In his convocation address last Friday, computer programmer and CEO of On Time Systems Matt Ginsberg immediately identified what he would not speak about: the role of technology now and the role of technology in the future.
“There is one question I really want to ask,” said Ginsburg, “Why are nuclear submarines so expensive.” Nuclear submarines cost over a billion dollars to construct or four dollars for every American taxpayer.
Nuclear submarines could cost that much for several reasons: because no one can build them any cheaper, because the shipyards do not know how to build them any cheaper or because the shipyards do not want to build them any cheaper.
Summarizing the last two possibilities, submarines could cost so much, because “there is a magical incantation that would make the submarines cost less (that) the shipyards don’t know” or “they know the magic incantation but refuse to say the magic words, because they want the boats to cost a billion bucks each.”
By “magic incantation”, Ginsburg refers to a schedule that could lower the costs of building ships. His company, On Time Systems, developed such a schedule for navy shipyards.
Its computer program ARGOS, which stands for A Really Good Optimization System, iterates through many possible schedules and selects the least expensive one.
In three hours, ARGOS creates the most cost effective schedule for building a submarine, which costs 16 percent of the original cost, or 150 million dollars less. In 12 hours, it can optimize an entire shipyard, reducing its costs by 12 percent or 500 million dollars.
Ginsburg cautioned though that “just because a schedule looks good, that doesn’t mean it’s actually used.” Unions or corporations might not like a schedule and prevent its implementation.
Cost effective schedules avoid layoffs, however. “They are deigned to be cheap,” said Ginsburg, “and layoffs are expensive,” so unions who want fewer layoffs will likely accept a schedule that achieves that.
Corporations also might not want a more cost efficient schedule implemented. Two corporations operate the six shipyards in the United States, Northrop Grumman and General Dynamics, and those two corporations share the profits on the ships they build and monopolize shipbuilding in the United States.
“Technology is a lot of the story but is not the whole story,” said Ginsburg, “just because you can build a ship better does not mean people will [do] it”
A large audience listened intently to his words as he explained the intricacies of schedule making in academia and applied it to shipbuilding in the United States.
“I think Matt did a great job combining technical elements of his algorithm work with difficulties that arise in the real world that have nothing to do with the actual problem itself,” said Casey Goodge ’14, “It was an excellent reminder that just because something can be implemented, does not mean that it will play out correctly or as intended in the real world.”