Chapter 135 22.5 Looking Ahead: Inside Action

In the course of investigating prediction and modeling mechanisms, I had the opportunity to visit the Jet Propulsion Laboratory in Pasadena, California.A state-of-the-art war simulation system is being developed there.I was invited by a computer science professor at UCLA to come to JPL.The professor has been expanding the field of application of computer power.And like many researchers who lack financial support, this professor has to rely on military funding to carry out his cutting-edge theoretical experiments.According to the transaction agreement, all he needs to do on his side is to pick a military issue to test his theory.

His test bed was to see how large-scale distributed control parallel computing—what I call "cluster computing"—could speed up computer simulations of tank battles, an application software program he wasn't too interested in.On the other hand, I'm really, really interested in seeing a top-notch war game. Once at the busy front desk of the lab, go straight to the security check.Since I was visiting a national-level research center, and the US military was on red alert at the border with Iraq, the security guards were quite enthusiastic.I signed some forms, took an oath of loyalty and citizenship, pinned a large badge, and was escorted upstairs with the professor to his comfortable office.In a small gray conference room, I met a long-haired graduate student who, in the name of studying how to use mathematical methods to simulate war, was looking for some innovative concept of the theory of cosmic computation.Then, I met the head of the Jet Propulsion Laboratory.He was nervous about my presence here as a reporter.

Why?My professor friend asked him.Simulation systems are not classified; research results are published in the open literature.The person in charge of the lab had a lot of rhetoric: "Ah, um, look, there is a war going on now, and we have been playing that plot in general for the past year or so-we chose that game purely by chance, not at all. There is no meaning of prediction—but now there is a real fight. When we start testing this computer algorithm, we always have to choose some episodes, any episodes, to try out the simulation effect. So we picked a simulated desert war, Those who participated in the war...were Iraq and Kuwait. Now that this simulated war is actually fought, we are more or less on the scene here. It is a bit sensitive. Sorry."

I didn't get to see that war simulation.However, about a year after the Gulf War ended, I discovered that the JPL wasn't the only place that accidentally rehearsed that war.The U.S. Central Command in Florida staged another more practical simulation of desert warfare before the war.The U.S. government simulated the Kuwait war twice before the war. Cynics believe that the U.S. government simulated the Kuwait war twice, which portrays its imperialist face and long-planned desire to launch the Kuwait war.In my opinion, the predictive scenes are not so much vicious as they are weird, bizarre and instructive.I use this example to outline the potential of the predictive mechanism.

Around the world, there are about twenty-four operation centers conducting this kind of war game with the United States as the blue army (that is, the protagonist).Most of these places are small departments under military academies or training centers, such as the war game deduction center at Maxwell Air Force Base in Alabama, the legendary global game room at the US Naval War College in Newport, Rhode Island, or Kansas. The classic "sand table" desktop deduction device of Leavenworth's Army Field Concept Department.Those who provide technical support and practical knowledge for these war games are some academics or professionals hiding in countless paramilitary think tanks. , California's Lawrence Livermore National Laboratory and other research areas in the corners of various national laboratories.Of course, these toy war simulation systems are all acronyms, such as: TACWAR, JESS, RSAC, SAGA, and a recent military software directory lists about 400 different military games or other Military models, and they are all listed for sale.

The nerve center of any U.S. military operation will be at Central Command in Florida.The Central Command, as an agency of the Pentagon, exists for the purpose of keeping an eye on a major battle situation for the US Congress and the American people like a falcon: the blue army versus the red army - the game between the superpowers, and the only one of them It is the Soviet Union that deserves to be the opponent of the United States.When General Norman Schwarzkoff came to office in the 1980s, he didn't accept that view.Schwafczkow—a thoughtful general—proposed a new perspective, and the phrase he used to express this new perspective was widely circulated throughout the army: "Soviet dogs don't hunt ".Schwafczkow then turned the attention of his planners to alternative scenarios.At the top of the list is the Middle East desert war along the Iraqi border.

In early 1989, a Central Command officer, Gary Weir, began building a model of warfare based on Schwarzkoff's intuition.He worked with a team of military futurists to collate data so they could create a simulated desert war.The simulation was code-named "Operation Inside Sight." Any simulations are only as strong as their underlying data, and Weir wanted Operation Insider to be as close to reality as possible.This means collecting detailed data of nearly 100,000 troops currently stationed in the Middle East.Most of this part of the job is extremely dull and tedious.War simulations need to know the number of vehicles deployed in the Middle East, the amount of food and fuel stockpiled, the lethality of weapons, weather conditions and so on.Most of these minutiae things are not readily available, and even the military is not easy to get them.All of this information is in constant flux.

Once Will's team had a plan for the organization of the army, the war pushers would compile CD-ROM maps of the entire Gulf region.And the basis of this simulated desert warfare—the territory itself—was transferred from the latest digital satellite photos.After this work is over, war game players will compress the terrains of countries such as Kuwait and Saudi Arabia onto CDs.At this time, they can input these data into TACWAR, the computer war game main simulation program. Weil started the desert war in the fictitious Kuwait and Saudi battlefields in the early 1990s.In a conference room in northern Florida in July, Gary Weir outlined the various outcomes of Operation Inside Look to his superiors.They looked at a situation where Iraq invades Saudi Arabia and then the US/Saudi strikes back.Unexpectedly, Will's simulation predicted exactly a thirty-day war.

Just two weeks later, Saddam Hussein suddenly invaded Kuwait.Little did Pentagon top brass know at first that they had a fully operational, data-rich simulation.Just turn the key, and the simulation program predicts possible wars in the region under endless changes.When news of the prescient simulation broke, Will smelled like a rose."If we wait until the invasion starts, we'll never get there," he admits. In the future, standard armament regulations may call for a command center with a box running all possible The parallel universe of war can be launched at any time. Immediately after Saddam's invasion, war actors turned their "inside view" to running simulations of "real" situations with infinite changes.Their attention focused on a set of possibilities around variables: "What if Saddam kept attacking the situation?" Iteratively calculating the predicted fighting within thirty days, Will's computer took only About 15 minutes.By running these simulations in multiple directions, Will's team quickly came to the conclusion that air power would be the deciding factor in this war.Further precise iterations show very clearly that if the air battle is won, the United States can win.

Not only that, but according to Weill's prediction mechanism, if the air power can indeed complete the mission assigned to them, then there will be no major losses to the US ground forces.The understanding of this conclusion by senior officials is that conducting precise air strikes first is the key to the low casualty rate in the United States.Gary Weir said, "Schwarzkoff was so tough on keeping our troops to an absolute minimum casualties that low casualties became the benchmark for all of our analysis efforts." In this way, predictive simulations give military command teams the confidence that the United States can win wars with minimal losses.This confidence led to heavy air strikes."The simulation definitely influenced our thinking [at Central Command]. It's not that Schwarzkoff didn't feel strongly about it beforehand, but the model gave us the confidence to implement those ideas," Weir said.

As a prediction, Action Insider has indeed achieved very good results.Despite some variation in the initial balance of forces and a small difference in the ratio of air to ground combat, the simulated thirty-day air and ground campaign closely approximated what actually took place.The ground battle unfolded largely as predicted and gradually.Like everyone who wasn't there, the Sims were amazed at how quickly Schwarzkoff wrapped up the final round at the front.Will said, "However, I have to tell you, we didn't expect to get this kind of progress [on the battlefield] in a hundred hours. From what I recall, we predicted that it would take six hours." Days of ground combat, not a hundred hours. Ground force commanders have told us that they expected to move faster than the simulations indicated. And they did. quick." According to the calculations of the prediction mechanism of this war game, the resistance of the Iraqis will be greater than their actual resistance.This is because all combat simulations assume that the adversary will use all of their available resources with all their might.But in reality Iraq didn't try that hard at all.Warfare actors once cheekily joked that none of the models would include raising the white flag into the weapon sequence. Because the war progressed so quickly, the simulators no longer had time to sequentially consider the next step in the simulation: that is, to forecast the progress of the war in daily mode.Even though the planners recorded the daily events in as much detail as possible, and they could always plan into the future, they felt: "After the first 12 hours, it doesn't take a genius to figure out the future."