Thanks Nic what I would like to do today is to tell you why I'm addicted to optimization. But I got to begin with with an apology. Normally one says it's a pleasure to be here. I have to say instead it's not a pleasure. Not to be here and in fact I I I wrote here's my excuse long before learning about this honor my wife and I had planned the fiftieth anniversary trip abroad and we just couldn't change that. So again please accept my apologies for not being here in person and I hope this video is Will. Well be a goods substitute. I'm very deeply honored to receive the first Cashion prize. I would very much like to thank the prize committee of more Groucho or cutting them Roski that I was part of Los and promised to Lockey and after I I learned from an e-mail from Thomas that I had I was the recipient I was curious to know who had nominated me so I asked Bill Cook and Bill said Of course you know I I was part of it because my colleague at Georgia Tech and in the nomination actually was done by Bill of my former students right corner. And and my very close friend Bill public. I really have a lot of people to thank and I will take a little bit of time to do this because without without help from from these people I could not have had the career that I've had first of all my thesis advisor Jack made and I came to Northwestern as a graduate student at nineteen thousand nine hundred fifty eight to use to get a master's degree in chemical engineering. I took as part of that I took a course from Jack in operations research. I had no. Very heard of operations research before and Jack was absolutely wonderful person who's who's influenced my whole life. When I finished my Ph D. I moved on to Johns Hopkins. Elena door was a senior faculty member there as a young faculty member it's kind of difficult to know exactly what to do and Ellie Ellie guided me through through with John Peter ropes and soon had to get into Hopkins the National Bureau of Standards now nest was doing a transportation product project on on how to make transportation more efficient in your These corridor from Washington to Boston. The technique my technical advisor on that project was was Alan Goldman you all of course know and his work and Alan had great influence on on me. After about eight or nine years at Hopkins I moved the Corps now and two years after I arrived at Cornell. Refocus and everybody knows of Ray of course the famous famous person for network flows. Rachael and the fact that you're there and are a really really I learned a lot from Ray about how to be a real scholar. Ha ha ha. You also want to thank all my all my colleagues who I've worked on who I've worked with in the last several years. I will mention every one of these people a little bit later on in the POC. And what I'm really most proud of. Is that over my nearly fifty year career at Hopkins Cornell and that. I have had fifty eight Ph D. Ph D. students that fit in. That's a little bit cheating because in the last couple of years I close to provide some of these students when with Alice Johnson and should be our man in Morgan's out as per actually. Actually I could put fifty nine in there because and you're somebody that you. You all probably know his name has been on the cover of Newsweek TIME magazine. Technically at Hopkins John Malone who got his Ph D. in sixty seven was a student of many Belmore but I really worked very closely with John. And one thing I can say safely is that John is certainly the richest person who ever got a Ph D. in an operations research. Well I've had some really wonderful students and I've learned. Of learn as much from them as as or more than they've learned from me so important part of my life my philosophy about doing research is is very simple. I really like to get my ideas from real problems. I like to get exposed to to real problems and when whenever you see these problems. There's always some new things come up that haven't come up before and that's where I like to generate generate more research in contrast I say to the person who reads papers and sees all this paper leaves out as as left this is an open problem. I don't think that's nearly as fruitful as as generating the ideas directly from real problems. I started as a as a chemical engineer I got a bachelor's degree in chemical engineering and went on to go with Western to get a master's degree but as I mentioned already in that in that first year I met Jack mitten. And Jack was also a chemical engineer and he suggested the thesis topic to me involving dynamic programming. You know standard dynamic programming usually involves a set of stages which which since early on were mostly the stages correspond to prime periods. You can think of that as a directive path in a graph. And so that you could think of dynamic programming has been had having been done on directed paths and graphs when you look at a chemical process with feedback loops and branches. You you really need a much more general situation and you can do to do dynamic programming on a die on a pretty general dive graph. That's what my thesis was about how to do that. It looked interesting little story here at the same time a very famous applied mathematician to time by the name of Gus an heiress at Minnesota was working on this very same problem. We will papers at about the same time and I got to referee his paper. Well he was a very famous guy but his paper was wrong and and and so I got into my first conflict. What goes on in. In research pretty early on but fortunately it was is resolved successfully and soon thereafter I wrote a little book on that I am a programming. Into programming it's been a field I've worked on all these all these last many years. And here's how it got started I heard of POC by early person operations research by the name of city pass and Sid came to Hopkins to give a seminar and talked about some pro bono work he was doing on a school district thing. Well and he said would you like to work on this kind of problem. And at the time it was really explainable destructing was really exciting. The Supreme Court had just passed its one one man one vote on the law and many of the states particularly those in the south were incredibly far from achieving that in Georgia for example there were Congress districts congressmen who represented fifteen thousand people and another congressman who represented six hundred thousand people and so the Supreme Court said all these districts had to be withdrawn and had to be redrawn and this was a natural inner to programming problem to be set up as a set partitioning model with with columns being instead inspectors of districts and the cost of a column could relate. For example to how far it was. From being a perfect district with regard to pop population on we. I got my student Bob Garfunkel involved in this problem man and and we developed an implicit in numeration algorithm no linear programming for solving this problem gotten got us both of us really interested in nature programming and in the early seventy's we were an intra programming book. I first N.S.F. but soon after that I left Johns Hopkins and moved to Cornell and really got involved in programming research my first famous at Grant was on packing partitioning covering problems and and here we began seriously to use structure to solve it in or to program problems with branch and bound algorithms that use linear programming we learnt we knew we had a pipe in the linear programming relaxations. That led to to customize cuts involving things like leaks and lift all that holds on and we got and this work was done with my student Les Trotter this was his thesis and and we got a very nice little theorem which said which seems to only apply to this problem when you solve the linear programming relaxation of a no packing problem just the edge constraints all the variables that are zero one in the relaxation can be fixed at those values. While we working on this focus and joined our faculty and he came up with this family of energy programs and he said George and I don't really do computation or programming but here's a really a better really hard problem that you can't solve and he gave us the Steiner triple system problems which I still persist as extremely difficult problems to solve today. Marshall Fisher was a visitor at Cornell and in the mid seventy's said he came along with a very. You know I see a problem that involve maximizing float in those days in those days the number of days it took to clear a check written from one bank paid to another bank the pen did a poll on how far they were apart and so for example if you if someone paid you a check. We're going to back in San Francisco and your bank was in New York it might take three or four days to clear. Well the big brokerage houses like Milledge the time realize that they could keep their money a little longer by paying their clients in New York from banks on the West Coast and vice versa and a question but you can't have located too many banks the question was to we don't have too many accounts of the question was were to locate and this is a became a classic facility location problem. Marshall and I got Gerard corner shows was another Ph D. student of mine as jointly to work on on on this stuff and we got some of the very early results on approximation algorithms using LP duality at that point I moved to Belgium to set up our operations research and you kind of metrics for a couple of years. I was there as the their research director. Speaking and working seriously with warrants Wolsey we extended the facility location work to maximising submodule functions and sure I'd had his Ph D. thesis on this topic and the the the original paper received the land just a prize. From. Then from or so of course now informs. Continued to work very heavily on energy programming throughout the eighty's during a good part of this time I was chair of the I.E.E.E. or department at Cornell and so had a little bit less time for research but really pushed forward on on writing and into programming and commentary a lot them say she and with Lawrence will see which which wrists from which. We also received on just a prize. At the time I was intrigued by Cottle's interpretation on of government cuts. You know that the idea that you could get comedy cut simply by rounding and so when writing the book. Lawrence and I said well you know what. What's the similar interpretation for mixed energy programs and we invented this idea of mixed in your giraffe ending which has become my a very important way I think of generating cuts or mixing to programming problems and continue to be used. Today and more and more I got involved in computation or programming there was there was no it was difficult to to to really do computational research because the codes that were out there didn't let you play with pieces of the code and so we had the idea. We need to build the research code for into programming one where you could really get into the guts of the code and change the branching rules change your cutting plan strategies whatever you want to do and so I started this project with a student called Gabrielson is Monday and at the same time. Morten saw this bird. Was at Georgia Tech I'd come to Georgia Tech by this time. Sal his perp was a post-doc a children's AK. And in the Minto. Code was created by a graduate student at the time was I have good and he he also pretty help write this code and of course he's the guru you've got a Ph teacher to tack and he's the do a good Robi that you all now know and also was very influential in writing and in writing the C. plex IP code which which had its early roots from Mento. In the late eighty's. Ellis Johnson joined our faculty at Georgia sat game. And and Ellis was had been working on airline problems. Well one nice thing about being in Atlanta as we had dealt there in the backyard and even nicer thing was the current the sky C.E.O. of Delta at the time was and I asked why you graduate and. And so we were able to get into Delta Delta at the time was using that crazy eighteen to the car market code to do its crew optimization that was. You know that code didn't even have an eye piece of our real life be solved already. And so we got very involved with Delta and then also with American Airlines and it's a united. Northwest and began to do serious research on on new models new solution techniques crew scheduling fleet scheduling disruptions became a we I think we did some of the very early work on disruptions and were able analyze and such solutions with the simulation model we built a very aligned operations this work that was done with. Cindy Boren her and had joined the our faculty then and she worked with us on on on this stuff and on the simulation stuff we had and from Cloud dissipated along with many graduate students. Well the work going on airline crew scheduling. Let us immediately to think more about how do we stop an intra programs with a huge number of columns and that led us to to the notion of well we named it branch and price other people had been doing similar stuff for example of the group and in Montreal. That this kind of approach into programs also proved fruitful in sports scheduling problems. I was the first for some time Georgia Tech's faculty representative to the A.C.C. in the N.C.A.A. and I heard about they were having difficult time doing their sports scheduling as T.V.. Began to the the E.S.P.N. and so on began to press harder for getting their games to the primes they wanted them. And so I I heard that my trip former Georgia Tech Ph D. will who is now at Carnegie Mellon was there when some work with Major League Baseball. We got into contact Kelly Easton Ph D. student George it's accurate or pieces on this topic and this eventually led to a couple little company we now have the sports scheduling group and we schedule major league baseball among others. Well that the work on computational intra programming has continued to the present. Most recently. Most recently Been thinking about how do to to deal with problems of uncertainty new to programming and together with my colleagues should be or are made in students. We've been working on polyhedral approaches to stochastic in order to programming. Even more I still want to work on on applied problems and one of the most fascinating ones I've worked on in the last few years an idea that that probably was before its day because the company is no longer in operation but on demand air transportation where you've got to solve it or two programs you change your schedule on the fly with each reservation you've got a chip solve it or two programs enormously quickly and we've built of scheduling it's scheduling software scheduling on it and software for Day Jet more recently work in heavily with X. on mobile in inventory routing and another copy attentional problem I like very much is I salute that we know how to branch out into programming so we're trying to work on learning methods or artificial intelligence learning methods for doing that. OK let me just close quickly with what I had in the last leid really mimics what I think we don't know how to do because that's what and and and that includes real time energy. Programming. You know you might have to solve a series of problems all very close to each other they differ slightly but they got to be solved in pen seconds. How do you solutions that you currently have. To to get new solutions and certainly and so uncertainty in energy programming is extremely difficult as some very nice work on robust optimization of course by by a. Number of skin. Ben Pollin see Miss but But what about the the chance constrained integer programming stochastic into programming. Those are still open questions as is extending much of the work we've done in little you're new to programming from non-linearity to program. Well like to thank thank you all for for telling this talk and again apologize for not being here. But I walked in questions and comments just sent me an email my email that you got my new address right there. Thanks again.