COMPUTATIONAL COLLOQUIUM
THURSDAY, OCTOBER 24th, 2002. TIME: 9:00am TO 5:15pm
UCLA FACULTY CENTER, IN THE CALIFORNIA ROOM
Please RSVP to Bill McKelvey (mckelvey@anderson.ucla.edu) if you plan to attend, so we know how many to plan for.
AND PLEASE DO PLAN TO ATTEND!!!
SPONSORS:
ABSTRACTS:
EDWARD CASTRONOVA: In March 1999, a small number of Californians discovered a new world called "Norrath", populated by an exotic but industrious people. About 12,000 people call this place their permanent home, although some 60,000 are present there at any given time. The nominal hourly wage is about $3.42 per hour, and the labors of the people produce a GNP per capita somewhere between that of Russia and Bulgaria. A unit of Norrath's currency is traded on exchange markets at $0.0107, higher than the Yen and the Lira. The economy is characterized by extreme inequality, yet life there is quite attractive to many. The population is growing rapidly, swollen each each day by hundreds of émigrés from various places around the globe, but especially the United States. Perhaps the most interesting thing about the new world is its location. Norrath is a virtual world that exists entirely on 40 computers in San Diego. Unlike many internet ventures, virtual worlds are making money -- with annual revenues expected to top $1.5 billion by 2004 -- and if network effects are as powerful here as they have been with other internet innovations, virtual worlds may soon become the primary venue for all online activity.
JOHN HILES: "Multiagent Systems, Connectors, and Networks: can we get them to do Blends?" The talk will describe our efforts to build agents and Multiagent systems (MAS) that benefit from current understanding of the control and coordination mechanisms found in living cells. Four stages of agent structure (along with the type of models or applications they can support) will be covered. A recently completed MAS produces story lines, logically connected sequences of scenes that tell a story without requiring predefined paths. Work underway now involves a combination of MAS, Connectors, and scale-free networks. This system will explore the possibilities for a MAS Blender - how would software construct Double-scope Integration Networks, or Blends?
JOHN KUNZ: "Using Computational Models to Support Research Questions and Methods." One of the most interesting objectives of research in engineering is to develop a scientific and theoretical base for engineering practices, even though practice normally is highly experience-based. A common method of science builds on finding theoretical mechanisms that explain behaviors found in the laboratory. A common method of engineering research observes practice and tries to find abstractions and generalities. At the Stanford Center for Integrated Facility Engineering, we have developed a general strategy of doing research in engineering that combines observations of practice with theory building and careful validation of theory in practice. Our research community has now used the method over nearly 10 years in a number of research projects. A unifying theme is the creation and validation of non-numeric computational models of engineering products, processes and organizations. This talk will describe several case examples and the broad research method.
JOHN PARKIN& ZIBA GHASSEMI: "Complexity Theory and its Implications for Architecture and the Design of the Built Environment - Agent-based modeling and its role in developing 'smart environments'." We will be presenting an overview of three major principals derived from complexity theory that are changing our approach to architecture and the design of the human built environment. This will begin to define a potential agenda for multidisciplinary collaboration. A specific focus will be on how designers could anticipate and build into a project from its inception, the idea that agent-based modeling would provide a powerful strategic decision making tool as the environment evolves as a complex system over time. These three principals are: Design as an Emergent Process: Our dynamic and interconnected context is causing a transition in the design process from a fundamentally formal, closed system, puzzle solving, tight fit mind set, to an increasingly emergent process. In this new setting a given building type may undergo numerous changes and refinements or event a major shift to a whole new higher order configuration - during the design process. This requires a whole new process platform. Design as a Transformative Experience: In a dynamic and complex setting a design process requires a much higher level of interaction, collaboration, learning and team building. This tends to make the process valuable in and of itself, rather than simply as a means to an end. This promotes the "learning organization" mind set and potentially puts a focus on agent-based modeling as a strategic decision making tool. Design as Adaptive Form: The emphasis is shifting to form - both physical and organizationally - that is intelligent and adaptive. The one time, perfect fit approach based on a snap shot of needs from a limited constituency is no longer viable. The focus is on building in intelligence and developing an integrated overall capability that provides the system with an "adaptive range".
BIOGRAPHICAL SKETCHES:
EDWARD CASTRONOVA, PhD, is an Associate Professor of Economics at Cal State Fullerton (business.fullerton.edu/ecastronova). His past research has a focus on the economics of government, incomes, and the evolution of culture. A lifelong gamer, he progressed from wargames and Dungeons and Dragons in his youth to titles like Civilization, Half-Life, and SimCity in the computer age. In April 2001, he stumbled into Norrath, the online world of the game EverQuest, and it immediately suggested itself as an object of study. His initial paper on the society and economy of Norrath was covered by the New York Times, the Los Angeles Times, the Wall Street Journal, the Manchester Guardian, Liberation, and countless weblogs and news sites; it was translated into Japanese and Russian, and became the all-time most frequently accessed paper at the Economics Research Network http://papers.ssrn.com/toptens/tt_ntwk_205.html, with more than 50,000 views and 14,000 downloads. Castronova is currently exploring some of the longer-run implications of shared virtual reality environments, and has begun consulting about them with a number of gaming companies, as well as the U.S. Defense Advanced Research Projects Agency (DARPA).
JOHN HILES is a Research Professor U.S. Naval Postgraduate School. Mr. Hiles directs the research program in Autonomous Software Behavior at the MOVES Institute, which is part of the U.S. Naval Postgraduate School in Monterey, California. Before coming to the school in 1998, he spent 27 years in the software industry in product development roles that included systems programmer, designer, manager and entrepreneur. His most notable products included operating systems, graphical user interfaces, and networks. In 1992 he moved into the computer game software at Maxis, the producer of SimCity. At Maxis and later in his own company he built game-like software that explained and explored real-world systems like industrial plants, professional occupations, and organizations. At NPS he established a curriculum and research program in Multiagent Systems. In 2001 his group successfully demonstrated a working software story engine for creating computer-generated stories. His research now involves software analogs to the molecular mechanisms that operate in the living cell to express and coordinate cell behavior. He applies these techniques to agent models of cognitive behavior.
JOHN KUNZ is Executive Director and Chief Scientist of the Center for Integrated Facility Engineering at Stanford University. Current and recent projects include developing computational models of engineering products and processes, organizational modeling, automated construction planning, developing symbolic process and instrumentation diagrams, and automated building code checking. Teaching includes a class in Symbolic Modeling in engineering, an engineering seminar series, various professional short courses in knowledge-based systems technology, advising the Ph.D. thesis research of many Stanford doctoral students, and supervising the research of professionals visiting the center. An important research and teaching theme is the use of different strategies to validate new formalized theories that have been implemented as computer models.
JOHN PARKIN & ZIBA GHASSEMI are principals in the design firm, Vortects, located in Santa Monica. Vortects is a firm that is pioneering an emergent process/ adaptive systems approach to design. John and Ziba have worked together on the design of numerous college and university projects including the USC Campus Center, the UC Davis Activities and Recreation Center (ARC) and Loyola Marymount University Recreation and Wellness Center. Ziba holds a B.Arch and M.Arch from the University of Arizona. John holds a B. Arch from USC and M.Arch from Columbia University.
Bill McKelvey
The Anderson School at UCLA
110 Westwood Plaza Los Angeles
CA 90095-1481
phone: 310-825-7796
fax 310-206-2002