NAM Traffic Simulator Help

[Tarkus]

Thank you for the kind words, b22rian!

At any rate, I've compiled a new Congestion DataView, which is attached here. 

It does have some three digit Uint32 values assigned for the upper data points, which I didn't have the chance to test since I didn't have any networks that were congested enough to see the upper end--so I'm not entirely sure they'll work.

The data points are similar in spacing, though I tended to go for 12.5% increments on the upper end.

-Alex (Tarkus)

[b22rian]

ok alex..

i had a chance to test this out on my big city 831 K sim B hard difficulty..

I think the yellows and oranges are working good..

But no red are showing up on networks im sure are congested enough to be red..

also i assume this one file will cover all 3 of the difficulty levels right ?

Thanks, for all the time you have invested in this today !

Brian

[Tarkus]

Brian, thanks for the test report!

It will work for all three difficulties, since it's designed to report percentages of the overall volume-to-capacity ratio, as opposed to absolute capacity values (like the Volume DataView that z released).

The reds should start kicking in around 250% in theory (around 6250 on a Road using A/B Hard), though because the reddish colors are set over 200%, they have the experimental three-digit hex values, which may or may not work.  If not, it proves that the DataView can't go over 200% capacity, and I'll need to adjust it accordingly.

-Alex

[z]

Alex,

This is certainly a very interesting idea.  And it seems to be producing yellows fairly consistently in areas with less than 100% capacity - something I didn't think was possible.  Congratulations!   

Three-digit hex numbers don't work, though; I had tried them before, as well as with your view, and the game just ignores them.  So it may be a Uint32 field, but the internal array is only 256 numbers big.

I should warn you of some of the experiences I've had, though.  The Congestion view will sometimes produce a wrong color, even a wildly wrong color, and this has nothing to do with the color ramp being used.  For example, in a city that I just tested, I have a road running at 50% capacity.  With the Maxis view, it shows as bright yellow.  With your view, it shows as dark orange.  Both are equally incorrect.  I've seen many other anomalies like this.  I simply mention this so that you can be careful during testing.

One thing you mention though:

So really, the Maxis graph has no differentiation between 0% and 118%, and the congestion coloration covers an 82% range.

This is not what I have seen - I see the Maxis graph green up to 100%, and then slowly start changing to yellow.  Once again, there are anomalies, so you have to look at a lot of data points.  I found my Traffic Volume View invaluable for checking this out - I think it would speed up your testing by quite a bit.

Then there's the question of where you want to go with this.  How far down do you want the colors to go?  Which colors do you want to use below 100%?  May I suggest that you coordinate your colors with my Traffic Volume View as much as possible?  Users always appreciate consistent interfaces.  And the colors used with my Traffic Volume View have proved very popular - I've gotten very positive feedback on my STEX upload.

In any case, what you've done is quite impressive.  I'll be looking forward to hearing more results.  And let me know if there's anything I can do to help.

[b22rian]

Three-digit hex numbers don't work, though; I had tried them before, as well as with your view, and the game just ignores them.  So it may be a Unit32 field, but the internal array is only 256 numbers big.

I should warn you of some of the experiences I've had, though.  The Congestion view will sometimes produce a wrong color, even a wildly wrong color, and this has nothing to do with the color ramp being used.  For example, in a city that I just tested, I have a road running at 50% capacity.  With the Maxis view, it shows as bright yellow.  With your view, it shows as dark orange.  Both are equally incorrect.  I've seen many other anomalies like this.  I simply mention this so that you can be careful during testing.

alex.

I definitely had a few network lines over 250 %, that still show as orange..
But I guess this is the same thing as Z saying with the 3 digit hex numbers not working..

For the most part the oranges and yellows work pretty consistently..but i have seen the same anomalies as Z
is reporting in the above,, also..

Then there's the question of where you want to go with this.  How far down do you want the colors to go?  Which colors do you want to use below 100%?  May I suggest that you coordinate your colors with my Traffic Volume View as much as possible?  Users always appreciate consistent interfaces.  And the colors used with my Traffic Volume View have proved very popular - I've gotten very positive feedback on my STEX upload.

I think this is a great idea.. Z's volume graph is outstanding ! 
But also I know you are busy with RL stuff and obviously you do so many different things with the game !!
when i saw your first posting, i was like wow !, tarkus is getting involved with this !....
We really appreciate your interest and getting involved with this and maybe Z can help you if hes not too
busy..  ?   he is really good at this stuff !! ..

Thanks a lot for the time you have already committed to this with your busy schedule..

Regards, Brian

[z]

On another note...

I was just going through the old "Commute engine tweaking for NWM" thread, which contains many of the original discussions and planning for the new traffic simulators.  I noticed that mott figured out how to get the Commute Time Graph to give accurate readings in real minutes (there's a scaling constant that can be adjusted), and he included a Commute Time Graph exemplar in an early version of his simulator (alpha 3) that implemented this fix.  In a post a little ways before the alpha 3 file, he says:

Oh, don't forget to copy the graph exemplars too.  The real-time thing doesn't work without it

The exemplars he is referring to here are the Commute Time Graph and the Traffic Congestion View.  The "real-time thing" is the proper display of the commute time in real minutes.

However, although the Traffic Congestion View made it into the NAM release, the fixed Commute Time Graph apparently did not.  Not only is the exemplar not in the traffic simulator files, but the readings on this graph show it is not adjusted properly.  And mott's numbers seem to be right; when I used them to generate a Commute Time Graph for my traffic simulator, the result was commute times that looked quite correct.  So what happened to the fixed Commute Time Graph?   

[remanh]

I can't get one way roads to cross borders.  I don't know if this is a traffic simulator problem or not.  Can someone help me?

[jplumbley]

Hello remanh,

One Way Roads (OWR) have never been able to crass borders.  This is due to a hardcoded function that MAXIS never included in the OWR Network.  The issue that OWR border crossings is that it would make MAXIS fix alot of regional things for pathfinding.  When, MAXIS created the game and a Sim travelled to a neighboring city tile that Sim was "assumed" to return via the same tile it left in (or in the case of two-tile networks the one directly to the left of that tile).  This was done most prob ably so that they neer had to keep track of regional pathfinding beyond the current city tile you are working in.

Sim City 4 claims regional play was one of the biggest advantages to the game, but in reality each city is exactly like one of those old Sim City games where you had one city tile and a bunch of dormant computer tiles next to it.  Only difference now is you are the one controlling the growth in those neighbor cities, even though they themselves are a separate entity.

[Mr_Maison]

Just out of curiosity, does fact that street capacities make the rural areas more realistic, lessening the traffic in those areas? I always had a problem with traffic jams where there are only like 5 mansions. In other words, will I see less cars in rural areas?

[GoaSkin]

Hello,

I am currently programming a new NAM tool (based on SC4Config) to make the configuration by shuffling plugin files obsolete. That means the program produces the configuration exemplars and saves them into plugins.

Not like in SC4Config, the user can only choose presets in the NAM tools without customizing anything (except the pathfinding heuristic if wanted and train lengths because the optimal value varies depending on the system performance).

If you want to have new simulator plugins to be included/emulated, please send me them early!

[jplumbley]

Hello,

I am currently programming a new NAM tool (based on SC4Config) to make the configuration by shuffling plugin files obsolete. That means the program produces the configuration exemplars and saves them into plugins.

Not like in SC4Config, the user can only choose presets in the NAM tools without customizing anything (except the pathfinding heuristic if wanted and train lengths because the optimal value varies depending on the system performance).

If you want to have new simulator plugins to be included/emulated, please send me them early!

You can download the April NAM for the latest ones.  April removed some of the older ones due to information learned last year at this time.  All the ones made by myself and Mott are included in the April NAM.