INFORMS Prize 2008

The INFORMS Prize is given to an organization for “effective integration of operations research into organizational decision making”.  It is given to organizations for sustained use of operations research.  The criteria are

  1. Variety of Applications of OR
  2. Competitive Advantage to the Organization
  3. Business Impact
  4. Business  Model for Success
  5. Endorsements (from top-level management)
  6. Overall Quality of the Application

INFORMS just announced at the INFORMS Practice Conference the 2008 winner to be GE Global Research, Risk and Value Management Laboratory, with Intel and MITRE as the honorable mentions.  The CTO of GE gave a very nice speech about the role OR plays at GE.

Some Miscellany from the INFORMS Practice Conference

Some random things from today’s INFORMS Practice Conference:

  • Sanjay Saigal, popular columnist for OR/MS Today and founder/CEO of Intechne, formerly of ILOG, just to continue a theme, chided me for not pointing to his blog. I actually read his blog, but he normally blogs on non-OR things. It is great (with a great title): check out “Another Argumentative Indian”.
  • I met Sandy Holt, who had given me a blog idea. If you see me, don’t hesitate to say “hi!”: I’d love to meet more of you who have emailed me over the past couple of years.
  • I had a good long chat with Cindy Barnhart. Cindy is the current President of INFORMS, and is an associate dean of engineering at MIT. For a person who seems pretty laid back, she certainly seems to get a lot done!
  • I also had a long chat with Irv Lustig from ILOG. Irv is extremely upbeat about CPLEX at ILOG, just as Alkis Vazacopoulos is very upbeat about Dash and Fair Isaac. As OR becomes more mainstream (in a form known as “Advanced Business Analytics”, with “Business Analytics” being “look at your data!”), it is natural that the software firms would act more like businesses, being bought and sold, and having turnover. Perhaps I overly worry about the various changes.

OK, time for dinner and an early bedtime. My buddy from graduate school, Chris Lofgren, now CEO of Schneider is speaking first thing in the morning.

ILOG Optimization Decision Manager Hands On

I sat through (OK, half of) the ILOG workshop on their Optimization Decision Manager.  The ILOG ODM can be seen as a front-end to the rest of the ILOG Optimization systems (like OPL).  I would think of it as a super-sized way of doing version control and what-if analysis.  I use ILOG software in a lot of my research (and consulting) and I generally do something like

  1. Optimize a system
  2. Change some of the constraints and data
  3. Not like the results, and
  4. What was 1. again?

Rather than do intermediate saves (hmmm… I wonder what “testjunk.mod” is?), ODM allows you to save scenarios with differing data, models, parameters, goals and so on and then compare data and results between models.  This makes it much easier to mess around with instances and keep track of what works and what doesn’t work.   Embedded within in ODM is the opportunity to make constraints “soft” (putting them in the objective, with differing weights) and to do goal programming.

I am not quite certain the goal user for this.  I like the idea of using this in my own work, where I am the user.  I shudder about giving this system to someone without some reasonable understanding of operations research:  the “end user” here had better be pretty sophisticated.

Overall, I am glad that I attended the half of the session I did.  I do think we need better hardware/software/display for demos:  squinting at what appeared to be 4 point font at a screen was not particularly illuminating, and the resulting headache chased me from the room halfway through.

Bixby, Gu, and Rothberg leave ILOG

I arrived at the INFORMS Practice Meeting, and one of the first people I met was Bob Bixby. I had heard some rumors, and noticed that the affiliation on his badge was Rice University, so I was eager to chat with him. I wrote about Bixby last year when he was the IFORS Distinguished Lecturer at the EURO Conference. I think Bixby has been the most influential person in our field over the last fifteen years or so, and that influence has been primarily through a computer code he created. Though he began his career as a top-notch combinatorial mathematician, he decided at one point to write the world’s best linear programming code. This code, called CPLEX, has had a tremendous influence on the practice of operations research, greatly expanding our field’s reach and influence. Bixby’s company was eventually bought by ILOG, a company that also does constraint programming and business rules systems (see Simon Holloway’s view of ILOG from someone versed in business rules), and ILOG has consistently improved CPLEX and provided support for the team developing it. Bixby has had a number of roles within ILOG, including Chief Science Officer.

Bixby, along with two on the CPLEX team, Ed Rothberg and Zonghao Gu, have left CPLEX in the last few months. This puts the IP/LP software world in a tremendous state of flux. On the positive side, three of the best people in our field will be able to spread their skills to other companies (Bob was cagey on where he is thinking about going, and I have not talked to Rothberg or Gu). On the negative side, there are now questions about two (with the Fair Isaac purchase of Dash Optimization) of the top codes that underly much of linear/integer programming based research and practice. Of course, CPLEX has a large (15 person perhaps) development team, and no one is irreplaceable, but that is a lot of change in a short period.

8:50 PM Correction. Ed Rothberg was not on the CPLEX team when he left ILOG. Further corrections as conditions warrant!

9:40 Addition.  Just so the following is not buried in the comments, here is the response from Irv Lustig from ILOG:

Further corrections to Mike’s post:

Bob Bixby has not been working on CPLEX for 4 or 5 years. Ed Rothberg has not been working on CPLEX for 2 years. Both of them had been working on ILOG’s FPO product. So only Gu has left the CPLEX R&D team.

As mentioned by Mike, ILOG acquired CPLEX over 10 years ago, and it is quite unusual in the software industry for a founder of an acquired company to remain more than 4 years after the acquisition. We were fortunate that Bob stayed for the 10+ years that he did.

I was also one of the original CPLEX developers, and I have not touched the code in over 10 years, moving on to other roles within ILOG. It is a testament to CPLEX and ILOG that we have replaced Bob, Ed and myself with new developers, and we have recently hired very good talent to improve the CPLEX product. The CPLEX product will survive just fine after these departures of our friends and colleagues.

OR at P&G

Let me be late in the OR blogging game and note that there is a great article on operations research at Proctor and Gamble on bnet. It is wonderful advertising for our field, including phrases like “P&G’s Killer Apps in OR”. INFORMS was strongly involved in the article, with quotes from Past-President Brenda Dietrich and Executive Director Mark Doherty:

P&G, GE, Merrill Lynch, UPS — the list of Fortune 500 companies getting into the OR game is expanding, says Mark Doherty, executive director of the Hanover, MD-based Institute for Operations Research and Management Sciences (INFORMS), an OR think tank. “In the private sector, OR is the secret weapon that helps companies tackle complex problems in manufacturing, supply chain management, health care, and transportation,” he says. “In government, OR helps the military create and evaluate strategies. It also helps the Department of Homeland Security develop models of terrorist threats. That’s why OR is increasingly referred to as the ‘science of better.’”

Having sat in on a few too many board meetings, I think calling INFORMS an “OR think tank” is going a little far. But the article does project the very best vision of our field.

Check out another take on the article at Punk Rock Operations Research (and I thought I saw it on another OR blog, but it escapes me at the moment).

Dash Optimization is acquired by Fair Isaac

Dash Optimization, makers of Xpress-MP (one of the two leading linear and integer programming solvers, along with ILOG’s CPLEX), has been acquired by Fair Isaac. Fair Isaac is an anaytical application company, known best for their credit rating systems (they do the FICO scores that companies use to determine whether to extend credit). This is an interesting move. On one hand, we have a premier OR company being acquired by a company that is not particularly known in the OR world (despite their analytical focus). Will Xpress-MP be forgotten, particularly as the credit market is troubled? Will Xpress-MP continue to be a available and supported for the broader world?

On the other hand, a very well-known company now has operations research software as a key aspect of their offerings. Will they be able to leverage this to extend the reach of OR to more firms? One nice line of their press release:

We will showcase the Dash products and introduce our clients to these capabilities at the forthcoming InterACT conferences in Vienna and San Francisco.

Having OR methods showcased by Fair Isaac could be a huge boon to the field. If companies think “Fair Isaac is investing in OR, maybe we should find out more about the field”, that would be great! But one potentially worrisome aspect of the press release:

With Dash part of Fair Isaac, we can do more to:

  • Incorporate an additional layer of analytic power into our industry-standard solutions
  • Create advanced custom solutions based on specific client needs
  • Develop innovative new solutions for large and complex business problems

How much will Fair Isaac keep within itself rather than provide core optimization products to the wider world?  The longer press release calms fears about the role Dash software will play:

Fair Isaac’s acquisition of Dash builds upon a longstanding partnership between the two firms.  Dash optimization technology is currently embedded in Fair Isaac’s Decision Optimizer, a software tool for achieving the smartest decision strategies given operational complexities, resource constraints and market uncertainties.

“Demand for sophisticated decision management tools is growing rapidly, and the addition of Dash optimization technology to our portfolio helps us to remain the market leader,” said Mark Greene, CEO of Fair Isaac.  “With their optimization capabilities and our own business rules management and predictive analytic solutions, Fair Isaac now has the industry’s most comprehensive decision management suite.”

Being part of a “decision management suite” seems good!

There is a discussion of this on the USENET group sci.opresearch. In the discussion, Bob Daniels, one of the founders of Dash Optimization, talks about the change:

The good news for Xpress-MP and its users is that ALL the Dash employees are
moving enthusiastically to Fair Isaac and the development budget is
considerably enhanced.

For those of you who don’t know, the name “Dash” comes from (Bob) Daniel and
(Robert) ASHford. Robert and I, and the team we have recruited, have been
developing Xpress-MP for over two decades and we weren’t about to sell our
“baby” to people who weren’t dedicated to Math Programming and Xpress-MP. Of
course, we the founders have moved to Fair Isaac too.

Best wishes to Bob Daniels and the rest of the Dash Optimization team in making this transition from a stand-alone company to part of a much bigger operation. And let’s hope that Xpress-MP continues to get better and better (and not just for the use of Fair Isaac!).

New CPLEX out

I got an email last week from a PR firm asking whether I would be willing to talk to some ILOG people about the upcoming release of CPLEX 11. Made me feel like a real journalist! Since I am unfortunately missing the Seattle INFORMS (the 12 hour flights from Auckland to LA are killing!), I was going to miss my normal chats with Irv Lustig and the gang, so I did talk with them a few days ago by phone.

You can read the press release about CPLEX 11. Since I have a few sports scheduling problems that are just not solving under CPLEX 10 (or anything else I have tried), I am looking forward to seeing if things are better under 11.

There were three things that struck me when talking to the CPLEX folks:

  1. Pretty well all of the improvements are in the integer programming parts, not the linear programming parts. This is in keeping with Bob Bixby’s talk at EURO (you can see a version of the talk in Seattle, too) where he said that LP improvements were enormous from 1996-2003 but not much has happened in the last few years.
  2. The improvements in mixed integer programming in this version are coming from improved search. This contrasts with previous improvements that came from better cuts. So, when CPLEX went from version 6 to 6.5, there was a huge improvement in speed due to the addition of Gomory cuts. I would guess that in the academic world, search has received about 1% of the attention that cuts have gotten. There is a well defined theory of facets, cut generation, and so on, and very little general insight into search. CPLEX 11 uses something ILOG calls “dynamic search”. If you use dynamic search, you have to give up all the callbacks that are normally used to provide user-defined search, so you are accepting the ILOG “black box” on search. This should lead to some pretty interesting testing: can problem-dependent search do better than the general search approach? The constraint programmers have always had search as a key component to their systems; it is about time integer programmers spent more time thinking about it. Is it possible to put search on the same solid foundation that cut generation is? Or will search end up being just a bunch of heuristics that happen to work well on many instances?
  3. Parallel is getting more attention (and parallel linear programming may be the step that makes LP faster). But it is clear at this point that having 8 cores (as I do!) won’t result in 8 times speedup. First, the root node isn’t parallel, so anything with an expensive root node won’t see much improvement at all. Later, once the branch-and-bound search tree is in full swing, the amount of memory access needed is large, so the bottleneck occurs in the pipe to the memory. I got the impression that this is something they are working on, which is good: lots of the upcoming speed improvements are going to be in parallel computing rather than higher clock speeds.

There are other new aspects to CPLEX 11 on the usability front, most notably the ability to generate all optimal solutions, not just one, which seems like a useful feature.

ILOG has put together some forums on optimization and constraint programming. If you hurry, you can be the first to post to them!