Saunders "optimizes" his time at Stanford

By Eric Grunwald
Office of Technology Licensing
Stanford, Spring 1992

Do you have a problem? Not a small problem, but a really big one? Perhaps Dr. Michael Saunders can help you.

You won't find Professor Saunders in the psychology department, however, and the answers he's likely to give you will have little to do with Sigmund Freud. Dr. Saunders is a Research Professor in the Systems Optimization Laboratory (SOL) in Stanford's Operations Research Department and co-author of the optimization program known as MINOS.

Dr. Saunders, born in Christchurch, New Zealand, first came to Stanford in 1967 as a graduate student in computer science. After finishing his Ph.D. in 1972, he returned to New Zealand to work as a government scientist. "I hated leaving Stanford," Saunders says. "It was a very traumatic experience."

Back in New Zealand he met up with Bruce Murtagh, a fellow New Zealander who had studied optimization in England. Large-scale linear programming was already a well-established tool, but the two decided to make the leap to large-scale nonlinear, and MINOS was born. Saunders did the main work on MINOS two years later, back at Stanford as a research associate.

MINOS (Modular Incore Nonlinear Optimization System) is now used to solve large-scale optimization problems having thousands or even tens of thousands of equations and variables. Over 800 copies of MINOS have been distributed to universities, companies and other organizations world-wide, and thousands more are used as "black boxes" inside a modeling system called GAMS.

After two more years in New Zealand distributing MINOS, Saunders returned to Stanford for good in 1979 as a senior research associate, joining Walter Murray, Philip Gill and Margaret Wright in SOL in what came to be known as the "Gang of Four."

Since then, Saunders has spent his time continually reworking MINOS and developing other optimization software such as LSSOL and NPSOL (also distributed by the Software Distribution Center at OTL) with the Gang of Four. Much of the software revenue has gone back into a research fund to support their work.

Saunders finished MINOS 5.0 in 1983, at which time he was still copying and distributing the tapes himself. "It was getting harder and harder to keep up," he remembers. "It was taking up far too much time."

OTL then agreed to take over the licensing and distribution of MINOS. "OTL has been very, very helpful," Saunders says. "They made it feasible to distribute much more widely," adding that OTL did work on commercial licensing that he couldn't have done himself.

Despite a lack of competition and any significant complaints, Saunders, who became a research professor in 1987, isn't at all complacent. "I lie awake at night," he says, "worrying that thousands of people are using MINOS, and there's really no guarantee it will solve all of their problems." He says his main task now is to improve the reliability and efficiency of his optimization programs.

But Saunders has another major commitment now: 2-year-old daughter Tania Michelle. "It was fairly late in the day," Saunders chuckles, "but it was much to my amazement and delight." The photos of Tania and Prudence, his wife of 24 years, covering the wall above his desk attest to that. "I don't think I'll need OTL's help on this one," Saunders laughs. "Tania seemed optimal from day one."