ST Alpha - "TFox: Remember my excel thing from yesterday? Worked great to a point then I ran into a problem." (Reading a Post)

SlickerTalk
Search Archives
The Leaderboard
The FAQ

Down to Post

Backboards:

Posts: 159

TFox: Remember my excel thing from yesterday? Worked great to a point then I ran into a problem.

Posted by loosilu (aka loosilu2) Dec 9 '10, 07:35

To recap: I am running two consecutive earthquakes on (many tends of thousands of) buildings, road segments, bridges, etc. Each structure has a damage probability curve, which results in one of five incremental damage states (none to complete). I generate random numbers then assign damage states to each structure.

the challenge is, the damage has to be cumulative. I can't undo damage caused by the first.

I did what you suggested and removed all the "complete" damaged stuff after the first round. but that left all the partially damaged stuff. and god damn it, in cases where the second random number was low, the damage undid itself. I need to come up with some kind of logic string to fix this. I haven't described the algorithm to assign damage states, so you probably can't offer a solution based on what I've just said. This is turning out to be more of a rant than anything else.

Responses:

Isn't it essentially: Probablity of X (where X is a state) + Probability of Y (where Y is the probability of the remaining "worse" states? - oblique Dec 9, 07:41 1
- hang on, that might be it. remaining states. by jove I think you got it. I was leaving in all possible states. -- nm - loosilu Dec 9, 07:46
Is the damage of Quake 2 influenced at all by Quake 1? -- nm - Strongbad Dec 9, 07:39 6
- Quake 3 was really the best of the series, by far -- nm - oblique Dec 9, 07:44 1
  - They totally f*cked up with Quake 4, though. -- nm - znufrii Dec 9, 07:52
- that was theoretically built into their high powered probability model. Either way, I dno't have to worry about it. I just have to use - loosilu Dec 9, 07:40 3
  - This is probably too simple, but I'm going to throw it out there anyway - Strongbad Dec 9, 07:43 2
    - that's exactly what I did. :) and uh, didn't work. -- nm - loosilu Dec 9, 07:46 1
      - Ah. :) -- nm - Strongbad Dec 9, 07:57
i thought you were just going to add the damage rates from both events together? -- nm - TFox Dec 9, 07:36 3
- that kind of doesn't work. I need damage states for both, not just a final state. -- nm - loosilu Dec 9, 07:38 2
  - I'm sure you can see it better than I can, but the beauty of a spreadsheet is you can just add a row or column to store and display interim values. - TFox Dec 9, 07:41 1
    - it's what happens in the second step. I run the random number generator. If the probability is low (usually it's zero, actually), - loosilu Dec 9, 07:45

Post a message top

TFox: Remember my excel thing from yesterday? Worked great to a point then I ran into a problem.

Replies are disabled on threads older than 7 days.