Sim model speeds decisions

Engi described his systems as combining protein computers--human experts--with silicon
computers.


Both are necessary in his global approach to making infrastructure models for
government officials who respond to hot spots domestically and overseas.


As yet, the Energy Department lab has no paying customers for the decision-support
technology. But Engi is convinced the work his team is doing with dynamic simulation
language models will prove invaluable to U.S. and international policy-makers.


"We're trying to build a core competency in the labs that can help answer
questions and resolve issues the government might have," he said.


A silicon computer can run through 167 policy options faster than a human brain
can--maybe not as well or as thoroughly, but much more quickly, Engi said.


"If you're tasked to go through 10 to the 50th power different possibilities, it's
going to take more than your lifetime to do it," he added. "That's where a
machine comes in handy."


Engi's team develops models using Powersim, a dynamic simulation language from Powersim
Corp. of Herndon, Va. The models run under Microsoft Windows NT. Users draw simple
pictures of the infrastructures they want to model.


"Imagine drawing circles on a page and labeling those circles with whatever the
state variable is and connecting one circle to another if there is a coupling," Engi
said.


Links between variables are rendered as differential equations describing the rate of
change of one variable as a function of the other.


Infrastructure models, integrated with other models, could help policy-makers see how
to maintain supplies of valued commodities such as water and grain while disrupting
supplies of negative commodities such as infectious diseases or weapons of mass
destruction.


David Harris, the project's technical leader, has integrated separate water, grain,
energy and greenhouse gas models into a single infrastructure model with data flowing back
and forth. Changing agricultural parameters directly affects the output of greenhouse
gases.


From the model, the Sandia research team demonstrated to policy-makers that they
probably need not worry about Chinese demand for grain imports exceeding the world's grain
export capacity in the next 10 years.


This year the team will build on that integrated model with policy gaming simulations
that attempt to model China's policies on greenhouse gas issues.


Then the team will try to build similar models of U.S. policies and international
responses to China's greenhouse gas production.


Several experts have helped the system developers rank the relative importance of
various infrastructure issues, which then become the design specifications for support
systems to manage those issues, Engi said.


inside gcn

  • pollution (Shutterstock.com)

    Machine learning improves contamination monitoring

Reader Comments

Please post your comments here. Comments are moderated, so they may not appear immediately after submitting. We will not post comments that we consider abusive or off-topic.

Please type the letters/numbers you see above