Invited speakers

Dr Rich Terrile

Rise of the Machines: The Fiction and Reality of Evolutionary Computation for the Design of Intelligent Systems

This talk will outline the progress toward creating an Evolutionary Computation group at NASA's Jet Propulsion Laboratory (JPL) to address the design of robust complex systems. What started out as research for a science fiction story ended up as a Center of Excellence at JPL called the Center for Evolutionary Computation and Automated Design. Our group has successfully demonstrated the feasibility, applicability and advantage of using evolutionary computational techniques to address a wide variety of complex problems in several areas in engineering and science. In the three years since we started our group we have published over 20 papers and won an award for demonstrating human competitive performance in four areas of robotics, automated design, trajectory optimization and gyro tuning. Progress, however, was not always easy with skeptical customers, major workforce reductions, funding threats and erosion of talent to outside offers, politics and salesmanship were as big a part of the effort as was the non-traditional technology. This talk will also discuss how close the real promise of this field is to the fictional perceptions of literature and movies.

About the speaker

Dr Rich Terrile is an astronomer and the director of the Center for Evolutionary Computation and Automated Design at NASA's Jet Propulsion Laboratory. He uses techniques based on biological evolution and development to advance the fields of robotics and computer intelligence. He has a PhD in Planetary Science from the California Institute of Technology and has developed missions to Mars and to the outer solar system. Rich Terrile is the discoverer of four moons around Saturn, Uranus and Neptune and took the first pictures of another solar system around the nearby star Beta Pictoris. His other interests include planetary rings, planetary geology, evolutionary computation and the development of medical instrumentation for tissue identification during neurosurgery.

Rich Terrile is a private pilot, a SCUBA and ski instructor, a rock climber and caver, and was an astronaut candidate. He has interests in fiction writing, photography, cinematography and movies, and works as a technical advisor on several major motion pictures, including work with Academy Award winning directors James Cameron and Steven Soderbergh. Featuring in several science shows, he has been a regular performer at the Ice House and the Comedy Store in the "Improv Inferno" improvisation group.

Prof Stefan Voss

Hybridizing Metaheuristics: The road to success in problem solving?

Over the years we have seen a potpourri of ideas coming up under the umbrella of metaheuristics. Often a new idea or a new paradigm is claimed to be "the" idea by the inventor while others see them as useless in the first instance. However, once hybridized, things begin to "fly." Especially in population based metaheuristics, many researchers have followed this trend, and so has EvoCOP: "Following the general trend of hybrid metaheuristics and diminishing boundaries between the different classes of metaheuristics, EvoCOP has broadened its scope ..."

In this presentation we highlight the notion of hybridizing metaheuristics by emphasizing simple and yet effective ideas. Of course this does not mean simply to use a certain method and marry it with local search. Instead we propose, for instance, to use a hybridization of reactive tabu search (RTS) and simulated annealing (SA) where the RTS may be viewed as an intensifying mechanism and SA is used for diversification. Moreover, we study the use of the pilot method when hybridized with well-known metaheuristics such as variable neigborhood search (VNS) and greedy randomized adaptive search procedure (GRASP).

The pilot method as a metaheuristic is a tempered greedy method aimed at obtaining better solutions while avoiding the greedy trap by looking ahead for each possible choice. Repeatedly a master solution is modified; each time in a minimal fashion to account for best choices, where choices are judged by means of some heuristic measure (based on a separate heuristic result, the pilot solution). We use an example from telecommunications network design to exemplify concepts. Designing a telecommunications network embraces the choice of the equipment as well as determining the routing. The aim is to find a least cost combination or network design. We present results for approaches which hybridize the pilot method and the VNS / GRASP metaheuristics. The problem instances cover a large variety of networks and demand patterns, including protection planning.

About the speaker

Prof Stefan Voss is Director of the Institute of Information Systems at the University of Hamburg, and was previously Head of the Department of Business Administration, Information Systems and Information Management at the University of Technology Braunschweig. He holds degrees in Mathematics (with Astronomy) and also Economics from the University of Hamburg and a PhD and the habilitation from the University of Technology Darmstadt, and has held guest professorships at the University of Graz, Austria, and at the University of Valenciennes, France. In addition to metaheuristics, his current research interests are in quantitative / information systems approaches to supply chain management and logistics including public mass transit and telecommunications. He is author and co-author of 14 books and 180 publications in numerous journals. Stefan Voss serves on the editorial board of several journals including being Associate Editor of INFORMS Journal on Computing and acting as Area Editor of Journal of Heuristics. He regularly organises workshops and conferences, and acts as a consultant to several companies. Stefan Voss holds a private pilot license, and with his partner, he participates in dance competitions as well as judging dance events.