Become a fan of Slashdot on Facebook

 



Forgot your password?
typodupeerror
Space Games

TheSpaceGame — Design Your Route To Jupiter 76

Posted by Soulskill
from the try-the-clarke-kubrick-corridor dept.
An anonymous reader writes "The Advanced Concepts Team of the European Space Agency is celebrating World Space Week (4-10 October 2010) with the release of 'The Space Game,' an online game for interplanetary trajectory design. The Space Game is an online crowdsourcing experiment where you are given the role of a mission designer to seek the best path to travel through space. The interactive game, coded in HTML5, challenges the players to devise fuel-efficient trajectories to various bodies of the Solar System via a user-friendly interface. The aim of the experiment is get people from all ages and backgrounds to come up with better strategies that can help improve the effectiveness of the current computer algorithms. As part of the events organized worldwide for Space Week, the first problem of the game is to reach Jupiter with the lowest amount of propellant. The best scores by 10 October will be displayed on the Advanced Concepts Team website and the three best designs will also receive some ESA prizes."
This discussion has been archived. No new comments can be posted.

TheSpaceGame — Design Your Route To Jupiter

Comments Filter:
  • by White Flame (1074973) on Monday October 04, 2010 @09:58PM (#33791080)

    It can make the Jupiter run in less than twelve parsecs.

  • Waste of time (Score:5, Insightful)

    by Angst Badger (8636) on Monday October 04, 2010 @10:07PM (#33791128)

    This is exactly the kind of combinatorial optimization problem that is superbly well-suited for solution by software and quite possibly the last kind of problem you want to hand to a bunch of humans, unless those humans happen to be programmers with backgrounds in celestial mechanics, heuristics, and genetic algorithms.

    As a way of driving public interest in the ESA's space program, it's not a bad idea at all, but if any of its users manage to come up with a better solution than the ESA's software, it's not a triumph for crowdsourcing, it's a sign that the ESA needs to hire new programmers.

    • Re:Waste of time (Score:5, Insightful)

      by Anonymous Coward on Monday October 04, 2010 @10:49PM (#33791352)

      They mention Monte Carlo by name in their video. If they know about random simulations formally, surely they know about genetic algorithms. They say at the bottom of the home page,

      We do not claim that computers are not able or are particularly bad at solving such problems. Rather, we think that 'watching' humans design complex interplanetary trajectories can be of help to improve the intelligence of computer algorithms.

      This is for publicity and for fun. It's the only explanation that makes sense without more information.

      Also, it's a decent example of the sort of thing possible with HTML5 crap, and it's GPL, so at least it's got that going for it.

      • I was really excited until I saw

        Browser Check: Check you browser

        Now I'm still excited, but it got knocked down a notch.

        • by mwvdlee (775178)

          Don't worry, it'll work on every browser except MSIE. Worked fine on Chrome6 for me. And it's a fun game to mess around with. I wish they had more tutorial missions though.
          p.s. While on the subject; is MS planning to support HTML5 canvas in IE9?

    • Since it is optimizing positions which can be easily encoded as floating point numbers, I would use Differential Evolution for the optimization, or Particle Swarm Optimization (or both).

    • Re: (Score:3, Interesting)

      by Synon (847155)

      This is exactly the kind of combinatorial optimization problem that is superbly well-suited for solution by software and quite possibly the last kind of problem you want to hand to a bunch of humans, unless those humans happen to be programmers with backgrounds in celestial mechanics, heuristics, and genetic algorithms.

      As a way of driving public interest in the ESA's space program, it's not a bad idea at all, but if any of its users manage to come up with a better solution than the ESA's software, it's not a triumph for crowdsourcing, it's a sign that the ESA needs to hire new programmers.

      Yes yes we get it already, computers will always be better suited for solving these kinds of problems. As such, I would like to point you to the front page of their website which states this-

      "The Space Game is a game and a crowdsourcing experiment run by the Advanced Concepts Team of the European Space Agency aimed to improve the methods for designing interplanetary trajectories. We do not claim that computers are not able or are particularly bad at solving such problems. Rather, we think that 'watching'

    • by idji (984038)
      it's buggy too. I have Leg number "NaN" with flyby altitude of 12 digit integer. I feel all i am doing is minimax. I spend 95% of my time finding the min position of a planet in an orbit - they could automate that - and the user then spends there time looking more at orbits than the 3rd decimal place after delta-V. They could also let it run 500 times in the browser to find a few low energy starting candidates. There should also be a visualization of the variant orbits so I can move closer to that global m
    • Dumbed-down GTOC (Score:5, Informative)

      by Zoxed (676559) on Tuesday October 05, 2010 @04:06AM (#33792452) Homepage

      Although I work in the Space Industry I am not a Rocket Scientist but it is my understanding that this is *not* a purely computer solvable problem and is explained on the GTOC website: http://www.esa.int/gsp/ACT/mad/op/GTOC/indexII.htm [esa.int]

      • by vlm (69642)

        it is my understanding that this is *not* a purely computer solvable problem

        since there is no obvious "best trajectory" for many deep space missions

        It is computer solvable but there are two problems:

        1) The ancient GIGO garbage in garbage out problem... Without a full description including solar sail effects, differential outgassing, etc, you can be pretty far off. Read up on the pioneer anomaly, not specifically for that anomaly but to see what all has to be included... lightwave IR radiation pressure from the hot parts of the spacecraft, drag from the solar wind, etc. All at best semi-predictable.

        http://en.wikipedia.org/wiki/Pioneer_anomaly [wikipedia.org]

        2) Overly

  • I was going to bash this and say they could just take the money and feed random trajectories in to the formula and get the same results, but I got to thinking, this is actually good. This can give you the top 100 or so options then you spend the cpu time tweaking this or that variable. You may actually get some better results that doing raw multivariable calculus with a lot of variables and unknowns could do.

  • by Laser Dan (707106) on Monday October 04, 2010 @10:08PM (#33791136)

    Crowdsourcing for this may be a good bit of publicity, but is really just wasting time.
    A genetic algorithm running on their "simulation" will find the best solution within the accuracy of the parameters very quickly. Run a couple of times to make sure it is the global minimum and you're done.

    Their competition has a hard limit on mission duration and the goal is minumum delta-v, so the fitness function is very easy to define.
    If anyone wants to win the competition, figure out how to write parameters to their simulation and read the delta-v and mission duration, run a GA for a while and you automatically win.

  • ... the first problem of the game is to reach Jupiter with the lowest amount of propellant ...

    This seems like a trivial thing to do. At escape velocity give engines a little burst, coast to the orbit of the destination, a little burst to decelerate, wait for the planet to get to that point. OK, its not very efficient with respect to time but that wasn't a stated criteria. Game design is hard, even for rocket scientists.

    • by joe_frisch (1366229) on Monday October 04, 2010 @10:26PM (#33791242)

      Not so simple. You can gain a lot by getting boosts off of other planets and moons. I suspect that minimum is just under earth escape velocity, with a boost from the moon. Then over lots of orbits you can use earth flybys to modify your orbit. Might then be a win to use either Venus or Mars. The optimal path might take a VERY long time.

      Don't forget the trick that burning your fuel deep in a gravity well is a big help - a near-solar flyby might also be an efficient route.

      Of course the "right" way to do it is just use higher ISP engines and go direct so you get there quickly and don't need to wait half a generation to get your data.

      --- Joe Frisch

      • by vlm (69642)

        The optimal path might take a VERY long time.

        Assuming an infinite supply of maneuvering thruster fuel, infinitely long lived electronics, infinite RTG half life, infinite radiation shielding, etc.

        Also theres some lovely and interesting theoretical orbits that unfortunately involve passing beneath the "surface" of the sun or the gas giants. Admittedly surface is a vague concept, but if you have to use aerobraking calculations you're probably doing it wrong since you're turning valuable velocity into heat. Or would be a perfect slingshot around Mars e

    • by Yvan256 (722131) on Monday October 04, 2010 @10:35PM (#33791296) Homepage Journal

      Game design is hard. Let's go shopping!

    • ... the first problem of the game is to reach Jupiter with the lowest amount of propellant ...

      This seems like a trivial thing to do. At escape velocity give engines a little burst, coast to the orbit of the destination, a little burst to decelerate, wait for the planet to get to that point. OK, its not very efficient with respect to time but that wasn't a stated criteria. Game design is hard, even for rocket scientists.

      Oops, I think I got that mixed up a little. IIRC its accelerate to maintain the destination's orbit, deceleration would be to get captured by the destination itself.

  • by imsabbel (611519) on Monday October 04, 2010 @10:20PM (#33791210)

    I mean, anything beyond jupiter would be a challenge. But jupiter itself? Hohmann transfer orbit, maybe with a sling around mars (would give very very low boost in deltaV, so not worth the launch window constrains IRL but ok for this)...

  • by assemblerex (1275164) on Monday October 04, 2010 @10:32PM (#33791276)
    You didn't need that DNA intact, did you?
    • What kind of radiation, and is it really that bad? Sorry, I'm not all that familiar with the Jovian system.

      • Europa is appx. 420,000 miles from Jupiter. Europa receives appx.1,000,000 rads/day from Jupiter. 800 rad is fatal.
  • I'm very confused why a computer couldn't just iterate through millions of iteration of the equation (http://en.wikipedia.org/wiki/Lambert's_problem) and pick the best.

    Heck you could even use my old game: http://code.google.com/p/exoflight/

    • by Anonymous Coward

      Their approach is even more difficult than you think. My understanding is that, first you have to find a privileged class citizen that has lost the use of his legs, then you have to substitute for him professionally, using his DNA samples to fool the screening systems. Do that for along time, and with a little right-handed luck, you'll be able to make your projected run. Maybe get a hot girlfriend in the process.

    • by Legion303 (97901)

      I thought of that as I was running the tutorials, but then I saw the competition. Calculations for one or two gravitational influences aren't hard, but once you hit three it starts to get hideously complex. The competition has five influences (Earth, Venus, Mars, Jupiter, and the sun) acting on the craft.

  • by gmuslera (3436) on Monday October 04, 2010 @10:55PM (#33791374) Homepage Journal
    is zero. Just need to get the right vehicle for that, i.e. one that looks like a monolith full of stars.
  • ...a release of slingshot [slashdot.org] implementation in HTML5?
  • Vernor Vinge? (Score:3, Interesting)

    by pontifier (601767) on Monday October 04, 2010 @11:23PM (#33791486) Homepage
    It sounds very similar to a game that was described in The Peace War by Vernor Vinge.
    • cool; I'm currently working my way through Marooned in Realtime, a sequel that stands on its own IMHO; ought to see how it goes with the original.

  • by DougF (1117261) on Monday October 04, 2010 @11:23PM (#33791490)

    ...the first problem of the game is to reach Jupiter with the lowest amount of propellant.

    I hate to be pedantic, but is the objective to arrive at Jupiter WITH the lowest amount of propellant, or is the objective to arrive at Jupiter USING the lowest amount of propellant? I suggest there is a big difference between the two.

    • Re: (Score:3, Funny)

      by Stihdjia (1870316)

      Thank you for your worthwhile contribution to Slashdot!

      I, too, was boggled by whether they meant fuel-efficiency, as previously stated in the article, or if this was a contest to find out who could design the most pointless trajectory. Will this puzzle ever be solved???

      • by tehcyder (746570)
        I had a go, and tried to create a trajectory that would use up ALL THE POSSIBLE FUEL IN THE UNIVERSE, just because I am evil.
    • by toolie (22684)

      That question doesn't mean you are pedantic, it means you are incapable of using common sense.

  • For the competition journey. What's a good figure?
    • I got 16.98km/s, which at the time was 9th in the rankings. It was of course quickly pushed off; #10 right now is 15.91km/s. Johannes Schnabl has had the top spot with 10.67km/s the whole time I've been looking. I wonder if he's one of these geniuses upthread running a GA?
  • Didn't we already do this, with a penguin? http://www.bigideafun.com/penguins/arcade/spaced_penguin/default.htm [bigideafun.com]
  • by seanthenerd (678349) on Tuesday October 05, 2010 @12:40AM (#33791762) Journal
    Check out Orbiter [ucl.ac.uk] - without a doubt the most realistic (and incredible) space flight simulator around. My little brother has basically taught himself orbital mechanics using Orbiter and online tutorials for the game (if you can call it that!) The real deal - Hohmann transfer orbits and spaceflight mechanics-type concepts I'd never heard of.

    When I saw "the space game", I thought for sure they were talking about Orbiter. If "designing your own route to Jupiter" is something you're interested in, do yourself a favour and check it out.
    • I'm with you. Orbiter's the best free program out there. AGI.com's STK software suite is the only commercial orbital navigation planning software I know of. A lot of defence contractors use it, as well as DoD, NASA and the AirForce. It's also used for Geospatial Intelligence and Space Battle Management.
    • by savanik (1090193)

      The TransX plugin can help you plan out the proper mission plan, too, with burn times, Delta-V requirements, et al. There's a bit of a learning curve involved, but nothing that taking a couple hours in the tutorials won't fix.

      I don't think Orbiter models effects like solar sails or thermal thrust yet, though.

  • by BlueStrat (756137) on Tuesday October 05, 2010 @03:21AM (#33792292)

    You'd have to do it in two launches, it won't work with only one.

    The first launch needs to terminate with significant mass at many miles per second in Washington D.C. prior to the Jupiter launch in order to prevent the whole Jupiter project being killed halfway through planning.

    Hey, just sayin'.

    Strat

  • In no timeframe? (Score:2, Interesting)

    by jlebrech (810586)

    So basically just avoid the other planets along the way, and go the slowest speed imaginable?

  • in the challenge, everytime i get the delta-v low (around the 22-25 mark), i'll end up moving one planet slightly wrong and the game just completely changes the trajectories and i'll have to spend another ten minutes tweaking it down to 25 again...

    nice idea and all, but the implementation just annoyes me enough to give up after a few tries

  • now just to go in the garage and find that warp pod i salvaged from the Ur-Quan dreadnought that crashed in my backyard.
    Prerequisites: start with 0 fuel.

    Step 1:hit quasi space and arrive as close as possible to our destination. (so far we have used 0 fuel.)
    Step 2: use Umgah Caster to call Google employees, sell data on behaviour of /. posters. -> buy some fuel
    Step 2b: proceed to destination.
    Step 3: ???
    Step 4: Profit!

What this country needs is a dime that will buy a good five-cent bagel.

Working...