Two university professors have come up with a computer program they say could be useful in predicting world leaders' decisions in times of crisis or negotiation.

The computer model attempts to simulate the thought processes of key advisers to decision makers, using the opinions, prejudices and attitudes obtained from the advisers themselves in actual interviews.

One of the men who put it together says the printout successfully predicted the U.S. would act in the 1973 Yom Kippur war, for example. Also, he said, the CIA has expressed interest in the project.

"This departs from the traditional approach of treating policymakers as rational beings," said G. Mathew Bonham, professor at American University's School of Government and Public Administration and the School of International Service.

"Really they're just ordinary people who often make irrational choices and are often mistaken in their view of the world. They may feel they have superior intuition, but they can be wrong in that, too," he added.

"Decision makers tend to believe that events are related causally . . . even when there is little or no evidence of a causal nature," the two researchers said in a paper on their computer model written last year.

The program Bonham designed, along with Prof. Michael J. Shapiro of the University of Hawaii, attempts to trace these causal relationships as the decision makers see them and to pinpoint the way new events will be seen and finally acted upn.

Bonham and Shapiro interviewed 15 American and two Russian policy advisers between 1971 and May 1973, on their views of the Middle East situation. Key phrases and cause-effect opinions each person expressed were isolated and coded for the computer.

For example, one of the Americans said: "It would appear to me that the continuous occupation of Arab territories by Israel is precisely this casual factor that makes Arab regimes open to offers of Soviet military aid and brings more and more Soviet penetration into the Middle East."

Bonham explained that the sentence contained three concepts in two relationships that were entered into the computer. "Occupation of Arab territory by Israel is perceived to cause Arab openness to Soviet aid. Then that is perceived to cause or provoke Soviet penetration in the Middle East. These relationships are each shown by an arrow with a plus (on a computerized diagram) sign indicating a positive or promoting action," he said.

Such diagrams for each sentence in the record interviews were used to build "cognitive maps" of the thought processes of each individual, and the computer combined them into maps for the group as a whole. The experiment assumed that a person's thoughts tend to follow maps or pathways of cause-effect relationships already believe in, Bonham explained.

New events or information are interpreted in light of that pre-existing cognitive map, he said, and decisions will tend to be made on the basis of it.

"The computer comes up with a number of alternative ways people will think about what is happening and what will happen as a result," he said. "Complex problems have many causes and effects, and as long as they don't contradict each other a person can hold several views simultaneously."

The computer automatically eliminates contradictory chains of thought, he said, saving those that are supported in the interview by historical analogy or that appear most frequently in the conversation.

The Americans interviewed before the 1973 Arab-Israeli war included members of the National Security Council staff, the Defense Department's International Security Affairs office, the State Department's Bureaus of Near Eastern and South Asian Affairs and Mideast specialists in State's Bureau of Intelligence and Research, Bonham said. Roger Davies, the U.S. ambassador killed in Cyprus in 1974, was one of them, he said.

"Lots of things have happened to them since - some have been promoted, others are out of government or out in the field," he went on. "They were all promised confidentiality."

The researchers said they encountered "a number of problems" in trying to interview Soviet decision makers and managed only to talk with two scholars of the Soviet Academy of Science in Moscow, Bonham said. He described them as two of the three "main advisers" to the Politburo on Mideast affairs at the time.

When the Mideast war broke out in October, 1973, Bonham said, "We had it all programmed, and it was quite exciting. We could see on the (computer) stimulation that the American didn't know how to react at first. The Soviet stimulation predicted exacted what finally happened."

For example, the computer predicted the Soviet advisers would probably view the 1973 war as a consequence of Israeli deep-penetration raids on Arab territory, Egyptian losses and subsequent Egyptian attempt "to liberate its own land," along with "retaliation against Israel no matter how dangerous," Bonham said.

Although such results might seem obvious without the help of a computer, it was significant that the computer vame up with them, Bonham said. "Still, the results are not as interesting as . . . the computer's reading on their view of what is likely to happen."

The predictions were as limited as the interviewers' data, and that is one of the things that makes the computer program intriguing, Honham said. "The decision makers themselves have transcripts or very careful notes on private conversations with Soviet leaders from many sources," he said."God knows what they've got, but they could turn all of it into cognitive maps. Our problem was access, but the computer model could be used for much more detailed information and it could be very useful for some crisis six months down the road."

As it was, the two American researchers asked the Russians what they would think if they were in the American position, "because if people aren't telling you what they really believe, it won't work," Bonham said. "We structured the questions this way because we thought they'd be a little more trughful then."

The computer, as a result, showed that the Russians in the U.S. position would try to initiate a cease-fire or do nothing rather than intervene militarily, which is, of course, what happened.

It showed American policymakers ranking seven policy options in the order that they seemed to occur: "posturing," "application of diplomatic pressure," a cease-fire initiative" and "a show of force," which all ranked higher than aid to Israel, lack of any initiative or military intervention.

The 1971-73 Middle-East model is outdated now, Bonham acknowledged. To remain useful, the program must be updated constantly with recent events and the decision makers' attitudes toward those events. This might be done with newspaper reports of remarks by the persons involved or with private conversations or more interviews, if possible, Bonham said.

He said the CIA's Office of Political Research contacted him in 1975 and asked to be briefed on the computer project. "I gave a presentation in the Bureau of Intelligence and Research at the State Department, too, but that was the last I ever heard of it," Bonham said.