TECHNOLOGY SPECIAL REPORT

Computer Simulations: Modeling the Future

A military planner at the Joint Forces Command in Norfolk, Virginia, watches a battle unfold without taking his eyes off the computer screen. The software is tracking a million vehicles spread over the Southeast Asian terrain — the most ever tracked in a military simulation.

Modeling and simulation have made momentous strides in recent years, and the military, medical science and other professions are on the verge of being able to use computing power to simulate reality for all kinds of applications.

“We are within sight of being able to create a large-scale, high-resolution battlefield environment detailed enough to let us experiment and see how a given system might perform,” Robert Lucas, director of the computational science division of the University of Southern California’s Information Sciences Institute, told TechNewsWorld. “Advances in both AI software and in networked computing have made virtual environments of previously impossible realism possible.”

Quantitative Determinations

In the recent past, simulations were imprecise, even frustratingly so, to computer scientists. But advances in the technology are helping to make reliable quantitative determinations possible.

Researchers at USC, working with developers at Options Technology, have linked supercomputers and intelligent software agents to perform a simulation for the Joint Forces Command in which 1 million vehicles move around a large, urban geographic area.

Previously, researchers were only able to simulate roughly 100,000 vehicles in an exercise of this type.

The computer scientists employed an array of technologies to make the project happen, including a 520-node cluster of Linux computers on the USC campus linked to other clusters at the Pentagon and in the private sector over high-speed networks. The simulation output was produced at the Pentagon’s Joint Experimentation Laboratory in Virginia — home to previous experiments of this kind.

Artificial Autonomy

Using artificial intelligence designed at USC, each of the simulated vehicles in the model was given autonomy, the ability to respond on its own to changes in its environment and the ability to travel over wide geographic areas — just like real-world vehicles, be they cars, trucks, tanks or personnel carriers.

Many of the vehicles in the simulation were “clutter.” That is, they were not directly involved in the military operation but had to be tracked by military planners, just as commuter cars driving around the streets of Baghdad would be monitored.

The next step for researchers would be a real-world exercise of the same scope to give military analysts the ability to control sensors in the field to track movements of vehicles and troops. This test is to be conducted this summer, and, if successful, the Pentagon plans to use the modeling technology for continuous training and experimentation.

Medical Modeling

Roger Smith, vice president and CTO of national security contractor Titan, told TechNewsWorld that other commercial contractors have been working on cutting-edge military simulations to model the spread of chemical and biological agents in the air and the impact of terrorist attacks on the national infrastructure.

But military simulation is not the only useful application for this kind of technology. Others are using simulations and models for medical research. At the Case Western Reserve University School of Engineering, researchers have developed an advanced simulation system, called the Model Integrated Metabolic System, that is designed to aid research about the heart, liver and brain under the stress of exercise.

Scientists at Entelos, a pioneer in predictive biosimulation, have developed mathematical models of different disease states — including diabetes, obesity and rheumatoid arthritis — that “replicate actual biological responses,” spokesperson Alan M. Zachary told TechNewsWorld.

The company’s software, PhysioLabs, can discern why one person with a particular disease will respond to medication, while another with the same symptoms will not. “The PhysioLab is the equivalent of a flight simulator for fighting diseases,” said Zachary.

Current clients using the simulator include drug makers Bayer, Johnson & Johnson and Organon.

Supply Chain Management

Another commercial use of modeling is in supply chain management. Consulting company Insight has developed optimization and modeling algorithms designed to model how a terrorist attack or a natural disaster might affect an individual company’s supply chain.

Other companies, such as Clockwork Solutions, are developing modeling and simulation technology for mathematically characterizing industrial systems to predict how component aging in factories might affect operation of those factories in future years. Clockwork Solutions’ simulation technology also can account for maintenance policies that could alleviate problems associated with aging components.

“The model quantifies expected asset performance and life-cycle cost, and identifies and ranks areas for potential improvement in component reliability, repair, spare part stocks and maintenance,” company spokesperson David Kanaan told TechNewsWorld.

Modeling’s Future

Scientists are continuing to develop new uses for models and simulations, including factoring census data and time to model the growth of the U.S. population. Researchers are also using simulations to compute the orbit of planets, regional transportation flows and stock market fluctuations.

Analysts expect that use of complex simulations likely will grow as the industry makes advances in clustered — or grid — computing. These technologies, and their successors, will bring simulations ever closer to reality.

Leave a Comment

Please sign in to post or reply to a comment. New users create a free account.

More by Gene J. Koprowski
More in Hardware

Technewsworld Channels