Forgot your password?
typodupeerror
Math IBM Technology

IBM Using Complex Math To Manage Natural Disasters 115

Posted by Soulskill
from the stopping-a-flood-with-natural-logs dept.
coondoggie brings us a NetworkWorld story about IBM's efforts to use complex algorithms to manage responses to natural disasters. Researchers are making use of recent increases in processor speed and algorithm efficiency to develop a scalable, flexible model capable of handling the complicated planning involved in reacting to a crisis. Quoting: "'We are creating a set of intellectual properties and software assets that can be employed to gauge and improve levels of preparedness to tackle unforeseen natural disasters,' says Dr. Gyana Parija. 'Most real-world problems involve uncertainty, and this has been the inspiration for us to tackle challenges in natural disaster management.' In the case of flooding, for example, the stochastic programming model would use various flood scenarios, resource supply capabilities at different dispatch locations, and fixed and variable costs associated with deployment of various flood-management resources to manage various risk measures. By assigning probabilities to the factors driving outcomes, the model outlines how limited resources can meet tomorrow's unknown demands or liabilities. In this way, the risks and rewards of various tradeoffs can be explored, IBM said."
This discussion has been archived. No new comments can be posted.

IBM Using Complex Math To Manage Natural Disasters

Comments Filter:
  • It won't save us (Score:5, Informative)

    by Ethanol-fueled (1125189) * on Thursday April 03, 2008 @07:02PM (#22958426) Homepage Journal
    High-performance computing won't save us from idiots in high places, even during natural disasters. Case in point: Michael D. Brown [wikipedia.org], who was in charge of FEMA during the Hurricane Katrina ordeal. From the wikipedia article:

    "...Some members interviewed felt Brown showed an imperious attitude, and nicknamed him 'The Czar'."

    Heckuva job, Brownie! More optimistically, I hope that their algorithms could predict the 4 or 5 "wild"- fires in Southern California which are all started mysteriously(on the same day) "in season."
    • Re:It won't save us (Score:5, Interesting)

      by Carnildo (712617) on Thursday April 03, 2008 @07:11PM (#22958526) Homepage Journal

      More optimistically, I hope that their algorithms could predict the 4 or 5 "wild"- fires in Southern California which are all started mysteriously(on the same day) "in season."


      Shouldn't be too hard. One of the things they discovered while studying line noise in telephone circuits is that the cause of the noise doesn't matter: it could be induction from nearby motors, bad connections influenced by the wind, or short-circuits triggered by someone dropping a screwdriver -- it all fits into the statistical patterns. In the case of fires, it doesn't matter if it's lightning, arson, or volcanic eruption, the pattens still hold.
      • Re: (Score:3, Funny)

        it all fits into the statistical patterns. In the case of fires, it doesn't matter if it's lightning, arson, or volcanic eruption, the pattens still hold.
        It leads one to wonder whether being statistically significant, is itself statistically significant?
        • by gstoddart (321705)

          It leads one to wonder whether being statistically significant, is itself statistically significant?

          Only if you can do it enough times. :-P

          Cheers
      • Re: (Score:3, Insightful)

        by Lurker2288 (995635)
        What you say may be true, but I can't help but think of Nassim Nicholas Taleb and his black swans. He would argue that while your wild fires might fit the existing statistical models, a REAL disaster, the kind with the potential to really knock the country on its ass, would be essentially impossible to predict. I'm not sure I entirely buy into his thinking, but you have to admit, it's usually the unexpected stuff that produces the greatest impact.
    • Re: (Score:1, Insightful)

      by Anonymous Coward
      I wonder if it's a coincidence that this news came just a couple days after the EPA announced it was banning IBM from bidding on future contracts.

      "Hey, look at what our scientists are coming up with using advanced supercomputer models to forecast hurricanes, forest fires, and climate change! Here are the predictions for the rest of 2008, 2009, 2010..." (holds printouts close to face) "...wow, these are potentially devastating consequences for the ol' USA unless our government leaders are unusually skill

    • by c6gunner (950153) on Thursday April 03, 2008 @09:19PM (#22959408)

      Heckuva job, Brownie! More optimistically, I hope that their algorithms could predict the 4 or 5 "wild"- fires in Southern California which are all started mysteriously(on the same day) "in season."
      What's your guess? Blackwater, with black helicopters, in the forest? I'm going with Col. Mustard, with the candle-stick, in the grasslands.
    • by gnick (1211984) on Thursday April 03, 2008 @09:49PM (#22959614) Homepage

      I hope that their algorithms could predict the 4 or 5 "wild"- fires in Southern California which are all started mysteriously
      If there's enough money in it, I can predict a mysterious fire just about anywhere at any time.
    • More optimistically, I hope that their algorithms could predict the 4 or 5 "wild"- fires in Southern California which are all started mysteriously(on the same day) "in season."

      The algorithms aren't designed to predict disasters, but to manage the response to disasters in progress. (Which isn't actually very easy.)
    • by hey! (33014)
      Well, to be a contrarian, Brown may have deserved censure, but I think criticism of him is exaggerated. In some ways, his most distinctive failings were political.

      True, the real issue was that FEMA wasn't ready, and he was supposed to be in charge making sure it was ready, but it's not like he's just one bad apple in the bunch. He might not have recommended enough, but the administration didn't even back him in what he had recommended.

      As for his imperiousness, well, what do you expect? When the higher up
  • by Runagate Rampant (602123) on Thursday April 03, 2008 @07:02PM (#22958430)
    sqrt(-1) = natural disaster!
  • Government (Score:4, Insightful)

    by boris111 (837756) on Thursday April 03, 2008 @07:04PM (#22958446)
    Can they model government indifference to the people's plight?
    • Easy (Score:5, Funny)

      by EmbeddedJanitor (597831) on Thursday April 03, 2008 @07:17PM (#22958584)

      if(contribution_of_lobbiests_impacted > 100000000)
      do_something(); /* better respond to them */
      else
      ignore_poor_people(); /* who cares */
      • Re:Easy (Score:4, Funny)

        by Anonymous Coward on Thursday April 03, 2008 @07:27PM (#22958660)
        PeopleWhoUseUnderscoresInTheirVariableAndFunctionNames = null; /* Okay, I'll get off your lawn. Sheesh. */
        • You didn't use Hungarian.
        • by dodobh (65811)
          person.add_attribute("make_reading_difficult") if (variable_names.get_style == camelCase);
    • Re: (Score:3, Insightful)

      Why is P modded 'flamebait'? He asks a relevant question. The response of government officials may unfortunately be the biggest factor in calculating how to deal with a disaster.
      • I was wondering the exact same thing. Flamebait seems a little harsh for a valid question.

        The mods are being rather abrasive and cruel. Maybe the new page style is setting them off?

    • Can they model government indifference to the people's plight?
      Is there any point to modeling another type of government?
  • by Carnildo (712617)
    Complex math? Aren't real numbers good enough for the job?
    • by sapphire wyvern (1153271) on Thursday April 03, 2008 @07:09PM (#22958510)

      Complex math? Aren't real numbers good enough for the job?
      You need the imaginary axis to quantify FEMA's competency. :)
    • by jdagius (589920)
      >> Complex math? Aren't real numbers good enough for the job?

      RTFA. It's not about imaginary numbers. That's "complex" as in "complicated" math, specifically stochastic algorithms, which are probability contrained problems dealing with random variables!
      • Re:Complex math? (Score:5, Informative)

        by jonadab (583620) on Thursday April 03, 2008 @11:51PM (#22960194) Homepage Journal
        > That's "complex" as in "complicated" math

        In mathematics, "complex" does not mean complicated any more than "proper" means correct or "rational" means sane or "group" means any old gathering or collection. These words have very specific meanings in mathematics, and using them for their general-English meaning, in the context of math, is at best confusing and at worst outright misleading.

        You can talk about a "complex algorithm", and people will generally understand you mean a complicated one, because the word "algorithm" lends more of a computer-science context. You can say "complex way of doing things" and convey the idea of complicatedness, because "way of doing things" is sufficiently general that it doesn't really imply any particular context at all. But saying "complex math" very much conveys the idea of the use of complex numbers (i.e., numbers with a real part and an imaginary part, either or neither or both of which may be zero for any given number) because the word "math" strongly implies a mathematics context and draws the math-jargon sense of the word "complex" to the forefront. Only someone who doesn't *know* what the word "complex" means in mathematics would think of any other meaning.

        It's like saying "hedge fund" and expecting people to get the idea that you're collecting money for shrubberies. Only someone with no idea what a hedge fund is would get that impression.
        • Re: (Score:1, Troll)

          by EveLibertine (847955)
          I wasn't confused, and neither apparently were most of the people in this thread, let alone everyone else responding who didn't mention this gaff. Your complaints are nothing more than nerd pedantry. Also, your analogy is awful. Furthermore, you are complaining about the title of a slashdot post, and it is well known that the editors hardly pay attention to the titles of these posts, let alone the content that follows.

          In short, I'd like to congratulate you on your Perfect Slashdot Post.
        • It's like saying "hedge fund" and expecting people to get the idea that you're collecting money for shrubberies.

          We are the Knights who say 'Ni'!

    • Re: (Score:3, Informative)

      by jdagius (589920)
      I repeat: RTFA. This so-called "complex math" is _not_ about imaginary numbers! It's about stochastic programming and _complex algorithms_ (i.e. complicated). Why don't you mod me up (for a change) so you people can get this straight.

      "The idea is to use high-level math techniques, which IBM calls Stochastic programming, to help speed up and simplify complex tasks "
  • Patents (Score:4, Insightful)

    by LeoDavinci578 (795523) on Thursday April 03, 2008 @07:10PM (#22958516)
    Oh man, I just loved this: "We are creating a set of intellectual properties and software assets that can be employed to gauge and improve levels of preparedness to tackle unforeseen natural disasters"

    Awesome, now they get to patent how to respond to natural disasters so that no one else can innovate... another victory for our wonderful patent process!
  • by icegreentea (974342) on Thursday April 03, 2008 @07:15PM (#22958556)
    Before everyone starts smarting at this thing 'predicting' natural disasters, please read the summary carefully (I know your not going to read the article). The math and system is designed to help deal with natural disasters that do happen (like optimizing your relief delivery path, plotting the best places to contain/fight a forest fire, etc etc). It is also used to evaluate how best, and how well current resources could be used in a natural disaster by predicting (yes, there is it, a prediction) most likely challenges, problems, scale and the like. I think it's useful.
    <p>
    The new thing with this apparently is that they're using a new mathematical model that previously was too computationally expensive to do on a large scale. Computers are powerful enough to use these models now.
    • And no matter how mathematically or computationally complex, much boils down to the quality of data and set of assumptions employed in the model.
    • by pipingguy (566974) *
      Sounds like a money sink-hole once the beaurocrats get a hold of it. In the old days we used to be able to do this sort of stuff with pencils and a really large napkin.
  • Clearly complex math needs to be computed on an iMac.
  • Natural disaster may be quantifiable, but do we really want some heartless machine deciding who lives and dies in the case of an emergency? Anyone see I-Robot?

    I don't mean to sound like a stereotypical paranoid geek, but we give too much power to machines they will start controlling our life.
    • Re:Quantifiable (Score:4, Interesting)

      by iminplaya (723125) <iminplaya.gmail@com> on Thursday April 03, 2008 @08:11PM (#22958906) Journal
      ...but we give too much power to machines they will start controlling our life.

      You get either a machine or a bureaucrat. Take your choice. At least with a machine, you can turn it off. Just try to get rid of an incompetent bureaucrat or crooked politician who appoints him.
    • The current system (letting nepotically chosen subordinates with no experience in their field) doesn't seem to be working out that well, either. At least with a computer program, you can examine the criteria and decide if it makes sense.

      It something like triaging patients in a mass casualty - in the case of limited resources and nearly unlimited casualties, you spend the doctors & nurses time where it will do the most good. Saving a 5 year old with massive head trauma may make you feel all warm and fu

  • Obligatory (Score:3, Funny)

    by fixer007 (851350) on Thursday April 03, 2008 @08:15PM (#22958942)
    I was told there would be no math...
  • by Anonymous Coward
    In the case of a natural disaster please check with your lawyer before responding.

    You may be infringing on a patent...
  • by creimer (824291) on Thursday April 03, 2008 @08:24PM (#22959010) Homepage
    Natural Disaster + Presidential Photo-Ops = Great PR!
    • by aleger (741695)

      Natural Disaster + Presidential Photo-Ops = Great PR!
      I would insert line: rollupsleeve(all politicians)
  • I wonder how long it would take for Charles Eppes to show up...
  • by AJWM (19027) on Thursday April 03, 2008 @08:32PM (#22959068) Homepage
    From TFA: "to tackle unforeseen natural disasters"

    But then it goes on to talk about mostly foreseeable natural disasters. If you live on a flood plain or a low-lying coastal area subject to hurricanes, you're going to get flooded. In an earthquake zone you're going to get earthquakes. Lot of vegetation in an area that has dry spells, fires. And so on.

    Legitimately unforeseen natural disasters would be things like a comet impact [imdb.com], volcanoes erupting in downtown LA [imdb.com], or perhaps alien invasion [imdb.com]. Oh wait, that last would be an unnatural disaster, wouldn't it? But come to think of it, the ones I just mentioned have all been foreseen too.

    I guess I just don't foresee a need for this software. Maybe they should work on software for foreseeable disasters.
    • by lexDysic (542023)

      From TFA: "to tackle unforeseen natural disasters"

      Legitimately unforeseen natural disasters would be things like a comet impact [imdb.com], volcanoes erupting in downtown LA [imdb.com], or perhaps alien invasion [imdb.com]. Oh wait, that last would be an unnatural disaster, wouldn't it? But come to think of it, the ones I just mentioned have all been foreseen too. I guess I just don't foresee a need for this software. Maybe they should work on software for foreseeable disasters.

      What about the Spanish Inquisition? No one forsees the Spanish Inquisition.

    • by wgaryhas (872268)
      How about a tornado in Salt Lake City, Utah (way outside the tornado belt)? Happened once a few years ago.
  • Monte Carlo (Score:2, Insightful)

    by alexhard (778254)
    From what I could gather from the summary, this sounds like a glorified Monte Carlo simulation, not exactly something newsworthy..
    • Re: (Score:1, Informative)

      by Anonymous Coward
      Well, stochastic programming is most definitely not a "glorified Monte Carlo simulation", which is based on random trials, and not guaranteed to find any sort of optimum. In fact, stochastic programming isn't really simulation-based at all. It's more like a set of optimization techniques for solving problems with parametric uncertainties.

      The idea behind it is that parametric uncertainties that can be characterized using a probability distribution. The optimization algorithm itself is often deterministic. (e
  • Someone has been watching too many episodes of numb3rs....
  • Hat. Old. (Score:4, Interesting)

    by PJ The Womble (963477) on Thursday April 03, 2008 @09:10PM (#22959340)
    I used to do Delphi stuff (I know) for a firm of insurance actuaries. They were writing code for (essentially) predicting how long it would take to pay out for natural disasters. They had some very clever Stochastics in there, along with some nice triangulation/vector stuff too: I remember the names Bornhuetter and Ferguson (sadly it's been a long time and there's been the odd small sweet sherry since, so life isn't that clear recently).

    What I do remember though, is that I mentioned to my superiors that a case-based reasoning engine would take a lot of the (non-discrete) math out of the whole thing. Because things happen and we learn from them. Has the nature of nature changed, or was I wrong in the first place?
    • Re: (Score:3, Insightful)

      by PJ The Womble (963477)
      It's just occured to me that my comment above is a nearly a good example for the discussion (on here? maybe) the other day about the desirability of more complex algorithms, versus the greater and greater amounts of data available, when data mining. Any thoughts?
      • Cripes, I can't buy a response on here these days! It's 03.45 here now, the wife's in bed and I'm to all intents and purposes stuffed. Who could have predicted that? Nor Bornhuetter and Ferguson for sure.

        She's a foot shorter than me but that stuff about a good big 'un beating a good little 'un every time is so much foo!

        Wasn't it Peter Cook, in his incarnation as Sir Arthur Streeb-Greebling, who said "Certainly I have learned from my mistakes. And if I had to start all over again, I'm sure I could repea
        • by Raenex (947668)

          Cripes, I can't buy a response on here these days!

          Ok, I'll bite. Sorry to ruin it for those who would rather watch you twist in the wind.

          The simple answer is that case-based reasoning is limited due to it's static nature. If you can boil your system down to "if this, then that", great. However, many systems are too complex or dynamic for that kind of analysis, and benefit from statistical techniques. Weather forecasting is a classic example. Even "expert" systems, such as for medical diagnosis, make use of probability.

          This is a huge field, and the ri

  • And so the Federal government will spend $millions, probably $billions, on no-bid contracts for all kinds of fancy gear, but still won't fix the levees protecting New Orleans.
  • by theodp (442580) on Thursday April 03, 2008 @10:33PM (#22959864)
    And thanks to IBM, responding to a crisis of 9/11 or Katrina magnitude strikes may constitute patent infringement. Big Blue has a patent pending for Optimizing the Selection, Verification, and Deployment of Expert Resources in a Time of Chaos [uspto.gov], which covers responding to 'episodes of profound chaos during hurricanes, earthquakes, tidal waves, solar flares, flooding, terrorism, war, and pandemics to name a few.' It's apparently this easy [flickr.com].
  • by YetAnotherOnlineAcct (1267424) on Friday April 04, 2008 @01:19AM (#22960498)
    Back in the mid 90's I was working on some logistics systems. I remember seeing a software package from some Swedish company that used stochastic programming. The software did spare parts optimization. Tell it what sort of spares you need (Jet engines, parts for oil well), where you need them (O'Hare, an oil rig), how much they cost (I don't know!), how much they cost to store, where you have warehouses, how much storage costs, how long it takes to get one if you need to order it from Boeing or something, what's the interest rate if you need to borrow the money to buy the parts, how long it takes to get from the storage site to where it's needed, and of course the failure rate of the parts, and a few more things I can't remember......

    In other words, it took LOTS of data.

    But, once you entered the data, you could tell the system "I want 99.999% uptime" and it would give you the most cost efficient way to buy and store the parts needed. Or you could start with a budget and find out what sort of availability you could afford. Depending on the size of the operation and criticality of the availability you could save a lot of money or really help with availability.

    It sounds like this software does something similar. In the spares optimization, you don't know which airport your plane will be at when it's engine needs replacing. In the disaster scenario, you don't know where the disaster will strike. In the first case, you're optimizing spare part allocation. In the second you're optimizing recovery supplies and equipment. Either way you get the best probabilities you can and optimize the best you can.

    Yeah, I know. I read the article. AND posted an on-topic, very un-funny comment. I must be new here. Well, this user ID is new, at least.

    • Yes, I'm replying to myself, but I didn't really make my point well.

      My point is that the technique isn't new.

      Using Stochastic programming to solve this sort of problem isn't new.

      Using it to solve this sort of problem can be done, but takes a lot of data and isn't easy to use.

  • stupid (Score:4, Insightful)

    by nguy (1207026) on Friday April 04, 2008 @01:39AM (#22960580)
    "'We are creating a set of intellectual properties and software assets that can be employed to gauge and improve levels of preparedness to tackle unforeseen natural disasters,' says Dr. Gyana Parija.

    Many research groups are working on simulation and prediction of behavior, natural disasters, preparedness, etc. But the first words out of an IBM researcher's mouth are "intellectual properties and software assets".

    Shame on you.
    • Re: (Score:3, Interesting)

      by Lurker2288 (995635)
      Why shame on them? They're not government, and they're not academics. They wouldn't be doing the research if there wasn't a way to make a buck off of it. If you don't want to shell out for their knowledge or their 'software assets,' then feel free to turn to one of the many other research groups who aren't working for a profit.
      • by nguy (1207026)
        They're free to make profit, but it's still a lousy attitude when the first words out of their mouths are about "intellectual properties" rather than about the results.

        And with an attitude like that, you can be pretty certain that they are going to try and patent miniscule variations on well-known technologies and are going to make a legal nuisance of themselves.
  • IBM is a bit of a natural disaster all by itself.....
  • i am very suspicious that a large corporation such as IBM is beavering away creating and no doubt patenting "intellectual properties" that may cover critical ares of need with the advent of global warming, global dimming and the potential water level rises. Seems somewhat opportunistic when it would appear to be global PTSD. (Just think of Katrina on a global scale).
  • Oh, yet another SD article completely composed of BS! Geeez, can we stop with the "I'm using really, REALLY complex (math/algorithms/hardware/physics/chemistry/lollipops) so therefore, my research can do this impossible task that nobody else can do. What's funny, in a year or two, all is forgotten about said "complex algorithms" ...blah..blah...blah...
  • by Money for Nothin' (754763) on Friday April 04, 2008 @07:29AM (#22961696)
    The current financial market crisis is considered by many economists and finance professionals to be due in part to the failure of such computationally-intensive risk-management models as those it sounds like IBM is creating.
  • I think you'll find that Led Zeppelin identified the problem some years ago:

    "If it keeps on raining: levee's gonna break.
    When the levee breaks: mama, you got to move."

    FEMA have been using this model for some time now.....

    Perhaps it does need a rethink.
  • Too bad they can't sell this to the FEMA anymore.
  • There are already many large private industries that extensively study the most efficient ways to distribute materials in a responsive way.

    Lessons from the Private Sector and the Coast Guard During Katrina [mercatus.org]

    Private-sector planning for the storm began days ahead of landfall. On the Friday prior to the Monday landfall, Home Depot activated the "war room" at its Atlanta headquarters, negotiating with various vendors to get needed supplies staged to move into the hurricane zone. Wal-Mart's response began slightl
  • The underlying optimization models and algorithms were initially prototyped on a large unnamed US Government program, where the key problem was how to efficiently deploy a large number of critical resources to a range of disaster event scenarios.

    Gee... I wonder what large, unnamed program deploying large numbers of critical assets under crisis scenarios might possibly be...hmm?

    Folks are right that this is nothing completely new, but if they've found a way to speed the computation of it along ("within an hour"?) with less machine resources required... then that is a breakthrough, is it not?

    Yes, there has been this sort of research going on in academic institutions and elsewhere, private and public sectors, internationally (and if you look at

  • // req vars
    $color = get_skin_pigment($victims);
    $sector = get_sectors_affected($disaster);
    $cost_per_minute = current_price_of_advertising();
    $bling = $db->get_var("SELECT (SUM donated AS bling) FROM data WHERE (in_array($sector))");
    $replair_cost = $num_dead * $cost_per_minute; // conditionals
    If ($color == #ff0000) {$pr = FALSE;} else {$pr = TRUE;} // logic
    If($repair_cost > $bling && $pr) {fix_the_problem();}
    else {call_larry_king();}

Lo! Men have become the tool of their tools. -- Henry David Thoreau

Working...