THE FUTURE OF HIGHER
(All chapters are intended for continuing revision)
Volume II - Chapter Five
(Last revised, Aug. 13, 2007)
LARGE-SCALE COMPUTER SIMULATING, MODELING, MAPPING
THINK BIG! For human survival research--and education for all--humanity needs global-scale research systems and tools in cyberspace/ hyperspace to control pollution, fight inflation, provide justice, and to warn of new dangers and threats that limit efforts to provide education for everyone in the world. (See Brewer 1979). Essential components include much larger and more sophisticated computer modeling, used in a global system in which computers and communication networks are combined with data intensive simulations and enhanced data storage systems. It is then possible to construct models of complex systems to explore dynamic implications of decisions while using a wealth of stored data. For example, the `Savannah environmental data atlas' has included airborne multispectral scanning data, aircraft video, photography (images as early as 1930) bibliographical data, gridded map data and much more, achieving a photomosaic of the entire site. What about one for the entire planet's ecological system? See: <www.outsights.com/systems/welcome.htm> and for its educational system? The ManyOne Foundation's digital earth project was in 2005 beginning to bring together vast simulations for 3-D viewing.<http://www.manyone.com/>
Takisihi Utsumi (2007) propses using a distributed computer simulation system to focus on the sustainable development of a socio-economic-energy-environment system. For example to see how researchers in Japan, the US, China, Russia, Kazakhstan, and the other relevant countries.canl construct simulation models that will be interconnected through broadband Internet to form a GloballyDistributed Socio-Economic-Energy-Environmental Simulation System. In Forman (2004) several experts explain how simulation games--`worlds in a box'--allow learners to do what lectures don't. Simulations enable learners to create a world where they can have all kinds of experiences that are not otherwise possible,. especially for the natural, physical and social sciences. Learners work and think in teams using their varied skills.(More in (3.3). Foreman (2004a) reported how simulation-gaming projects "anticipate advanced online learning worlds that can be dedicated to different subjects, populated by sing users and teams, and pedagogically structured for deep and rapid experience-based learning." Educational simulations will be able to adjust themselves to learning styles, he said, and to skill levels "to accommodate" learning teams "from all parts of the world."
A European river basin management system--universities and industry cooperating--was designed "to integrate geographic information systems, large data bases, simulation models, expert systems and tools for visualization and sophisticated graphics displays" (McLeod 1993) such as maps of the spread of underground water. Since no single software system could suit all river basins, users could select models and tools needed for the management of their own system. They could exchange models or add new ones in dealing, for example, with contamination. Bringing various dynamic simulation models together, the aim was to include sub models of components describing rivers and reservoirs, coastal marine areas along with sociopolitical factors to provide a more realistic view of problems and options. This would make possible a comprehensive approach to river basin management.In time a global water management model for the entire planet could link all such models and systems.
Climate system models also include many complex subsystems, such as atmosphere, ocean, cyrosphere (ice where data from tens of thousands of years may be required), land masses and biosphere and their interrelationships. Because of technology and data limitations, each subsystem is often researched alone and conclusions are thus based on inadequate data. A longer-range prediction must include many more factors and may require data from over thousands of years. "In a fully developed climate model the variables of all subsystems need to be predicted as well as the interactions between the subsystems." (Gueltzow 1995). As if not complex enough, all kinds of human factors must be included also, such as the effect of auto emissions on the ozone layer. So the problem of dealing with vast amounts of data in many complex social subsystems can add greatly to the complexity of a research design.
According to the World Meteorological Organization, the Global Telecommunications System of the World Weather Watch is an integrated network to interconnect meteorological centers. It exchanges satellite data as part the World Climate Research Program that is also sponsored by the International Council of Scientific Unions and the International Oceanic Commission (IOC) of UNESCO. It seeks to develop global models of the present climate and tries to predict climate variations on a wide range of space and time scales. The management and control of such vast amounts of data is an unsolved problem. Technology adequate to store, manage, distribute, retrieve and use such comprehensive data does not yet exist. Supercomputers are required for larger and larger simulations but the weather system is so complex that even networks of supercomputers do not yet have the power needed. Guetzlow's summary progress report said that "the total requirements for computer power exceed present capabilities by a factor of 100-to-1000."
Japan's `earth simulator' (Wired, March 2003) is a major step ahead as it collects "data from around the globe to predict planetary changes like el Nino, earthquakes and global warming."
Global society needs more adequate conceptual tools to deal with complex problems that tempt the world's leaders, as in education, to simplistic solutions. It is going to be possible with more powerful tools now being developed. Utsumi (1996) proposed combining existing models to create more holistic models for examining various scenarios that might resolve global crises. It is now essential, he says, for "scientists and simulationists to cooperate--interdisciplinarily and internationally--in the establishment of an interactive modeling system for global policy analysis." This would involve the establishment of "suitable ground rules, reasonable assumptions and a common set of premises." He speaks of interconnected global-scale computer simulations as complex as the neural network of a global brain. (2.1). Computational simulations, models and modeling can also now be shared and accessed on the Internet and World Wide Web.
Atmospheric research (Cleveland 1993) "has been driven well by progress in modeling global climate and regional ecology," promoting feelings that researchers are within reach of the ultimate goal of models of the whole earth system which is of unimaginable complexity. The 1990s saw scientific conferences on plans for simulating the earth system, social and economic systems, medical decision-making, virtual communities, business reengineering, energy systems, and more. A whole earth system will require enormous increases in computational speed and storage. )
Shubik (in Brewer 1979) described four kinds of models: verbal, mathematical, graphic picturing, and computer/digital. Global planning will require an extensive hierarchical system of models, simulations, games. Utsumi proposes reaching out to the real environment and connecting to its sensors and effectors; not a complex program in a giant computer, but a complex of programs in a network of computers.
Writing an editorial "Simulation in the Service of Society" for professional simulationists, McLeod (1997) asked colleagues for suggestions for modeling for sustainability. A reply from Australia told about the ECUMENE (environment, population, development) project there, a multidisciplinary examination of environmental futures for fifty years ahead. (10.2). It has had several models running to examine sustainability, and the impact of many items like refrigerators, trees, energy use. It was hoped that this modeling could be tied in with "a suite of global models". Meanwhile its best international links had been to projects in Holland and to the QUESTIONS project (Syndromes of Global Change) at the Potsdam Institute of Climate Impact Research. The latter had defined about twenty syndromes--that is combinations of social, political, economic and environmental factors--which exist in a designed local space.
"Developing the complex computer models that are necessary for effectively managing human affairs through the next century," Maxwell (1998) proposed, is going to require "high performance collaborative modeling." An illustration has been an `Open Modeling Environment' that makes use of distributed computing, multiple space-time representations, graphics and linking of simulations. Its use does not require knowing much about computers and computer programming. So it can be available to all who are studying ecological systems. Such simulations are important because "collective global economic activity is altering the self-repairing aspects of the global ecosystem." The speed at which human beings can negatively influence the environment far exceeds the ability of scientists to anticipate the effects over long periods of time and over large areas of the globe. So computer simulations--such as geographic modeling of ecosystems --become urgently important.
Tremendous amounts of computer power are required, however. The kind of modeling which Maxwell proposed is possible by linking many powerful computers. Each computer--or system of linked computers--can be assigned a part of the project. This process has not been used much in the life sciences, Maxwell reports, because the creation and debugging of programs for parallel computing is very complex. Also, it is very time-consuming for a scientist to learn new computer languages. So Maxwell has described a geographic modeling (SME) environment that links parallel computers with "a generic object database." This makes it possible to put together "reusable model components in simpler and easier ways. This design enables many networking scientists, in various locations, to work together in creating and using what Maxwell calls "a global-scale ecological/economic model." This will make possible more significant large-scale joint work among "a distributed network of collaborators," using a variety of computer systems.
Jay Forrester (1999) in an address on systems dynamics modeling, said that people commonly think in a linear way in which a problem leads to action and then to a solution. In fact, we operate within closed feedback loops in which "in which the human mind is not able to anticipate the resulting behavior" since it is unable of understanding how the complex parts of a system will interact to produce unexpected behavior over time. In the next fifty years, he said, we can anticipate an enhanced understanding of our social and economic systems.
Maxwell and his colleagues "expect that effectively managing human affairs" in the 21st century is going to require bringing together new information technology with spatial data (dynamic spatial modeling), high performance computing and virtual reality. Thus a 'Virtual Earth' infrastructure can be created "to provide a "dynamic, virtual representation of many aspects of the planet's ecosystem, economics, and social/cultural/political systems." This is possible because of the ability now to "capture, store, process and display an unprecedented amount of information about our planet and a wide variety of environmental and cultural phenomena." This can include high-resolution satellite imagery of the entire planet, digital maps, and economic, social and demographic information. Thus a more holistic understanding of earth system dynamics will be possible. It will facilitate responding to manmade or cultural disasters. As the system evolves 'what will be possible in 2005 "will seem primitive in 2020.".
This book has no space for the technical details of what is now coming into existence, much less for estimates of what is going to be possible in the next twenty years. What is important is the maturing of systems adequate for large-scale modeling "at reasonable cost and speed." Linked with data bases, Maxwell reported, "one has a powerful, easy to use, spatial modeling environment. . . which can result in a fundamental paradigm shift in complex systems modeling." He believes "that effective management of human affairs throughout the next century will require. . . complex and reliable computer models."
Mesarovic (1996) said that great care must be given to the human factor in modeling. The term `human dimension,' he said, demonstrates the inadequacy of the role that may be assigned to people. One would not speak of the `ocean dimension' of climate change. Similarly humankind is a subsystem of the global climate system. The term has usually been limited to the effect of society on the environment and the impact of environmental change on people. Humans, however, are not observers from outside the climate system. Also, the `input/output' paradigm for observing the role of humans in the environment originated in the old Newtonian paradigm. It assumed that scientists would be able fully to predict the future once enough data and knowledge were available. Modeling from that theory "leaves no room for uncertainty or indeterminism. Mesarovic says it can be useful in the short term but the findings are usually wrong "when the predictions are really needed."
An alternative is modeling which has its analogy in biology and human nature studies. "Symbolic human nature modeling provides a framework to take into account nonnumeric non-measurable aspects of reality." The need to represent phenomena from different scientific disciplines--in the modeling of global change--leads to the concept of `integrated modeling' in which all relevant disciplines are taken into account. But, Mesarovic says, there are serious shortcomings to developing separate models in different disciplines and then attempting to link them together. When the sub models are complex it is hard to be sure that the integrated model is true to real life.
The alternative to such hardwired linkages of models--for use in education research-- is to develop a "goal-seeking framework to incorporate the human within the model," and an adequate description of that would be too complex to report here. However, it would lead to ever more complex models, involving many disciplines to cope with uncertainty and complexity. It would be analogous to the human body, Mesarovic says, in which the various parts cannot be separated from the total system. In a global climate system, for example, it is a mistake to assume all the regions of the world will converge.
So the National Academy of Sciences prefers a `bottom up' engineering approach rather than a `top down' integrated model system. It is not enough to model resources and the supporting environment without including the social and human domain with all of its uncertainties. (i.e., see <http://www.cares.missouri.edu>) This means that identifying realistic policy alternatives cannot be accomplished with computer models alone. They may provide useful information and guidance, but "decision-makers face a more complex process in creating effective policies." Mesarovic therefore proposes "the cybernetic viewpoint in which all available tools and information are used." He illustrates with the GLOBESIGHT system that "plays the role of a consultant." Its Information Base contains historical data and its Modeling System includes many procedures for exploring the consequences of policies. The Issues Base can preserve and compare results for future reference and for the extension of analyses. In this system the "human and the computer walk hand in hand, step by step along alternative, feasible future paths." So the scenario that emerges is not one that was known in advance. From this perspective Mesarovic's GENIe project had a holistic view of global change based on a multidisciplinary foundation. It blended physical and social sciences and uses information technology "to assess a plethora of uncertainties intrinsic in the integrated assessment of alternative global futures."
How can scientists and scholars be helped to learn of research in other fields that is relevant to what they are doing? It is difficult to keep up with all the research in one's own specialization. Sandia laboratories (Steinberg 1997) has been working on a "map of all research" which would "reveal connections that were previously hidden." Three million papers were being examined to analyze connections and present them visually in a three-dimensional landscape. For example, a map of a `mountain range' showing research in biology connected "to an area in Physics by a narrow ridge." This suggests the potential of similar mapping in social science (13.6), interconnected with models of cities, and ultimately all of human society in order to interconnect many kinds of research and planning.
Global geophysical data can be "a matter of life and death." (Steele 1997). Yet despite having spent ten billion dollars a year on overhead reconnaissance, Steele said, maps do not exist for a high percent of the world that are accurate enough for coordinated peacekeeping activities, to effectively monitor economic activity, or to understand "the environment across time and space." Mapping and spatially arranged data bases--possible with one-meter imagery--are important to show "how the world actually is," rather than how scientists and the rest of us think it is, (Morton 1997). Many sorts of information are far more effectively transferred in iconographic and pictorial form. In a rich-data environment, Morton said, charts and graphs convey much information in a concentrated space. Where a graph or diagram is an abstraction that can distort reality, a satellite image "presents a vision." He illustrated--how this can help anyone get a clearer understanding of many aspects of reality--by pointing to the clouds crossing the map on the TV weather report. "Once a qualified meteorologist has made sense of it, an ordinary person can participate in the analysis...and (could) make a counter-analysis." As humans got a new view of the earth from outer space, satellite mapping can give us "a new view of (political) borders" andof current events in the news, "...not a view from nowhere, but a view from everywhere for everyone."
Another role for modeling in socio-political applications has been proposed (Mayer-Kress 1994a) to deal with crises. "A new paradigm is introduced into conflict management" by the use of intelligent work stations, the global Internet and the theory of nonlinear dynamics, chaos, and complex adaptive systems. Putting these together provides "a basis to solve global problems in a complex world, A global network of highly connected and powerful computational models," he says, "can then find solutions to problems that could not be solved in any other way." Much as young children learn to solve complex problems through imaginative play, he has proposed that "groups within the global community will use imaginative computer simulation/games to find novel solutions to global and regional problems." Ironically one way to experiment, enlarge and develop these global-scale tools is to play games with them. (2.7.1) In other words, simulations can unofficially try actions that officials are not yet able to consider. From now on it will be essential to use computer modeling for making important decisions, models that incorporate much more knowledge about people and institutions. Utsumi would continue to enlarge global-scale models to cope with more and more complexity by interconnecting such whole world models as the FUGI model, developed by Akira Onishi at Soka University in Japan, and LINK at the University of Pennsylvania. (12.4). Many helpful questions can be raised, for example, when the unexamined models which politicians have in their heads are compared with statistical models. They can clarify differences, disagreements and possible changes. Yet decision-makers have been slow to create or use such tools designed precisely for the tasks they face. Instead they have sought to adapt tools already prepared for business or the military.
Not all modeling and gaming should be immediately practicalfor education research. Sometimes audacious, impractical alternatives should be examined when the `right' ways of doing things are leading to disaster. Computer modeling, as interconnected supercomputers become more powerful, can explore the `impossible' and break the logjams that obscure solutions to crises. They can explore dreams and visions and can stimulate creativity and imagination in education in a way analogous to the capability of the human mind, as seen in the intuitive flash, the connection of previously unrelated thought processes. The value of global modeling and simulation tools and systems, as discussed here of course, will be determined by their success in helping researchers ask the most fundamental questions in relation to the most desperate of human problems.
Arbib (1997) reported on his examination of efforts to use `world models' to predict ecological and economic changes affecting the entire population of the world. Although not as adequate as weather prediction models, he says, world modeling has already significantly effected the way many people view the world's social problems. World modeling is increasingly valued because "we are heading toward a worldwide crisis for the human race," illustrated by potential environmental catastrophe and therise of terrorism. Something must be done, world leaders say, but what? Can computer simulation be effective in exploring different courses of action to discover the consequences? Rischard (2002), for example (3.10), how biotechnology and biodiversity have become pressing issues as a result of "the spectacular explosion of discoveries." Yet, he points out, "biotechnology is still in its infancy, "bringing up many more questions and possibilities than answers and outcomes." It is an area where global rules will be needed.
Arbib found that he could learn much from the models prepared for the Club of Rome. However, the predictions of such primitive models "are to be trusted much less than the weather forecast for a week from today." It is crucial to put much more effort into model validation. A good model must be based on good data. Those used in outer space projects "are reasonably good because they can be based on fairly accurate data, gathered by satellites." A continual photographic record would be useless if it were not for computer technology that can examine and manage huge amounts of the data in ways that would be difficult for humans. Still, powerful new tools are needed to model whole world, including all the social and human factors.
Those who study the weather, Arbib said, can devise models that are increasingly more realistic. This is true because, using computers, they can process far more information "from the huge data base they have on magnetic disks." Also they can test their models the next day to see how accurate the previous day's forecast was. They can look at data from all over the world for that day; but even with such powerful models and vast information they often fail. So new and larger-scale approaches are need to model helpfully the entire ecology of the planet. See: Freedman 2003)
The first Club of Rome model, "Limits to Growth," sought to examine the interactions of population growth (2.11.4), pollution (2.10.6), food production and natural resources depletion. Then it played out various scenarios for preventing disaster to evaluate possible policy changes. Unfortunately, Arbib says, it did not deal with many crucial factors, such as regional differences. So the Club of Rome decided that the world had to be modeled as a system of ten regions. This too, however, turned out to be "far too simple."
Models can be improved "by building up a massive data base of very accurate information." The modeling of economic and ecological trends, however, are at best valid only for a few years into the future, Arbib says. Better modeling is needed in addition to more hard data. A model of a spaceship is more accurate because it can be based upon Newton's laws of motion. World models, however, require the development of "tools that are still controversial." More adequate models must be continually revising themselves and must also take account of the totally unexpected. Model validation, including a continuing examination of the failures of past predictions and why they failed, he says, "is a never ending process." It must be closely related to the real world.
Models of five problem areas at the International Institute for Applied Systems Analysis (IIASA) have been linked to study the effect of emissions upon the environment and climate change. The database used in 1996 covered more than fourteen hundred technologies, ranging from energy extraction to energy use. An economic database estimated the costs for eleven world regions. An energy database provided data for evaluating the use of different types of energy within alternative scenarios. Four models were linked: a macroeconomic energy model, a systems engineering energy model, a transboundary air pollution model and a macroeconomic food and agriculture model. The designers of this project found that mathematical modeling provided a powerful tool for forecasting and controlling the behavior of large-scale economic and environmental systems.On simulation for learning, see: <http://projects.edte.utwente.nl/pi/Book/Contents.html>.
A `game' approach has been found adequate to deal with systems that are driven by `players' (agencies, corporations, government regulations. IIASA <http://www.iiasa.ac.at/> works on and interrelates increasingly larger models and projects. Biological evolution has been found to provide an analogy for examining economic and population dynamics. When more important variables are added, such as bio-geochemical cycles in the climate and radiation, even more power will be needed that will be adequate for storing an increasing number of models and their data for future comparisons.
The future of gaming, modeling, digital mapping and simulating seems almost limitless. One simulationist has said that current modeling may be compared to making mud huts, where future modeling will be more like building space-age cities. Fishwick (1997) said that simulation is a fundamental discipline for studying complex systems by constructing digital worlds. Its future lies in relationships with related disciplines. He used the term `extension' of simulation for combining it with artificial intelligence, neural networks, computer graphics, artificial reality and more, including fuzzy logic. This can make it possible to "embrace a new way of doing science through simulation."O'Looney (2000) discusses the use of GIS (Geographic Modeling Systems) in e-governance. <http://www.cpsr.org/conferences/diac00/proceedings/WORK/John.htm>
It is hoped that artificial intelligence will be useful in enlarging simulations to create multimodels, he proposed, in which non-specialists can use natural language and not just mathematical language. Modeling a gasoline engine is quite different from modeling human thought. Expert systems, he says, provide a good illustration of codifying `metaknowledge' about a system and not just low-level aspects of it. The idea of fuzzy logic "is to approximate human decision making using natural language. Simulation is the only way to study chaotic behavior and to understand the implications of "bottom up investigation complexity" as in theoretical biology.
Simulationists, Fishwick said--with the aid of parallel and distributed computing--must start thinking of themselves as "digital world builders instead of workers building isolated models useful to a few people." They can make their research findings clearer for use by more scholars through graphics and soon through virtual reality (VR). As graphics--visual feedback--reduces the man-machine communication bottleneck, he said, immersive interface technology explores how the machine and human users "can be more harmoniously coupled" so that the human has the feeling of being more immersed in the computer environment rather than being separated from it.
While VR "represents the future of simulation" there must be improvements in the human-machine communication. Not only the researcher, but the user of a simulation will be aided, Fishwick said. He illustrated by proposing an automobile manual that presents not only text but also a simulated experience of driving the car. "The key point is that simulation is then a natural part of the information access and not merely a generator of information." The tools are in place today for embedding simulation models and their outputs into the World Wide Web, he said. Web based simulation will become the predominant mechanism for running any simulation. By "combining computer simulation with other disciplines," Fishwick concluded, "we obtain extensions that are used to better solidify the current foundations for simulation methodology." At the same time, simulation becomes a foundation for work in many other fields of research.
"What do you want the world to look like in another 50 to 100 years?" asked a participant at a Frontiers of Science symposium. Sponsored by the National Academy of Sciences, it was attended by a hundred of the brightest young scientists from a wide variety of disciplines. One group of them, on the Internet, examined the idea of preparing a model of each country (18.1) which would bring together better information on all of its problems and crises. It would include comprehensive data on population, health care, nutrition, trade, agriculture, natural resources, energy, pollution, military technology, debt. . . and everything else! To link all such national models together to see how the worldor education should look in a hundred years would require a comprehensive model of the world as it is now. Critics of current modeling point out, for instance, that the problems of poor nations are almost totally left out of economic modeling. "If one sees the function of global models primarily to challenge present worldviews, what is needed are grand designs (Dror 1994) "to examine the dynamics of situations and of environments, of uncertainties with some "skepticism on the present stage of the art." The promise of modeling justifies giving a high priority to its enlargement. (Barney 1991)
Finally, what about models to examine possibilities for the the global research system itself? In the last chapter we asked about computer models of individual scholars with Howard Gardner's seven kinds of intelligence. Isn't cyberspace now the place for modeling alternative kinds of co-laboratories which use collective intelligence; of new models of professional and disciplinary associations and how they might better define their research role? Could simulationists explore some entirely new research structures which would focus on collaboration in in search of ways to cope better with humanity'sglobal education needs? Some such possibilities are discussed in the next chapters.
CRESS (The Centre for Research on Simulation in the Social Sciences) has been building a data base on simulation research potential with links around the world: <http://www.soc.surrey.ac.uk/research/cress> What does this suggest for lifelong education, not only for better research but also the creation of more adequate models of a global education system itself?
The Future of Higher (Lifelong) Education: For All Worldwide: A Holistic View