Return to Video

04ps-10 Stochastic Motion Solution

  • 0:00 - 0:02
    Here is my solution.
  • 0:02 - 0:05
    As I go through all different actions a, as before,
  • 0:05 - 0:10
    I now create a new inner loop of going through different action outcomes.
  • 0:10 - 0:14
    This lists is (-1, 0, 1),
  • 0:14 - 0:17
    and I set the actual outcome to the adjacent action in the action list.
  • 0:17 - 0:21
    You might remember the action list is a list of different outcomes.
  • 0:21 - 0:27
    By incrementing it by 1 or decrementing it by 1, I can pick a slightly different action in that list.
  • 0:27 - 0:30
    Of course, I have to do the modulo 4 on the right side.
  • 0:30 - 0:36
    Then the limitation is similar to before. I project the outcome into new coordinates--x2 and y2.
  • 0:36 - 0:39
    Now I need to assign the probability with this outcome
  • 0:39 - 0:42
    where if they modify a 0, we take the success probability.
  • 0:42 - 0:49
    If it's not 0, we take 1 minus that divided by 2, because there are 2 possible undesired outcomes.
  • 0:49 - 0:52
    Then the test proceeds by checking whether this is a legal grid cell,
  • 0:52 - 0:55
    it's inside the grid, and the grid value is 0.
  • 0:55 - 0:59
    Then like before, I add the value of the grid cell
  • 0:59 - 1:03
    by now multiplying by the probability of that specific action outcome.
  • 1:03 - 1:06
    Otherwise, I do the same for the collision cost.
  • 1:06 - 1:12
    Finally, I take my cumulative value of v2, which I initialized with the cost of motion.
  • 1:12 - 1:14
    You can't see this right here, but it's filled up.
  • 1:14 - 1:17
    I update my value function just like before.
  • 1:17 - 1:19
    You can see the quote over here.
  • 1:19 - 1:21
    This is what you should have programmed.
  • 1:21 - 1:26
    The key difference to our example in class is the inner loop over here
  • 1:26 - 1:29
    where I go over different possible action outcomes,
  • 1:29 - 1:32
    compute the actual action outcome,
  • 1:32 -
    and then do the probabilistic addition of these outcomes rather than just studying one outcome.
タイトル:
04ps-10 Stochastic Motion Solution
概説:

dummy description

more » « less
Team:
Udacity
プロジェクト:
CS373 - Artificial Intelligence
Duration:
01:38
Amara Bot added a translation

English subtitles

改訂