electrical Engineering, University of Colorado, Boulder, Colorado, USA

Dipl.-Ing. 1965 Electrical Engineering, University of Ljubljana, Slovenia

Awards

  • Grant by Mentor Graphics ($15,000) - support for the computational resources of the laboratory - 2005
  • Two grants by DARPA ($1,000,000) - Globally Distributed Microsystems Design: Proof-of-Concept (The Vela Project: a collaborative project engaging 6 universities) - 1995-2000
  • Grants by SRC/Sematech ($400,000) - National Benchmark Program in Microelectronic Systems Design Automation - 1994-1998
  • Grants by ACM/SIGDA ($84,000) - Benchmark Archival and Distribution - 1992-1994
  • Grant by Siemens Corp. ($62,000) - Partitioning and Testability of FPGAs - 1992
  • Travel and Scholarship Grant by German Science Foundation, a collaborative research sabbatical at the Technical University of Munich, Germany - 1992
  • Travel Grant by Saudi Arabian Ministry of Education for a short course on CAD methods for VLSI design at at King Fahd University, and an invited talk at VLSI conference in Dhahran, Saudi Arabia - 1991
  • Best paper award - IEEE International Conference on Circuits, Computers, and Design - 1990
  • Travel Grant by Brazilian Engineering Society for a short course on CAD methods for VLSI design at University of Sao Paulo, and an invited talk at VLSI conference  in Porto Allegre, Brazil - 1990
  • Dr. France Preseren's Award for best Diploma Thesis in Electrical Engineering, University of Ljubljana, Slovenia - 1966

Publications

In 1953, Metropolis introduced well-defined criteria for the next step of a random walk on vertices in a graph, with vertices represented as states. Since then, new randomized search heuristics continue to evolve with examples of solutions to a wide range of combinatorial optimization problems. Metaphors from physics (simulated annealing heuristic) are increasingly augmented with heuristics that are inspired by metaphors from nature (behaviors of ants, bats, bees, bird flocks, cuckoos, etc). The abstract from a 2013 journal article ’Metaheuristics – the Metaphor Exposed’ provides a tongue-in-cheek introduction to the current state-of-the-art:

... a true tsunami of novel metaheuristic methods, most of them based on a metaphor of some natural or man-made process. The behavior of virtually any species of insects the flow of water, musicians playing together -- it seems that no idea is too far-fetched to serve as inspiration to launch yet another metaheuristic. ...
Experiments that compare runtime performance of algorithms based on metaphors from nature have led to inconclusive results all too frequently. Two important factors that can lead to such results: (1) problem instances may be "too easy" ; (2) statistical tests include too many censored observations.  As an antidote, we rely the metaphors of self-avoiding multiwalks and uncensored first-passage time stopping criteria, Furthermore, all presentation of experimental results strive to meet the standards defined by Tufte in Beautiful Evidence.

Most of our articles  have been indexed and are accessible  at Google Scholar,  dblp,  arxiv , researchgate, or academia . Original articles, along with companion slides of talks, stand-alone talks, patents, and co-supervised  MSc/PhD theses , are posted  for ready download from  www.openFB.org