Issues Regarding Computer Modeling and Simulation
Computer models and simulations are used for state of the art research, testing and planning in areas ranging from the design of commercial jets to the simulation of crowd movements in new sports complexes.  While these models represent crucial research tools and instruments for analysis, models have also been known to misrepresent data and present false information as well.  This page seeks to explore this topic by first examining some of the better known uses of computer models and simulations, then looking at the limitations of computer models in general and the impact of computer simulations on the general public.  Specifically, the page is divided into the following four articles:  At the bottom of each article, links to information about the subject matter discussed are provided.  Enjoy!
 
Weather Modeling
 Modeling the Universe
One of the most important uses of modeling and simulation technology is in the area of weather pattern simulation and forecasting.  A wide variety of weather models are at the disposal of meteorologists, many of which are available online.  These simulations promise accurate predictions of weather forecasts based on the vast array of satellite, radar, and other data available to weather forecasters.  Millions of citizens watch the product of these simulations on the TV news each night on the news.  NASA meteorologists depend on models to predict weather conditions for space missions.  Oil companies rely on accurate weather forecasts via computer models to insure oil rigs worth billions survive the dangerous tow to their destination.  Certainly the effects of such computer models are far-reaching. 

This, of course, begs the question of whether we rely too much on such models to predict weather patterns and whether the forecasts they generate are indeed accurate.  In fact, weather models tend to be quite accurate given the amount of complexity of the systems they are intended to model.  Several factors contribute to the reliability and accuracy of such models: 

  1. The ability to test and verify the predictions of the computer models against the actual events
  2. The capability of satellites, radar, and other tracking means to gather large amounts of data for computer models to interpret
  3. The speed and storage capabilities of modern computer systems
  4. The availability of a large amount of historical data regarding weather patterns
The result is an accurate weather prediction and a confirmation that computer models, when developed and used in the appropriate manner, can be an invaluable tool in analysis.  Weather forecasts once considered barely reliable for a single day are now trusted to be accurate for 3 or more days, and the accuracy of weather prediction continues to improve.  When arguing the pros and cons of using computer models to simulate the real world, proponents invariably point to weather prediction as a demonstration of the benefits of such tools. 

More information regarding computer models and weather forecasting in general is available in the USA Today article Weather Forecasting.  It contains a good amount of information on the topic of weather forecasting models in addition to topics regarding the subject of weather forecasting in general. 

Links: 

 		 Computer modeling and simulation is certainly not limited to the somewhat uninspiring roles of business models and weather forecasting.  In perhaps its most extravagant role, computer modeling is used to help solve mysteries on a literally galactic scale. 

Recently, physicists have been introducing computer models to help verify the existence of elusive cold dark matter in the universe.  Supercomputers and computer modeling are to assist in the replacement of nuclear testing with what is claimed to be the computer simulated equivalent.  Researchers at Cornell are amongst the many scientists exploring the evolution of the universe with the aid of computer models.  With more powerful microprocessors and storage devices emerging on a continual basis, scientists and engineers are starting to embrace the computer simulation as a means for modeling the complex systems of the real world. 

The difficulty with assessing the validity of such systems lies, in part, with the fact that modern science and technology is far out of the reach of the average individual.  Whether a computer model can replace a full-fledged nuclear test is a question of importance to many, but which can only be answered by those intimately familiar with this advanced technology.  However, there certainly are guidelines by which one may judge the validity of any such simulated system.  Some of these guidelines follow: 

  • What real-world outcome, if any, can the system being simulated be judged against?  (For example, does a simulated nuclear bomb blast draw comparable data to the actual blast given the available combination of inputs?)
  • What theoretical basis or construct is used to formulate the model?  (For example, does a model of weather patterns rely more on emulating actual weather systems or on statistical analysis with the aid of historical data?)
  • What assumptions does the model make in an effort to reduce simulation complexity?  Are the assumptions reasonable?  (For example, is a structural model of a building that presumes a perfectly rigid building foundation overly presumptuous?)
When faced with highly advanced topics such as particle physics, we are often left at the mercy of scientists and engineers to determine the validity of such models.  However, acting as a professional in the field, one is obligated to measure one's own model against these criteria prior to obsessing over the masses of data such models can generate.  Perhaps the evolution of the universe can be accurately modeled within the confines of a supercomputer.  However, given the criteria stated above, making the transition from an interesting concept to a verifiable representation of the universe as we know it is no trivial task. 

Links: 

 
Limitations of Computer Models
While computer models have been very successful at assisting scientists and engineers with solving complex problems, the power of the computer model to simulate reality has often been misrepresented as well.  As pointed out in the Wired magazine article, "Simulation has its limits", the computer model possesses the dangerous capability of affording one a false sense of security.  The NTSB forgoed the use of more advanced forms of testing in its simulation of the events leading to the tragic explosion of TWA Flight 800 in favor of a more expensive direct approach.  By performing direct testing via sensors on a similar Boeing 747, the NTSB acknowledged the limits of the computer model in providing us a simulated representation of the real world. 

The question then becomes simple.  What, in the age of exotic 3-D renderings and virtual reality is preventing the complete obsolescence of real-world testing?  Why bother to test or to study in the physical world when the processing capabilities of modern computers seems almost limitless?  Here are some answers to those questions: 

  • Processing and storage capabilities of modern computers are not limitless.  Many problems involving highly complex systems are beyond even the fastest supercomputer's abilities.  The process of reducing a problem's complexity such that it may be investigated on a modern computer may involve assumptions which are clearly unacceptable.  Scientists modeling the heart and engineers modeling complex aircraft systems, amongst others, still lack the power needed to fully simulate them even though significant progress has been made.
  • Models only represent aspects of systems which we can both understand and quantify.  In the case of the post-accident investigation of TWA Flight 800, the investigators were clearly looking for an aspect of the Boeing 747 which was not apparent to them prior to the accident.  The NTSB deemed a physical simulation appropriate as they were not convinced an available computer model would confirm the true cause of the accident.
  • Models always involve assumptions.  The real world involves systems of infinite complexity.  To make a problem computable within a reasonable amount of time, clearly some assumptions must be made.  In some instances, the assumptions required to do this may invalidate the results of the model.
  • Models are only as good as their creators.  A model based on false theories or bad assertions may present data which at first glance appears correct.  However, models must be examined with the same scrutiny given to any other scientific or engineering tool.
  • Simulations, like any other programs, are subject to bugs.  The danger of bugs leading to serious problems with computer simulations is made worse by the fact that in some cases scientists and engineers do not know what to expect when it comes to the results of the simulation.  They generated the simulation for the purpose of finding those results (this is not to say that invalid output is unrecognizable ... one should always ask whether the results do or do not "make sense").
In any event, a high level of confidence is required prior to adapting a computer simulation scheme in place of a physical testing program.  Though computer modeling of real scenarios has been of great benefit to engineers and researchers in recent years, the physical testing program will continue to have its place in the years to come. 

Links: 

 
Simulations in Everyday Life
Computer modeling and simulation is no longer strictly the domain of scientists and engineers working on supercomputers pushing the state of the art.  With the advent of smaller and faster microprocessors and storage devices, the power of the advanced simulator is available to the home PC user.  Here are some examples of simulators available (or soon to be available) to the home PC user: 
  • Flight simulators such as Microsoft Flight Simulator.  These applications are steadily improving in their realism and graphics capabilities, though the lack of realistic flight controls and surroundings (i.e. your own home plane cockpit) is apparent.
  • City/sociological (game) simulators such as Maxis' Sim City 3000, soon to be available to the PC user.
  • 3-D walk-around simulations such as ID Software's Quake II, which allow players to walk around in a 3-D game board attacking evil creatures of all sorts.
Though many of these simulations fall into the category of games and entertainment software, they certainly are not limited to it.  Less entertainment-related software such as business forecasting simulations are available for the PC.  Even accurate weather forecasting models can be had on the PC, such as NCAR's MM5 for Microsoft Windows/NT.  Though the most advanced in the arena of computer simulations are limited to high-capacity machines out of reach to the average individual, certainly the computer simulation is beginning to reach the typical PC user on an ever-increasing scale. 
  
What problems, if any, do these simulations pose?  It is difficult to argue that images emanating from a 15" or 17" monitor could be misconstrued as the real thing.  However, with technology continually advancing, the prospect of a virtual reality universe unto itself is fast approaching.  Highly sensory experiences such as that from a roller coaster ride or a bumpy plane ride are now available to theme park visitors across the country.  The most apparent threat thus far has been that people will replace outdoor sports and recreational activities with their less strenuous simulations on their PC's or their Nintendo 64's. 

However, the threat may in fact be more serious than we, at first, perceive.  Here are some more pressing concerns regarding the influence of simulations on our daily lives: 

  • What impact, if any, will simulations have on our social behavior.  Will the advent of such programs lead to further detachment from neighbors and members of the community?
  • Do simulations lead to a dangerous false perception of reality?  For example, do game simulations involving guns and violence lead to one's belief that guns and violence are an acceptable means of solving conflicts?
  • Will simulations lead to a false sense of one's capabilities?  Will people wrongly believe that real-world feats are possible after performing the same feats on a simulation?
Movies and television shows have begun to explore the world as it would be with the proliferation of advanced simulation technology.  They have been portrayed both as invaluable tools of society and as instruments of social destruction.  As simulations improve in quality and realism, society will be forced to confront these issues on a grander scale. 

Links: