Dynamic Global Modeling: Overview | Demo | Technical Description of Modeling | Resource Directory

Technical Description of Modeling

Summary:

• Better quality forecasting is necessary and possible
• Basic idea
• Prediction requirements
• Comparison with other approaches
• Development steps

Better Quality Forecasting is Necessary and Possible

Rapid changes in the environment and in human relationships present a complex of problems that may threaten human values, institutions, welfare, even species survival. For a representative list click here. Within this complex, more specific challenges include those of avoiding, ameliorating, or resolving violent conflict, achieving sustainable development, and creating and maintaining a just society.

To gauge the severity of such problems and to cope with environmental and social changes requires a highly realistic picture of the present state of the world and of alternative scenarios concerning where it might be at various future moments, depending on what private and public policies are chosen. Such a picture must include anticipating and understanding changes regardless of level—international, national, or local—in a global context.

Achieving such high quality understanding about the future involves forecasting based on realistic information. Of course, global forecasting is not, by itself, a sufficient answer to those challenges; moreover, forecasting is always limited by systemic complexity and by imperfect information. Nevertheless, in meeting the challenges of change, planners and other concerned parties must know, before the fact and as reliably as possible, the consequences of taking no action as well as of various actions and policies.

At present, such forecasting is not nearly as good as it should and can be. The information used is over-specialized and too disjointed. The method of using information results in forecasts of unnecessarily poor quality. The consequences are plainly evident in the many failures to anticipate or effectively respond to war, acts of political terror, political disintegration; and in the chronic violent conflict, economic difficulties, urban anarchy, disease, and environmental degradation of local neighborhoods. These difficulties have happened in so many places on every continent of the planet that to name any few would slight dozens more.

To remedy this situation requires vast improvements over the present content, form, and availability of information but, more fundamentally, it requires that we give sufficient attention to what the modern scientific mentality offers, concerning prediction, especially concerning the prediction of social and other complex phenomena. See below, Prediction requirements.

Complimenting such public policy challenges is the intellectual challenge of achieving a major improvement in scientific understanding of social dynamics, which is the likely corollary of such effective global forecasting .

In sum, improved global forecasting offers the indispensable key to our acquiring necessary knowledge about change and its relationship to action or inaction; thus it is a necessary component of meeting these needs.

As the previous paragraphs suggest, meeting these practical and conceptual forecasting challenges can be aided by insights from modern scientific mentality. By modern, we mean, literally, post- 16th Century. Over the past 4 centuries science has accumulated a vast fund not just of specific physical knowledge but, also, knowledge about how to discover more knowledge. This knowledge is available for application to the global human societal-environmental system. It is time, overdue, for such an application to take place.

For additional discussion, see The Concept of the Center for World Studies, Inc , at this site, the Resource Directory: Materials Online.

[to top]

Basic Idea...

... can be summarized as follows. The global modeling concept of GVI will

• Furnish numerical predictions of behavior using a dynamic global model. See Overview.
  
  
• What behavior?
  • Events and changes throughout the world – in war, politics, economics, culture, climate, disease and disease spread, other environment, and other areas that significantly affect each other as well as human welfare.
    
    
• What kind of model? GVI model will represent the world as system of ?
  
  ...interactions (flows of information, energy, commodities, materials, persons, other living things)
  
  ...among distinct units (regions, nations, provinces, localities, groups).
  
  
• How far into the future?
• To be discovered as modeling work progresses.
• Most likely varies from one behavior to another.
• Start with 1-year lead time.
  
  
• What other limits, besides distance into the future, affect such a model?
  • To be discovered as modeling work progresses.
  • Most likely varies from one behavior to another.
  • Future changes in culture and technology look very difficult.

[to top]

Prediction Requirements

Summary

The essence of the GVI approach to global modeling is to prefer the approach of “Newtonian” science over the approach of “Aristotelian” science. (The quotes are because we are using these names to label two opposing viewpoints strongly associated with those two people, without necessarily claiming that they personally held viewpoints exactly as stated here.)

2 approaches to science, Aristotelian compared with Newtonian:

• Aristotelian approach is first to identify causes.
• Newtonian approach is first to predict changes.

Stated more completely, the Newtonian approach is to attend, first, to predicting change. Concern with causation is not excluded, but need not come first.

One implication of this is that dynamics and couplings (connections between one element and another) may transgress conventional ideas of causation. For example, with his laws of gravity, Newton transgressed the conventional causal distinction between “Earth” and “Heaven”. The modern version might be to transgress the conventional distinction between “social” and “physical” phenomena.

This Newtonian point of view leads to a priority of emphasis on the 3 more specific types of requirement shown below as key to a modern scientific global study.

• Dynamics:
  • Represent change over time (dynamic modeling).
  • Recognize system complexity (chaos, “memory” of history).
    
    
• Couplings:
  • Represent relevant factors & couplings among them.
    
    
• Data based testing:
  • Calibrate & test from historical past and other observations, using numerical, reproducible data.

More on prediction requirements

There is a fundamental distinction between two different conceptions of “science”: focused, respectively, 1) on identifying causation versus 2) on dynamic modeling, as the latter term is understood in modern physical science. The quantitative study of global human society has been almost exclusively focused on the former—identifying causation. Attention has also been given to a form of dynamic global modeling that contains some of the aspects among the three mentioned above—dynamics, couplings, and data based testing—but omits others, as dynamic modeling is understood in modern physical science.

This contrast between the approach of modern physical science and the current approach to quantitative global human society study presents a problem. The problem is, almost all the predictive achievements that we currently associate with science are achievements based on the modern physical scientific conception with its particular approach to dynamic modeling; thus, the very thing that experience would suggest is most fruitful to scientific understanding has been the comparatively neglected alternative in quantitative global human society study. One can scarcely avoid the suggestion that the current approach may be “missing the boat” in regard to its quite valid quest for a scientific understanding of global human society.

Nevertheless, the essential starting elements to a dynamic modeling-focused approach to global human society are, in fact, already available in the form of roughly a couple dozen separate fragments or strands, originating in present forms of global modeling and in social, computer, and physical scientific inquiries; strands that presently are more or less well established as separate inquiries in their own right. The task is not to invent from scratch but to combine these separate existing strands into a dynamic global modeling program formulated from a modern physical science viewpoint.

We turn, now, to separate discussions of the 3 prediction requirements.

Dynamics

A wide assortment of modeling resources, or tools, are available for dynamic global modeling. To see some of them, click here. Additional information can be found in the Modeling Tools entries of the Resource Directory at this site.

More significant than availability—the list evidently is long—is to use these elements in productive combinations. Modeling tools are like parts to a car lying on the floor of a garage. Probably they will not self-assemble into a functioning whole; assembly must be done by human intervention.

Couplings

Couplings are connections between pairs of elements in which the state of the first element affects the state of the second. They are illustrated by the list of problems and factors making up the speculative (but reasonable) matrix of couplings you can see by clicking here. Note that a working forecasting model would be based on data-derived and tested coupling values, not on the speculations appearing in these illustrations.

An example is illustrated in a second matrix, Some Possible Couplings—Asian Crash, 1997. This matrix shows how one factor—climate change, in the form of a return of El Nino weather—may have affected a second—economic instability, in the form of disrupted agriculture and urban work in due to extensive fires and soot in Southeast Asia. The coupling (the causal connection) is that the excessively hot, dry weather of El Nino made wood- and brush- lands more vulnerable to fire.

A further discussion of couplings is available at the previous cited link.

Data based testing

Data requirements are extensive, since they must represent a wide spectrum of human societal and environmental factors. A sufficient collection of specific indicators may run in the hundreds or thousands; however, a working forecasting model may by able to make do with a careful choice of far fewer. This can only be discovered though actual experimentation with, and testing of specific models.

The following are several different ways to get a more specific impression of global modeling data requirements.

• A detailed discussion of data requirements can be seen by clicking here.

• To see a sample listing of indicators at a more detailed level, click here.

• Still another way of viewing data requirements is to again look at the couplings matrices given in links, above; each factor or problem area is tacitly pointing to some bundle of indicators minimally required to track the named area.

• One can also examine the various entries in the Resource Directory, the section entitled Issue Areas.

Finally, the following links illustrate some of the data that have been developed, used, and tested in various models.

• Data on characteristics of nations, including number of military personnel year-by-year. To see a logarithmic-scale plot of number of United States military personnel, click here.

• For data indicating the conflict involvement of a group of nations, click here.

• Conflict involvement is of a type of a type called “events data”. For illustration of another use of events data, click here.

Data and analyses such as shown in the above examples are in common use in analyzing physical phenomena. The distinctive aspect, here, is their use in analyzing social phenomena. Global Vision, Inc. is committed to exploring these and other kinds of cross-applications from physical science methods to social prediction.

[to top]

Comparison with other Approaches

There are 2 kinds of comparisons:

• What is distinctive about the approach of Global Vision, Inc., compared with other, more conventional or traditional forecasting activities? To see some answers, click here.
  
  
• How does GVI propose to relate to other forecasting activities?

The answer to the second question is that exploring and making use of existing approaches and data are essential elements in the GVI approach. You can see this in many places throughout this page and its links. To see a list of some studies and groups that are relevant to global modeling, see throughout the Resource Directory. The listings in this directory define the minimal scope of resources on which GVI is drawing and with whom it is ready to work, to develop effective global modeling and forecasting.

Moreover, the Resource Directory is by no means complete. As time and information permit, GVI will put additional entries into it. These additions will define a corresponding enlargement in definition of relevant resources.

Finally, Global Vision, Inc. connects to traditional approaches in a different way, in that its own active participants span, so far as feasible given their number, the range of relevant areas. See the biographical pages under About Global Vision, Officers / Advisors.

[to top]

Development Steps

There are 2 parts to the GVI model development plan:

• Create a demonstration global model.
  
  
• Incrementally expand the demo model into a fully functioning global model, as identified on the Overview page. Below and elsewhere we will refer to this as the Reference Model.

Since the discussion of the demo model is much more detailed, we discuss the Reference Model first.

GVI Reference Model

Factors affecting plans to develop the Reference Model include:

• This step awaits the results of the demonstration global model project. We need to see how the more limited model performs.
  
  
• As mentioned above, to accomplish the integrated forecasting goal, the Reference Model may need to track—and certainly to consider—hundreds of variables describing different types of events and changes.
  
  
• The complete version of the Reference Model will require very considerable funding. See Summary of Project to Create ...
  
  
• There is no one point at which to regard the model as complete. Extensions and improvements will continue indefinitely.

These factors point to the following conclusion: Barring receipt of very substantial funding in a single gift, GVI will adopt an incremental approach to building the Reference Model.

In addition to the single large proposal ( Summary of Project to Create ..., linked above), proposals have been written involving various more specific aspects or applications. These are accessible, via the Materials Online section of the Resource Directory.

Demonstration Global Model

Summary

• Initial application to modeling numbers of military personnel. For results see Resource Directory, Available Materials Online, articles by Williamson and Bueno de Mesquita, and by Williamson (“New Considerations About Military Personnel Dynamics” , “Predicting Changes in USA Military Personnel Using Neural Network Modeling”).
  
  
• Second application, to modeling changes in GDP growth rates, is currently underway.
  
  
• Shell development is currently underway.
  
  
  • A shell orchestrates computational steps; and receives, stores, and outputs information.
    
    
• Incremental addition of further capabilities is planned.
  
  
• Modeling techniques include conventional linear regression and artificial feed-forward neural network.

For a summary discussion of the demo model, click here.

For a more mathematical discussion of the demo model,* click here.

For further discussion of artificial neural networks, click here.

To see the information flow of an artificial neural network, click here.

Expected Results from the Demo Model:

• Reactions and suggestions will be solicited from users.
  
  
• Idea and possible social value of GVI concept of global forecasting will be demonstrated.
  
  
• Will serve to test the various techniques and approaches, and to explore the limits imposed by data quality, by chaos, and by other aspects of system complexity.
  
  
• May help attract financial support for a larger-scale fully functional computerized, integrated global societal-natural environmental forecasting tool.
  
  
• Results of the project will be presented at scientific conferences on social-environmental forecasting.
  

[to top]

Home | News | About Global Vision | Our Mission | Dynamic Global Modeling | How You Can Help | Contact Us