rrlangly

Member
  • Content count

    21
  • Joined

  • Last visited

Community Reputation

111 Neutral

About rrlangly

  • Rank
    Member
  1. I'm familiar with A* and have worked it on a 2D problem (cars). I am now looking to learn about shortest path problems for a 3D space (working with flying a drone) and was hoping someone could tell me what algorithms (simplest) to look at.   I'm wanting to take off, fly to an ordered number of way points, and return.   thanks,
  2. math required for Adaline neural network

    And what areas/courses would this mostly be taught under so I can find the books to study in-depth?
  3. math required for Adaline neural network

    I'd have to say no.
  4. math required for Adaline neural network

    I understand this conceptually from reading high-level descriptions, but I think I need to get to the point where I can work out a problem and see it through. So the answer is no.
  5. Hi,   I've implemented an adaline neural network by looking at example code, but at times it seems buggy and I really need to understand it better to know exactly what my problem is.   Eventually, I'd like to implement other types of networks, but just to start off with and for adalines only ...   What is the minimim math I need to understand this. My current plan is to buy some math books just to cover what I need and dig in.   Any suggestions/recommendations much appreciated.
  6. trying to plot linear equation

    You are correct, it is in a loop and generates random points from -1 to 1, and then this linear function is used to "categorize" the data. To answer the "y1 ignore", I should have put this after the line "y1 = " (which I think tells me what side of the line the point is on). [code] if (y < y1) output = 1; else output = -1; [/code] I substitued x = -1 and x = 1 and got these results, but I don't think I'm doing this correctly. [code] y1 = (-5 * x - 2) / -4 y1 = (-5 * -1 - 2) / 4 y1 = (5 -2) / -4 y1 = 3/-4 y1 = (-5 * x - 2) / -4 y1 = (-5 * 1 -2) / -4 y1 = (-5 -2) / -4 y1 = -7/-4 [/code] I'm trying to learn simple neural nets. I'm just not good at math but I can see I have to start somewheres.
  7. I believe this code is generating random data points and plotting based on a linear equation, but I'm not sure how to solve this equation manually to check how points are lining up. Can anyone point me to something so I can understand how to plot points from this equation? [code] x = ((double)rand()/(double)RAND_MAX) * 2.0 - 1.0; y = ((double)rand()/(double)RAND_MAX) * 2.0 - 1.0; y1 = (-5 * x - 2) / -4; [/code]
  8. I'm not very good at math, but heres what I'd like to do ... I have an object located on a 2d grid, so I know its current point. I get direction by moving the point a short distance in the forward direction so I know where its pointing. My target location is a secondary point. My goal is to find direction to the next point. So basically, I just want to turn my object to the correct angle and move it to a specified location. Can anyone help w/ how this is done?
  9. help calculating G cost of A* algorithm

    I see my path taking shape, but trying to debug areas where the parent is wrong. If a better path was found because it has a lower G cost, and then after G, H, and F scores along w/ its parent has been updated ... do I then drop it from the open list to closed list? Or does a better path found mean that better path node stays on the open list? So far the only time I see that I ever drop a node from the open to closed list is when a node is checked for the low F score. And up until my path is screwed up, I can see when I backtrace for parent nodes, they're all on the closed list. So my path looks good from the start to point of screw-up and all those nodes are on the closed list. At the point of screwup, I see nodes on the open list.
  10. help calculating G cost of A* algorithm

    Ah, ok. So I think I'm starting to get it. When I had no obstacles in my path, The path was always the lowest F score and that's what I would store (and print) for my path. And it worked fine. Then once I started adding obstacles, I would see huge blotchy type areas and I wasn't using parents as I should, rather, I was still thinking like a path w/ no obstacles. I haven't fixed this yet, but at least now I think I understand what should be happening. Thanks for your help.
  11. help calculating G cost of A* algorithm

    I guess that's good. Good that you think it should work like that, bad that I don't understand whats happening. Although now it just keeps running, it never finds a path. I had it working before, even though at some points the path solution before a wall looked like a christmas tree. But I'm thinking now that that was my fault in the way I was printing the nodes. So at the point where I explained that it goes from node 1675 and seems to reset at 1174 (about ~10 rows back up) ... how can it continue to find a path if several of the nodes that it tried to find a solution with the first time are no longer on the open list. They're on the closed list now and unavailable to be traversed again. I didn't read anything about resetting any lists in the tutorial.
  12. help calculating G cost of A* algorithm

    This is confusing me when my node runs into invalid terrain. Lowest F score is 1675, so I drop it off of the open list and add to closed list. [code] 1574 1575 1576 1624 1625 1626 1674 1675 1676 1724 1725 1726 (this row is a wall) [/code] Path is travelling downward (1575, 1625, 1675). So now I search for a better path of nodes on open list by lowest G cost. G cost for each of these nodes is: [code] 382 378 382 392 388 392 402 398 402 X X X [/code] Current node is 1675. 398 + 10 = 408 for horiz/vert nodes, no better path found. 398 + 14 = 412 for diagonal nodes, no better path found. Now I search all adjacent nodes to 1675 that are not already on the open list. This adds nodes 1724, 1725, and 1726 (the illegal terrain nodes). Because they're illegal, they DO NOT go on the open list. So now I start a new iteration and its time to select the lowet F score on the open list and switch it to the closed list. The problem is, my lowest F score on the open list is now like 10 rows up. And from here on out my nice clean path that I could see it traversing has just transported itself across my grid and things just get whacky. Any ideas?
  13. help calculating G cost of A* algorithm

    Ugh, my code works, it finds the target, but I know it's not right. I think I'll try to find another tutorial to see if it explains certain things differently that might clear things up for me.
  14. help calculating G cost of A* algorithm

    Here's a question I have ... Why doesn't the open list just contain the current working nodes (current node + adjacent nodes to current node) instead of all the nodes that all current nodes ever run into. There can never be a better path than an adjacent node near a current working node so why keep the history of all inspected nodes on the open list since they can't be a better path. Guess that means the open list would only ever contain about 8-9 nodes at most. Or am I leaving something out?
  15. help calculating G cost of A* algorithm

    [quote name='alvaro' timestamp='1331611165' post='4921563'] Yes, that's called the Manhattan distance. That explains why you are exploring diagonal movements first. Anyway, is it clear now how you handle updates to G? [/quote] I won't say its clear, but you've given me much to think about. Let me chew on this a bit. Thanks.