2007 Complex Systems Proposal

Title:

Emergence and Complex Systems

Principal Investigator:

Russ Abbott

Organization and Cost Center:

Computer Systems Division/5814

Technical Category:

Information Sciences

Status:

Continuing

Associated SMC/NRO Technology Topics: 2006-1604

Priority 1

Advanced Computation Technology

2006-1609

Priority 1

Application of MEMs Chemical and Biological Sensors for Real Time Threat Detection and Identification

2006-1610

Priority 3

Architecture Design Center

2006-1611

Priority 2

Architecture-Centric Evolution (ACE) of NSS Systems

2006-1612

Priority 2

Automated Information Discovery and Access

2006-1617

Priority 2

Chaos-Based Communication

2005-1621

Priority 2

Complex Query Analysis for Disparate Information Sources

2006-1626

Priority 2

Automated Relationship Discovery

2006-1627

Priority 1

Data Structuring and Mining for ISR Systems

2006-1641

Priority 2

Information Fusion for ISR Systems

2006-1644

Priority 3

Integrated Tasking and Information Management

2006-1660

Priority 1

Models for Integrated Information Processing

2006-1746

Priority 1

Pico Satellite Joint Warfare Space Operational Utility

Program Funding: $125,000 Problem Description:

Complex systems are systems that exhibit what are called multiscale effects: phenomena appear at a macro level that cannot be predicted from a closed form analysis of a micro level. Only by actually running the system (or a model of it) will one observe these effects. This multiscale property is believed to be central to a very wide range of current research problems in fields ranging from physics (e.g., the so-called “collective effects” in condensed matter physics), chemistry, biology, computer science, and mathematics, to the social sciences. The appearance of macro regularities from micro interactions is known as emergence. Even though emergence is very poorly understood, commonalities appear across disciplines: the most frequently cited is that events tend to exhibit scalefree, i.e., power law, distributions. Two serious problems of special significance to Aerospace result from our lack of understanding about emergence. (1) Looking “upwards,” we are unable to build simulations within which emergence can be captured. Simulations that exhibit emergence

have existed for a long time, but these simulations are not able to recognize that emergence has occurred nor can they work with the emergent results. (2) Looking “downwards,” we cannot know what phenomena we are excluding when we select what is necessarily an arbitrary base levels for our models. Progress to Date on Continuing Projects:

During this past year we focused on three objectives. 1. We raised the awareness of the field of complex systems within the corporation by organizing a series of CSD Tech Forums on the topic. We have also begun to apply this perspective to the analysis of work in the areas of horizontal integration and systems of systems. 2. We initiated the idea of and are well on our way to producing an international symposium on complex systems and systems engineering. The primary objective of the symposium is to bring insights from the field of complex systems to the systems engineering and systems architecture arenas. The organizing and advisory committees include people from Aerospace, from other FFRDCs (MITRE, RAND, and JPL), and from industry and academia. The conference is scheduled for January 11 & 12 2007 at the RAND corporation. Participation will be by invitation only. To date we have recruited as participants a range of major contributors from both complex systems and systems engineering. A continually updated list of participants may be found here. 3. As a way of building momentum for the symposium as well as a way of raising the visibility of the corporation in this area, we have organized sessions, participated in sessions, and presented papers at a number of conferences, including: the Conference on Systems Engineering Research (CSER), The IEEE Systems of Systems Engineering (SoSE) Conference, The Understanding Complex Systems Symposium (University of Illinois, Dept. of Physics), The International Conference on Complex Systems (New England Complex Systems Institute), and INCOSE 2006. We have also had a paper (“Emergence Explained”) on the foundational notion of emergence accepted for publication in Complexity, one of the primary research journals in the field. Work to be Done this Year:

We will focus all our attention on complex systems. We propose four primary tasks. Formalize our characterization of emergence. Three types of emergence are generally recognized. •

Nominal emergence is emergence that is the sort one gets when putting together a system from well understood components in well understood ways. The result of such a construction does indeed have properties that the components do not have, but there is nothing surprising about it. This is what might be called traditional (and unchallenging) engineering.

•

Strong emergence is the phenomenon of finding what might be called new forces of nature at a macro level that didn’t exist at the micro level. This form of emergence is generally dismissed as unscientific, e.g., vitalism.

•

The intermediate form, called weak emergence, is the appearance of phenomena that can be explained once they appear but that were not anticipated. It is this form of emergence that characterizes most work in complex systems and that needs further explication. We have made a significant start on that work (and have presented our results at conferences), but the work is not complete. A onesentence summary of this understanding is that this sort of emergence uses epiphenomena to do real work, which is why it has been so difficult to grasp.

Elaborate and describe in more detail the consequences of this new understanding of emergence for issues related to simulation and modeling. This work involves two issues: (a) characterize under what circumstances the no-base-level issue becomes important for modeling and simulation and (b) begin to work out how to build models that are capable of recognizing emergent phenomena and of incorporating them back into the model. These two issues are both very difficult, and we expect at best only to make an initial dent in dealing with them in the coming year. Clarify the implications of the complex systems perspective for Systems of Systems and Capability-Based Acquisition. A fundamental insight of complex systems is that when a new system is created, it becomes part of a larger world, which can then exploit that system—for good or for ill. When we build stovepipe systems we often do our best (in some cases intentionally; in some case not) to minimize that effect. A complex system of systems perspective would urge us to build systems, to the extent possible (and to the extent consistent with security concerns—which, along with their differing reward structures, are major reasons this effect is so much more pronounced in the commercial sector than in the military sector), to be part of an infrastructure (of services) that grows increasingly powerful and sophisticated. This task will clarify and elaborate this perspective. Establish a complex systems focal point within The Aerospace Corporation and begin the process of helping people from different disciplines who are working with complex systems to share their knowledge and expertise. The field of complex systems is widely recognized and widely established in the scientific community. Significant numbers of multi-disciplinary institutes and centers have been established at Universities and research labs across the country and around the world. It is understood that the benefits of interdisciplinary work flows from the ability to apply tools developed in one discipline to problems that arise in another discipline when the problem of the second discipline resembles problems that have arisen in the first discipline—which is often what happens when dealing with complex systems. Status of Field and Competitive Position of Aerospace:

When Aerospace invented the notion of systems engineering, we initiated the study of complex systems as an engineering discipline. Unfortunately, we have allowed our lead in the field to disappear. Although there has been some recognition of the importance of complex systems at Aerospace (especially some of the work that has been done in agentbased modeling and the use of genetic algorithms), that recognition has tended to get lost in the crush of other work or to be limited to specific projects. This is in contrast to other FFRDCs, SETAs and other scientific and research organizations which have enthusiastically adopted a complex systems approach.

The systems we deal with increase continually in size, complexity, and the multi-scale range of phenomena that they span—from quantum physics to social interactions. We are in an excellent position to re-establish ourselves as the premier engineering organization with expertise in what is now recognized to be one of the most important areas of study of the 21st century. Success Criteria and Metrics:

We plan to produce two kinds of products. 1. Papers explicating our results for the first three tasks. 2. An organization within the corporation (either formal or informal as appropriate) to facilitate corporate-wide discussions of and cross-disciplinary interactions about complex systems. Capital Equipment Requirements: No capital equipment is required.

Other Corporate Investment required to Support this Effort: None. Project Scale: Information Integration Invent, Assess, Diagnose Percentages of this Effort: Invent: 75%; Assess: 25% SPOs Benefiting Directly from this Effort: AFSB/SBC, NRO/DDSE, SMC/ISS, SMC/ISG, SMC/GPH, SMC/MC*, NSG, CCO, and others