Inefficient walking distances with large area and many portals

[mrplow]

I know that the script doesn't scale all that well, but when it does complete the linking is pretty inefficient, requiring multiple agents at the same portal at the same time, where it would obviously be faster to have just one agent doing the linking.

I've resorted to editing the links_foragent*_of_7.txt files manually

output.pkl http://pastebin.com/jpm5d240

[jpeterbaker]

I'll take a look. What number of agents gave you this problem? It could make sense to have two agents could be at the same portal simultaneously if they both have spare time. Or maybe there's a bug.

[mrplow]

Thanks, I'm trying it with 7 agents. Its not that because I noticed that one player had to race across town to make a link where another agent already was, then he had to race back to where he started from.

[jpeterbaker]

Sounds like more than a depth-of-search issue. I'll be bug hunting (unless you want the exercise).

[mrplow]

Sorry but I most definitely wouldn't even know where to start, I'm more of a script kiddie than a real programmer

[jpeterbaker]

I meant exercise running across your whole farm repeatedly. I wouldn't ask anyone to dig into my code; it's very messy.

[mrplow]

Ha ha I see, well I already manually modified the walking plan, switching agents around, keeping the linking order. But then again I may have made a mistake, having such a long plan those happen

[mrplow]

Here is an example: Links 25-37: Agent 5 is at the Boys Scouts Monument then has to travel to The Air and The Land only to have to travel right back to Boys Scouts Monument, although agent 2 is at Rick Hansen Man In Motion Totem which is right beside the Boys Scouts Monument, then agent 5 has to travel to Spear Fisherman Totem. It would be more efficient for agent 5 to travel to Spear Fisherman Totem after making link 31 and agent 2 could do link 32.

[jpeterbaker]

I assume the plan that produced the "output.pkl" you posted exhibits this problem?

[mrplow]

yes

On Thu, Jan 23, 2014 at 2:58 PM, Jonathan Baker notifications@github.comwrote:

I assume the plan that produced the "output.pkl" you posted exhibits this problem?

Reply to this email directly or view it on GitHubhttps://github.com/jpeterbaker/maxfield/issues/6#issuecomment-33179097 .

[jpeterbaker]

I couldn't get the current version to produce your routing problem around links 25-37. What version do you have? Do you still have the problem with the current version? I hope not since the routing algorithm is supposed to be deterministic.

[rwehe]

This wasn't my problem. I believe you have the wrong email. I had the other (most recent issue) that you already resolved. Thank you for that! Cheers On Feb 4, 2014 9:19 PM, "Jonathan Baker" notifications@github.com wrote:

I couldn't get the current version to produce your routing problem around links 25-37. What version do you have?

Reply to this email directly or view it on GitHubhttps://github.com/jpeterbaker/maxfield/issues/6#issuecomment-34136459 .

[mrplow]

I'll check when I get to work tomorrow On 2014-02-04 8:19 PM, "Jonathan Baker" notifications@github.com wrote:

I couldn't get the current version to produce your routing problem around links 25-37. What version do you have?

Reply to this email directly or view it on GitHubhttps://github.com/jpeterbaker/maxfield/issues/6#issuecomment-34136459 .

[jpeterbaker]

@ryanieer I don't know why you got an email for that. I posted to the routing issue thread.

[mrplow]

Go ahead and close the issue jon, looks like the new updates have fixed the issue

On Tue, Feb 4, 2014 at 8:58 PM, Jonathan Baker notifications@github.comwrote:

@ryanieer https://github.com/ryanieer I don't know why you got an email for that. I posted to the routing issue thread.

Reply to this email directly or view it on GitHubhttps://github.com/jpeterbaker/maxfield/issues/6#issuecomment-34137850 .

[jpeterbaker]

Awesome. Thanks for the report.