Decision Help For The PhD-Impaired?

Ilog has rolled out a solution allowing
line-of-business owners to take part in decision-optimization work that has
thus far been the exclusive province of optimization experts.

The Optimization Decision Management System (ODMS) allows executives to
review predictive mathematical models in terms they can understand, and help
translate strategic decisions into tactical and operational decisions.

Thus, they can perform what-if analysis and scenario comparisons so they can
understand their options and balance conflicting business goals, such as
service level agreements and various costs such as labor or fuel.

Heretofore, these business owners had to rely on PhD-level mathematicians to
develop those predictive models.

According to Jeff Kilbreth, director of optimizations product marketing for
ILOG, ODMS allows business owners to peek into the black box and give
quicker, more accurate feedback to the math wizards.

“Historically there needed to be a mathematical god to build the models.
ODMS is making it easy for business people to participate in the model
development process and trust the results of the models,” he told

ODMS allows the developers to generate a working application for which
executives can provide feedback.

“With a working decision support application, the non-mathematician can
easily understand all of the costs, constraints, resources, and goals” input
into the model.

This allows them to vet the assumptions made by the developer.

“They tell the optimization expert what is missing or incorrect and then
they do it again. Typically three prototype applications gets you to a final
model and therefore a final application,” said Kilbreth.

“Business users won’t use the information unless they’re sure they can trust
it,” he added.

Most companies using this type of solution are in manufacturing, where ODMS
can be used to resolve supply chain issues like optimizing truck routes or
raw materials utilization.

Solutions of this kind are also being used to resolve complex
decision-making puzzles in other industries, such as scheduling staff in
large hospitals, pricing routes in the airline industry or calculating
portfolio weightings in financial services.

Ilog solutions are already imbedded as rules engines within supply chain
applications from vendors like SAP and Manhattan
Associates .

But the modeling used by those applications tends to be too generic for many
larger companies with specialized problems to resolve.

Ilog solutions have also been adopted as stand-alone applications run by
operations research professionals at large enterprises.

Ilog says 88 of the world’s 200 largest companies use its solutions to build
their own optimization models.

Bob Parker, vice president of research at Manufacturing Insight, noted that
ODMS will help companies actually implement better decisions more quickly.

“What this gets you is you can iterate the decisions much faster,” he

ODMS doesn’t solve every problem in the decision chain, however.

“You still need to write the new decision back to the transaction system so you
operate against that [new] plan. That process is still not streamlined,”
Parker said.

Still, he said, the solution addresses a very real problem.

“It’s something that’s been needed in the market in this area for quite a
while,” he told

Manufacturing Insight includes decision optimization modeling as part of the
business intelligence market, which Parker said drives approximately 20
percent of external IT spending.

“We see that doubling over the next five years,” he said. “Demand for this
is really going to grow.”

Ilog currently generates $50 million from decision optimization solutions,
and Kilbreth said the company expects to double that figure within three

Kilbreth said the solution costs anywhere from $250,000 to $500,000,
depending on the amount of customization of the ODM-generated screens, the
complexity of the optimization model, the complexity of the integration work
and whether the company has any operations research skills in-house.

News Around the Web