Jump to content
  • Advertisement


  • Content Count

  • Joined

  • Last visited

Community Reputation

146 Neutral

About uncutno

  • Rank
  1. uncutno

    Max Priority Queue sixe for A*?

    I use linked lists (not pointers but with indexes), and since all nodes could be in any list, all nodes got a link (to another node or to NULL)... this way, no data is allocated or released, all i need is a List-Head[x,y] and then node[x,y] points to node[i,k] or null... this way, all nodes can be in one list, and no data is realocated anytime (not even in worst case)... makes the List code a little more complicated, since i have to make sure that all lists are correct and in order, but when this List-wrapper is done, its lightning fast!
  2. uncutno

    Making new connections in NNs

    often, all nodes in one layer is connected to all nodes in the next layer... instead of not connect, you connect all from the start, and then you set the weight to 0
  3. ok, you could do it the simple way; whenever you have calculated a path, store it, then when you calculate a new one, you see if you already have it. To decide what path to throw away, use an algorithm like the ones used to decide what pages to swap... this way, paths that was used often (or got parts of it used often) would stay, and more rare paths would be recalculated... now you need to give your paths some sort of keys, like how do you know if path X-Y is part of any of the paths in the cache? This is only usefull if the same paths are used alot, if they not you will only add overhed while looking for paths in the cache... This is sort of a hybrid between pure A* and preprocessed nodemaps, you could for example divide your map into areas, and then chach the shortest ways from one area to another area... later, when a new path from area X to area Y was needed, you could reduce the searchspace dramatically... look up nodemaps for bots to find stuff about this... A path from base A to base B would be needed all the time (good for caching), but a path from tile X to tile Y would only happend once a while (bad for caching)... When i created my last A* pathfinder, i had a pretty small area, so a search didnt take to much time, but still i ran it asyncronized in another thread, and even if the reaction time was 0.5 sec instead of 0.02 sec, it was great (the game seemed to be much more responsive instead of lagging a little every time i calculated a path...)
  4. uncutno

    Max Priority Queue sixe for A*?

    think about this: What if there is no path? Dont you want it to try all possible routs before giving up? If so, you need to store all nodes. 300 * 300 * 10 = 900kb, which is a small amount of memory when you get pathfinding in return...
  5. if you cache paths, there are alot of algorithms from virtual-memory-page-swapping which could be used: FIFO / SecondChance / etc... By storing one path, all sub-paths are there, so it shouldnt be to hard to make 1MB of memory of cached paths improve your overall speed. Then, if everything is hashed :-)
  6. Frames is as waigly defined as any other knowledge representation theory, like semantic nets which masicly is nets with no defined semantics :-) Frames should not be seen as a way to structure your data, realy nobody cares how you implement your frames, the point is that you have a concept of frames, containing a fixed set of fields, or any kind of fields, and you do compaer them with method X, resulting in a form of classification. So: As engenearing tool, one can find out how frames have been done before, and how your implementation should be, but in the end FRAMES is a knowledge-representation theory, and NOT an engeneering blueprint! IMPORTANT!
  7. Hi, i'm an AI researcher, ill give you answers. The main feature (or underlying operation) of frames is to compare them. If your problem is known, you could have a static set of fields in the frame that cant change (if you want dynamic fields it gets more complicated.) I think the easy way would be to use java, and have a MyFrame class with all fields as hadrcoded member variables, and a specialized Compare(MyFrame anotherframe) method. You could use prolog, but since you only want to compare subsets of the model (two frame at the time), you need some complex declarations. If you know prolog, use it. If you dont know prolog, dont care about it. Javas Objects can be used as Frames and should give clean and "to the point" code.
  8. uncutno

    Neural Network Help

    ok... a node have 5 other nodes as input. each of the 5 connections have a weight. you multiply the input nodes values with the connection weights, and then pass this new number trough a "neuron" function; for example sigmoid. The interesting stuff happens when you connect them together into a net. Read: (a suggestion on learning order) #1 (1 hour) Example of AND/OR/NOT/XOR (this will show you that you can represent logical rules with the ann) #2 (1 hour) ANN Backpropagation (this will show you how to train a network) #3 (5 hours) Make a small ANN that learns the logical operators (AND/OR/NOT/XOR) with backpropagation #3 (1 lifetime) Experiment with different problems and different implmentations to build yourselfe a better understanding of what you can use ANNs for. Pros: ANNs can handle noize (successfull experiments exists) ANNs can handle missing data (successfull experiments exists) ANNs can find patterns in what appears to be noizy data (successfull experiments exists) ANNs can transfer its training data to new and unseen problems (successfull experiments exists) ANNs are fast when trained (successfull experiments exists)
  9. Memory mem = new Memory(size); mem.use(); delete mem; 8-P
  10. Look up : subsumption architecture : Brooks R.A. have nice papers about this. Until this day, I have not found a better way to controlle a unit in a complex state space (you dont need it to play chess, but ill guess that your players can mover freely). The architecture makes testing more easy, and alow for a wery robust and layered AI. The cool thing is that each layer is a little more abstract, so that the lower layers takes care of player movement and ball handeling, higher layers can be abstracted away from such details, and can cover the team part. Because the AI is so robust, it can often survive in enviroments that changes to variations not forseen by the developer. I always like to look at how it works in the real world. If we look at the team part of most ball games: -each player often have a possition and a role. -other players support the player with the ball. -players sends messages by shouting (think how cool it is if you connect these to your sound system)
  11. uncutno

    Game Idea: The Brand

    I would not say this is a political game, just because it cares about political issues... (Any War game, is in that way political). I want this to almost be an action game :-) More comments? Why are political games narrow? Are they like educational games? (Often wery booring), or are you actually talking about the genre as where you controlle policies? Currently, events will pop up on the mam, and all events can be handeled by dragging a anti element onto it. The world will be divided into 5 continents, where each continent have a popularity rating, and every event affects this rate. The rate then decides the marketshare, and how much profit you make (after removing production/legal/commercial costs)... So: A Story in a local newspaper, can be removed right away, by digging dirt on the reporter. If you are to slow, and the story makes enough interest, it will lower your popularity, and finaly a court will be anounced. Then you have 3 months to try to bribe the judge, or find bad stuff, and then after a while the jury-game starts, and either you have to pay a big fine, or you go free... The less legal your product/production/commercials are, the more often reporters write stories. Stories in big papers spread to other continents. You can always throw lawyers at trying to change the law of the continent, so that your product get more legal. The game will consist of these 3 main forces: Make your product better/cheeper to produce => Make your product more illegal ,more popular and more addicting Make the laws weeker => Make your product less illegal More illegal product => More stories/reports/demonstrations More stories/reports/demonstrations => Less popularity More/Better commercials => More popularity In your lab, you can test and develop "tools" to affect all the factors showed abow, for example: chemicals, new-laws (that you can lobby), new commercial tricks (this does not need to be fancy, just a list of options, and then you use money to research on them, or to research on new options!) Any other game dynamic you think belong here? Another fun thingie I was thinking about was the pop-stars, and how they loose their credability when you hire them, so that the higher credability the pop-star have, the longer he/she lasts, but the more expencive... Im hoping to develop a small model of different types of people, so that fast cars affects males, pop-stars affects the youth, clowns affects the children => affects the parenths...
  12. uncutno

    Game Idea: The Brand

    Quote:Original post by Timus Or better, Start a game where you're Stalin, ya damn commie. :-P Hehe, im afraid i cant comment on that :-) (Dont want to mess up this thread) Quote:Original post by Timus What depth would it have in gameplay? Ok, this is one of my main concerns. One solution i was thinking of was a set of mini games. Pretty simple, but fun and fast. For example, in the court house you would have a jury, and then a set of emotional words that would affect different members of the jury (you had a fast : remove jury member part first), where "family" would make all the family people warm, and "facts" would make all the smart people warm, and then you had 20 sec to get the majority of the jury warm. The better lawyer you hired, the better they were to get empathy from the jury, and if the oponent lawyer was bad, you would win. :-) Wouldnt it be fun if you won a trial by saying "love","family","understanding","enviromental friendly", :-) and all the jury members got tears in their eyes. Another mini game could be the "dirt digging part", where you could manipulate pictures so where they did something bad. Another fun part I am thinking about is your test lab, where you have all kinds of peoples with 100 of wires connected to their head, wich you could show comercials, or make them taste your soda, and then you would have a readout on how it affected their mode for your brand. Here you will also develop new technologies, like a tech tree climbing thingie. I was thinking about making the whole resource management pretty intense, so that you always have laws to fight, and marketshares growing and shrinking, so that the reactions on what you do would be pretty big. A big fall would not make you loose the game (Think that you have a crisis, and you are loosing millions every second, but then you trow together a new comercial, or give money to charity, and then everything turns :-) ) : The point is that a fast turning world affects how much you earn, but its the stock-holders that fire you / get angry, so they give you some time to make up your loses. Besides, you could always fire all your emploies :-) I was also thinking about that you could make your work conditions horrible, but compencate by sending inn "motivators" holding seminars for them, so that they got happy. Later in the game, you only have "temporal" workers, and all worker unions are dead... The truth is somethines horrible :-)
  13. uncutno

    Game Idea: The Brand

    Quote:Original post by CIJolly Give this McDonalds game a go: http://www.heavygames.com/themcdonaldsgame/showgame.asp ... seems like they removed it (great idea for new gameelement : indy bands / gamedevelopers / moviecreators creating anti-propaganda, that you have to fight with law... / removing the right of speach could be a great sub-goal)
  14. uncutno

    Game Idea: The Brand

    Great response! I realy like that the goal of the game is to take over the world, (you just need to sell some soda first). I also like the ethic score. So questions: -Should this be a high action 10 min game, or should it be a 10 hour strategy game? -What happens when you have taken over the world? (How to make it replayable / Open ended , just like SimCity) -At what level should the micromanagement stop? (I like the idea that you do all the fun stuff, and the rest is done by excecutives (expencive are good, cheap are crappy) ). The simulator itselfe could be based on a set of attributes that different age groups value differently, and that everything you do, adds or removes points from
  15. Ok, i have this idea, and i was hoping for some feedback. The title is "The Brand", and its a simulator, where you as chief over a company want to make better market shares, by branding your product. Now, after reading the book "No Logo", I always wanted to try to be one of the big evil companies that do all these bad things, but still get away with it because they hire pop-stars. Take Nike. They hire low paid workers in poor countries (with tax free sones) to make thier shoes. As soon as their tax free time is up, they drop the contractor, and get a new one. When their contractors do something bad (to be able to beat other contractors on price) they drop the deal, and suddently they are no longer responsible for what they did. At the same time they use 70% of their income for comercials, that is: if you buy shoes for 100$, 70$ goes straigt to comercials trying to make you buy more. So i always wanted to be one of these giants (McDonal,Nike,Coca-Cola,Starbucks), and to be the evil company that loured the people into bying my product. You would make the comercials : throw in a picture of a fast car, and males suddently like your product. You would have lawyers : standing every accuse you are charged over and trying to loby laws that you dont like You would hire private detectives : to digg up dirt about anybody telling bad stuff about your product You would have labs : Trying out different chemicals that made people adicted to your product You would destroy the careers of reporters digging up stories. You would abuse poor countries to lower your expences. This would all be pretty simple and fun, in the same way that it was fun to be the bad guy in dungeon keeper, and the political message would not be that this is ok, but that you should think twice before liking what britney spears like (Im trying to say that playing such a game should be fun, even if you dont support it :-) ) Im picturing a world map, where you can add comercials, hire lawiers and make them do stuff, fight reports, a lab where you create your product... All of these elements could have a sort of tech-ladder, where you get better stuff by paying more. The competition would be the oponent "Normal"-product, the law, and your stock-holders wanting the money. So what do you think? Any ideas? (I wish this section was more about discussing game ideas, then discussing small elements of existing and established genres.)
  • Advertisement

Important Information

By using GameDev.net, you agree to our community Guidelines, Terms of Use, and Privacy Policy.

GameDev.net is your game development community. Create an account for your GameDev Portfolio and participate in the largest developer community in the games industry.

Sign me up!