Engineering as Collaborative Negotiation
Stephen C. Lu
sclu@usc.edu
Yan. Jin
yjin@usc.edu
The IMPACT Laboratory
University of Southern California
Denney Research Building, Suite 101
Los Angeles, CA90089-1111
Background
Computers have revolutionized the ways we perform engineering tasks over the past several decades. Before the invention of computers, engineering was viewed as a largely heuristic-based process where human experiences played important roles. Digital computers which perform analytical computing, geometric drafting and database searches have changed engineering tasks from heuristic activities into a highly technology-centered process. While more powerful computer tools are made available to engineers, the rapid increases of demands on complexity of highly coupled engineering systems have widened the productivity gaps in recent years. It is clear that the current paradigm of viewing engineering problem solving as a collection of computation, drafting and search activities has begun to show its limitations, and a new paradigm in engineering is needed in order to close this productivity gap.
Large scale engineering problems are inherently complex and highly coupled with many sub-tasks worked by teams of engineers and systems. In developing a new automotive, for example, a large team of engineers are involved in design and manufacturing processes. Different designers may work on different versions of the design at a given time. While some engineers may choose to modify their design, others may be off-line working on other urgent things. Using the conventional computer tools, engineers can simulate different design alternatives and find a locally "optimal" design. The collection of these local "optimal" solutions, however, is often not the best solution to the overall design problem. To prevent inconsistency among sub-solutions, designers must communicate/negotiate with each other sufficiently so that the needed information flows timely, knowledge about design is shared, and design activities are well coordinated. Conventional computer tools has reached its limit since they do not address the issue of negotiation or collaboration among engineers. In some cases, the use of computer tools actually discourages negotiation or collaboration among engineers, and consequently causes degradation of overall productivity.
Engineering as Collaborative Negotiation
Our observation has led us to propose a new engineering paradigm -- Engineering as Collaborative Negotiation (ECN). This ECN paradigm acknowledges the evolution of engineering problem solving from a heuristic through technical to a social process. Engineering is not only about finding technical solutions for sub-tasks. Rather, a large amount of engineering activities are related to communication and negotiation among engineers as a complex social process. The conventional computer tools failed to improve team or corporate productivity because they are designed to "optimize" sub-tasks as a technical problem, not to enable engineers to negotiate with others as a social process. Unlike a purely technical process in the past, a social process involves communication, coordination and negotiation among engineers and computer systems. The computer tools developed to date failed to provide adequate engineering support because they are not designed to support those non-computing and more socially oriented tasks such as negotiation and coordination.
Viewing engineering as collaborative negotiation suggests a new way of looking at computer tools, i.e., computing is for negotiation, not optimization. The ECN view opens new domains for computer applications in engineering and poses new requirements for CAE tools as well. Following is a partial list of requirements for tools to support engineering as negotiation:
- Leave rooms for others: A good negotiator always leaves rooms for others. From the technology viewpoint, This implies that CAE tools should attempt to provide intervals of decision values rather than a single "optimal" value.
- Sufficient conditions: One of the purposes of negotiation is to find a globally coherent solution. Therefore, CAE tools should provide the bottom-line solutions (rather than optimal solutions) for individual sub-tasks, so that engineers can maintain their negotiation positions to jointly achieve a globally optimally solution as a group.
- Least commitment: A way to maintain negotiation position is not to commit too early if you are not completely sure. In early stages, only incomplete models are needed since no commitment are needed on precise values.
- Direct dialogue: CAE tools should support direct and interactive dialogue among engineers and systems rather than the conventional one-way and iterative trial-and-error process. This requires bi-directional computer models.
- Early stage support: In most engineering processes, the opportunities for negotiation are more in the earlier stages than in the later ones. More negotiations at the earlier stages will have positive impacts on the overall performance and reduce the chances of expensive regrets at later stages. This requires early stage supports by new CAE tools.
- Redesign support: The result of negotiation often involves "regretting", i.e., someone has to re-do their tasks or retract previously committed values. We need CAE tools that can facilitate, or partially automate, redesign tasks.
- Explaining why: In negotiation, one needs to explain "why" for his/her proposals. The computer tools, therefore, must capture and maintain the rationales of each engineering decision to justify negotiations. Rationale database is no longer a side product but rather a main part of the CAE tools.
Supporting ECN: Research at the USC IMPACT Laboratory
The new ECN paradigm of provides both opportunities and challenges for the development of new CAE and IT technologies for engineering. At the USC IMPACT (Improve Productivity with Advanced Collaboration Technologies) Laboratory, we are conducting basic research to understand how engineers negotiate with each other, explore ways they should negotiate, and develop IT technologies and tools to support engineering as collaborative negotiation. Because we view engineering as a social-technical process, the research on supporting ECN requires a truly interdisciplinary effort. We are collaborating with social researchers and computer scientists to explore this emerging topic. Communication and organization theories, AI, and object-oriented modeling technologies are the foundations of our research. Our current research falls into the following three thrust areas:
Behavior and Methodology: Research in this area focuses on understanding how engineers collaborate in practice in order to develop new ways for them to collaborate better. The research topics include observation of team work behaviors, engineering organization modeling, community of learning agents for collaborative engineering, and team project and quality management. The results will be theories of human collaboration behavior that integrate and extend the existing communication and organization theories. The findings will be the theoretical foundation upon which we model engineering as a social-technical process, and will be used to guide the following research areas.
Collaboration Infrastructure: ECN will not happen in vacuum. It needs physical and organizational infrastructures, such as office furniture to facilitate local collaborative activities and enabling information technologies to support communication over distances. Our research in this area attempts to develop a uniform collaborative infrastructure that embeds both physical and information facilities and supports engineering as a social process. As a collaborative research project, we are currently working with companies to develop ideas and prototypes of collaborative workspace for the future. Based on Internet technologies, we are also developing new product and process models to support information exchange and agent-based structures for collaborative negotiation support.
Enabling Tools: Given collaboration infrastructure, engineers need IT tools for negotiation and collaboration. Our research in this area focuses on two types of tools. One is "ECN-compliant" tools for task solution that satisfy the ECN requirements described above. Supported by manufacturing industries, we are developing an adaptive and interactive modeling system (AIMS) that can trade model accuracy with computation time in a controllable way for the early stage design support. AIMS has been tested successfully for the automotive power train design. The second type of tools we are developing is intelligent agents for work process support. Instead of finding solutions for specific tasks, intelligent agents act as collaboration and negotiation assistants for engineers. We are currently working with an automotive company to develop an intelligent agent based framework to support collaboration in automotive inner panel design.
Summary
We have observed that engineers involved in large scale engineering projects usually spend more time to negotiate or collaborate in a social community than to compute their local sub-tasks in isolation. Providing advanced CAE tools for sub-task solutions will not impact on the overall engineering productivity greatly since the maximum share of possible improvement is very limited. Unlike conventional CAE approaches that view engineering as pure computing processes and ignore the social aspects of negotiation and collaboration, the ECN approach treat the negotiation as the main theme of engineering and opens a new way for computer applications in engineering: computing for negotiation. We believe that the paradigm shift from engineering as computing to engineering as negotiation will have an important impact on the development of future computer applications.
Authors Biographical Sketch:
Stephen C-Y Lu
Stephen CY. Lu is the permanent holder of the David
Packard Endowed Chair in Manufacturing Engineering in the School
of Engineering at USC, with the responsibility to direct the school-wide
research programs in manufacturing and design systems. He is
a tenured full professor at the Department of Mechanical Engineering,
the Department of Industrial and Systems Engineering, and a faculty
associate of the Department of Computer Science at USC. He heads
the IMPACT research laboratory, and serves as the Director of
USC's Center for Automation and Manufacturing Systems. He is
also the Director of the Asia Pacific Institute for Industrial
Leadership at USC. Before joining USC in January 1995, he was
a full Professor at the Department of Mechanical and Industrial
Engineering and the Department of Computer Science at the University
of Illinois at Urbana-Champaign (UIUC). He was the founding Director
of UIUC's Knowledge-based Engineering Systems Research Laboratory
(KESRL), and pioneered in the development of advanced computer
technologies for automation and integration of complex engineering
and manufacturing systems. In 1993, he was a Visiting Professor
of Mechanical Engineering at the Massachusetts Institute of Technology
(MIT), and the Technical University Berlin, Germany. The results
from his industry-driven R&D efforts over the past 15 years
include over $18.0M project funding, $3.5M research facilities,
4 industrial patents, 170 technical publications, 3 books, 17
Ph.D. (3 of them won NSF/NYI), and 18 MS students. Dr. Lu has
received numerous national and international awards and recognition,
including the 1987 Presidential Young Investigator (PYI) Award
from the National Science Foundation, the 1988 Outstanding Young
Manufacturing Engineer Award from the Society of Manufacturing
Engineers, the 1989 Outstanding Young Man of America award, and
the Xerox Senior Faculty Research Award. He was appointed as
a Distinguishing University Scholar by UIUC in 1989 for his excellent
contributions to scholarly activities. In 1993, Dr. Lu received
a Fulbright Scholarship in the Senior Research Professor category,
and the prestigious German Alexander von Humboldt Research Award
for Senior U.S. Scientists - the first ever precipitant from the
manufacturing community in the award history. Dr. Lu has been
awarded honorary professorship from three Chinese universities.
Yan Jin
Dr. Yan Jin is Assistant Professor of Mechanical
Engineering in the University of Southern California and the Associate
Director of the USC IMPACT (Improving Productive with Advanced
Collaboration Technology) Laboratory. He received his Ph.D. degree
in Naval and Information Engineering from the University of Tokyo
in 1988. Since then, Dr. Jin has done research on knowledge based
systems, distributed problem solving, organization modeling along
with their applications to computer aided design, computer integrated
manufacturing, collaborative engineering, and project management.
Prior to joining USC in the Fall of 1996, Dr. Jin worked as a
senior research scientist at Stanford University for 6 years.
Dr. Jin's current research interests include collaborative design,
product and process modeling, computational organization modeling,
and agent-based systems. He is on the editorial board of a new
journal International Journal of Emergent Mechanical Engineering
Technology. For more information about Dr. Jin, please visit
his homepage: http://bcf.usc.edu/~yjin.
 |