• Welcome to SC4 Devotion Forum Archives.

NAM Traffic Simulator Development and Theory

Started by z, August 02, 2008, 05:07:50 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

z

#100
Quote from: sumwonyuno on December 21, 2008, 08:06:16 PM
To z,

I'm still testing out the simulator.  I've been having some and quirky behavior (patches of chronic abandonment and a few instances of commuters flocking to a city tile with no jobs), but I can't discount nor blame the simulator as cause.  Also, there are still instances of mass commuters taking the shortest route to/from a neighbor connection, and a few places where the spreading behavior is occurring where it is not desired.  I can post pics if you are interested.

I'm quite curious about the congestion view and your new traffic view.  Under a non-z-simulator configuration, I didn't have much of a use for the traffic view, and the congestion view alone (usually) told me about trouble spots.  Now, my cities are mostly full of green roadways in congestion view.  But the route query and traffic view tell a whole other story.  In both rush-hours, as expected, major arterials and bottlenecks are red as can be, while side and cross streets show up orange and yellow.

I'm nowhere knowledgeable as you or the NAM team about the traffic simulator.  Is there a sort of ultimate limit on streets relative to the stated 100% capacity of a roadway?  In other words, is there a % limit?  I know of the 65535 showable digit limit for highways, but you could still add vehicles to it and development would still continue.  I'm asking because of these patches of chronic residential abandonment.  It happens across all 3 wealth types in the same area.  This is an issue I've also had with Simulator A and B.  I do not believe it is due to a lack of available appropriate jobs since I have job lots within the same city with no workers.  Also, I can still build in other areas, up to a point.  I think I can attribute it to bottlenecks and a huge influx of neighbor commuters.  I don't believe I have eternal commuters because the region I'm building is quite linear.  I've tried using the higher capacity versions and I've settled on Medium.  Raising the capacities somewhat lessens abandonment, but it increases the chance of commuters flocking to a wrong neighbor connection and spreading where undesired.

Some of what you're describing sounds like it might be caused by unbalanced demand.  So first, I have three questions:  1) What is your regional population?  2) Are you using CAM?  3) What are your tax rates for this city?

Next, it would be very useful if you could post a picture of the output from either the Census Repository Facility or the Census Repository Vault for the city in question.  If you don't have these buildings, you can find them here; if you're just installing this for the first time, the easiest thing to do is use the Vault.  Also, posting a picture of your RCI demand graph would be helpful; although the data is basically contained in the Census Repository, there are certain patterns that I find easier to recognize in the graph form.

The case of mass commuters taking the shortest route to or from a neighbor connection is inherent in the built-in traffic simulator engine, and cannot be fixed.  If it could, then the eternal commuter problem could be fixed as well.

You say there are "a few places where the spreading behavior is occurring where it is not desired."  A picture with an explanation would be very helpful here.

As for the traffic volume view, your results are a bit puzzling here.  They would make sense if you were using Simulator A or B, but I gather that you're describing a situation with Simulator Z.  Are you using the optional Traffic Congestion View that comes with the NAM?  If so, try removing it and see what your results are.  To find it, look in your Network Addon Mod subfolder in your Plugins folder, and see if the file NetworkAddonMod_Congestion_Data_View.dat is there.  If it is, remove it.

If that's not your problem, please post some pictures with queries, and we'll see what we can figure out.  I have not heard of this problem before with Simulator Z.

As for capacities, all networks have a hard limit of 65,535 in terms of a volume that can be displayed during a commute period.  However, as you gathered, you can have higher volumes than that; they just don't show up in the query (or in the Traffic Volume View, which simply reflects the query.)  Other than that limit, there are no other hard limits in terms of absolute numbers or percentages.

I've seen the same type of chronic abandonment that you have in certain areas.  You mentioned that you've noticed it in a number of simulators, which makes sense to me.  Either it's a demand imbalance problem, which the output of the Census Repository will show, or else it's a limitation in the basic traffic simulator engine, upon which all traffic simulators are based.  Based on my own experience, I don't think it's likely to be in the traffic simulator engine.  But again, your pictures will tell us a lot.

One final question:  You said you had the same abandonment problem with Simulators A and B.  Was the problem using these simulators more, less, or about the same as when you use Simulator Z?  The answer to this might help to pin down the problem.

sumwonyuno

#101
To z,

1. My current region population is 790,577.
2. Nope, I'm not using CAM.
3. Tax rates are all 9% except for R$$$ at 12%

I re-ran all of the inhabited city tiles in the region for about 2-3 years each and then resaved them.  I was quite surprized that the traffic was generally performing the way I wanted to, and many of the quirky behavior was gone.

However, there is one persistant quirk.



I don't know if it's some sort of joke the Sims are pulling off.  1600 cars are going into that dead end street and turning around.  I know this isn't a normal because this area is the SimCity equivalent of my neighborhood.  To the right are neighbor connections, jobs are southwest.  There is a closer offramp to jobs, but the roadways there are more congested.

I've removed the NAM congestion views and data views.  I had a THL Data View More Detail dat, but there was no difference when removing it.  But since I resaved the city tiles, congestion view has worked as expected; there are much more areas with yellow, orange and red.

Yes, using Simulator Z has greatly lessened the spread of abandonment and dilapidation.  Compared to now, previously with A and B, there was a smaller population (< 500,000) and more jobs.  I had less congestion.  Abandonment occured in many cities, both mixed-zone job centers and suburbs.  Many commuters just commuted to neighbors.  It almost reached the point of eternal commuters, and I had to implement stop-gaps, but doing that further spread abandonment.  I think there might have been some 'residue' in some of the cities I could have not resaved with Simulator Z.  Abandonment is now limited to 3 city tiles (one job center and two consecutive tiles to the right).

This is from the job center tile in question:






I think a huge fault of mine in regards to demand, and it probably has to do w/ the Residential Halver and Industry Quadrupler plugins.  When I first started the region in January 2008, I did not have them installed.  After starting to build out the job center tiles, I installed them, but later removed the Industry Quadrupler.  I did manage to find a thread at Simtropolis where RippleJet said along the lines of halving an existing residential population would turn them into commuters and would mess up demand.  However the demands are quite high (R$$$ is only low in this city tile due to taxes; it's well above 5000 elsewhere), and it's not like I can't build development.  I've managed to "get rid" of abandonment in all of the other city tiles.  It's just that in the 3 city tiles, certain areas still tend to abandon and/or dilapidate.

Thanks for helping.


The City & County of Honolulu, a Mayor Diary based on Honolulu, Hawai'i.

mark's memory address - I've created a blog!

z

Well, the good news is that there's nothing to indicate any problems with the traffic simulator here.  Also, I don't see anything here that can't be fixed.

Quote from: sumwonyuno on December 22, 2008, 05:58:16 PM
I think a huge fault of mine in regards to demand, and it probably has to do w/ the Residential Halver...

Yes, that will do it.

QuoteI did manage to find a thread at Simtropolis where RippleJet said along the lines of halving an existing residential population would turn them into commuters and would mess up demand.  However the demands are quite high...

But from the end of your previous message:

QuoteRaising the capacities somewhat lessens abandonment, but it increases the chance of commuters flocking to a wrong neighbor connection and spreading where undesired.

This sounds to me strong evidence of what RippleJet was warning about.  The idea of the Residential Halver is good in principle, but a number of people have found it just doesn't work well with the game.  You should definitely get rid of it.  Also, if RippleJet says something, especially if it has to do with demand, please, believe him.

So I think a lot of your problems, including the transitory ones, are due to the Residential Halver.  I would also not recommend reinstalling the Industry Quadrupler.  Instead, I would strongly recommend installing CAM 2.0 when it comes out.  I wouldn't install the current version at this point, simply because it's going to be completely incompatible with CAM 2.0.

The THL_DataViewMod_MoreDetail.dat is a very nice mod, and I've never had any problems with it.  You should feel free to reinstall it without worrying about any problems.

It sounds like removing the NAM Congestion View fixed your congestion problem.  This file is known to be broken, and is scheduled to be fixed for the next NAM release.  I think that what you're calling the NAM data view is actually my Traffic Volume View; if so, you can safely replace it.

I gather from your message that when you installed Simulator Z, your population increased almost 60%.  This is consistent with what other people have seen with this simulator; it's a result of the increased pathfinder optimizations, which is also what's responsible for lessening the spread of abandonment and dilapidation in your city.  But if your transportation network is unchanged, it's not surprising that this would cause additional congestion.  You may need to do some upgrading here; adding some rapid transit would help a lot.

As for that street in the first picture, the cars aren't actually turning around; the built-in pathfinder prohibits Sims from ever retracing their routes.  The animated cars may be turning around there, but their actions rarely tell you what's really going on.  Still, the volume of 3185 at the end of the street is surprisingly high; there's probably something illegal going on there. ;D  Seriously, if you want to know where all the traffic is coming from, you'll need to do a route query on the buildings around there.  But again, this is not a traffic simulator problem; the simulator is simply taking the Sims that live there and routing them out to their jobs.  The incoming traffic looks a little spurious; my guess is that there's some building there that has a few jobs that are being filled.

Swamper77

Quote from: z on December 22, 2008, 11:52:08 PM
As for that street in the first picture, the cars aren't actually turning around; the built-in pathfinder prohibits Sims from ever retracing their routes.  The animated cars may be turning around there, but their actions rarely tell you what's really going on.  Still, the volume of 3185 at the end of the street is surprisingly high; there's probably something illegal going on there. ;D  Seriously, if you want to know where all the traffic is coming from, you'll need to do a route query on the buildings around there.  But again, this is not a traffic simulator problem; the simulator is simply taking the Sims that live there and routing them out to their jobs.  The incoming traffic looks a little spurious; my guess is that there's some building there that has a few jobs that are being filled.

Z,

The traffic going to the dead end street to turn around could also indicate a pathing problem. In their case, it looks like there is a pathing problem with the diagonal One Way Road intersecting the orthogonal street.

There were a few problems like this when Rush Hour came out. In those very few cases, Maxis had omitted paths to allow the vehicles to travel in specific direction. The vehicles would go down a dead end street to get to a side of the intersection that would allow them to travel in their intended direction. I don't have the pictures handy, but I do recall there was an issue specifically with a diagonal road intersection that would not allow vehicles to turn left from the diagonal road. The old Transit Bugfix files fixed these pathing issues. They are currently in the NAM's files somewhere as all the old Transit Bugfixes have been incorporated into the NAM long ago.

-Swamper
You can call me Jan, if you want to.
Pagan and Proud!

z

Swamper, what you say makes a lot of sense, and really answers all the questions about what's going on here.  It would mean that Maxis modified their implementation of the Manhattan A* pathfinding algorithm to allow path retracing in certain cases, but I gather from what you say that this has definitely been observed with dead-end streets before.  And as this particular configuration is a bit unusual, especially with the custom NAM off-ramp, and since most people don't do route queries that often, I can see how this would have been missed until now.

So who gets to fix this?

Tarkus

sumwonyuno, would you be able to show a pic up close of that Diagonal Elevated Highway/OneWayRoad exit ramp there with the DrawPaths cheat (included in Buggi's ExtraCheats.dll file)?  That would help us ascertain what's going on with the actual transit path files and see if they are the cause of your problem.

-Alex

b22rian

Quote from: sumwonyuno on December 22, 2008, 05:58:16 PM
To z,



I re-ran all of the inhabited city tiles in the region for about 2-3 years each and then resaved them.  I was quite surprized that the traffic was generally performing the way I wanted to, and many of the quirky behavior was gone.



I know this sounds kinda amazing.. But ive noticed when doing testing .. most changes do occur in the 2-3 time
frame you mentioned.. but I have seen some changes continue to evolve up to about 5 years of game time or
so.. the pathfinding is quite complex, so I can't say im too surprised about this.. But I would recommend going
about 5 years just to be sure of any sort of change to traffic..conditions..

thanks, Brian

z

#107
Brian is right here.  During my simulator testing, I had to switch between simulators (or different versions of the same simulator) many, many times.  It usually took about five years for the switchover to be complete.  During this time, the routes used by the game would gradually change from the routes used by the old simulator to the routes used by the new one.  The reason that the change takes so long appears to be due to the anti-herding behavior built in to the Maxis traffic simulator.  It will only change a certain number of routes at a time, so as not to flood the most popular routes with Sims, only to have to redistribute them on its next pass.  This gradual change can easily be seen on the Traffic Volume graph; the changes are most apparent using the five-year timeline.  A while after the simulator change, you'll see an initial adjustment of the traffic volumes, which shows up as a square 'U', followed by a stair step pattern in many of the traffic volume lines.  In the cases I looked at, each step was a bit less than a year long.  This pattern typically lasts for about five years in a big city; when the lines flatten out, and there is little or no change for over a year, the transition is complete.

sumwonyuno

To z,

I do realize I have steered the thread away from the simulator by talking about demand (not that I've been warned or anything).  I will post in the appropriate topic if I do have any more demand issues, since we've been able to contain it as another problem.

The good news at least is it isn't your traffic simulator that caused any of the problems.  By the way, the 60% increase in population was because after I switched to simulator z, I built out into suburbs.  Sorry, I'm mixing up cause and effect.  I didn't mean to imply that I had a static-sized region while testing.  But I don't mean to discount that it has helped in making more desirable traffic patterns so I can focus on aesthetics and development, rather than worry if the Sims are jamming the "wrong" roadway.

Yes, the transportation system in many city tiles are essentially "unchanged", since I build infrastructure before unpausing for the first time.  The region I'm creating is very much lane-deficient, has a lack of alternative routes and abundant bottlenecks.  The single avenue that serves my eastern suburbs is already red (9000 vehicles westbound, on medium) and this connects to the only freeway through town (dark orange at 10000 westbound).  Eastbound traffic will be even worse when I finally create them.  I will remove the Residential Halver, and see how the simulator handles twice as much traffic as before.  And congratulations, you've figured out the main plot of a mayor diary I'm writing about this region!

Now, to everyone, here are the pictures:

Draw paths:


Route query; there are only about 30 "local" commuters on this street:


Querying the offramp:


I've built many practice replicas of my neighborhood and this issue does seem to always happen.  Automata is set to low, for performance.

Condemned some houses to extend the offramp:


I've tested both scenarios for 5 years each, and yes I've started to notice the behavior in the traffic volume graph.


The City & County of Honolulu, a Mayor Diary based on Honolulu, Hawai'i.

mark's memory address - I've created a blog!

Tarkus

sumwonyuno, thanks for posting the DrawPaths pics!  It does indeed appear there is something up with the pathing on the OWR Diagonal/Street intersection that's causing your problem.

-Alex (Tarkus)

sumwonyuno

#110
I can't reliably use population or job numbers in many city tiles, in the aftermath of removing the Residential Halver.  There has been a predictable wave of abandonment.  According to the graphs, populations have "doubled".  As for traffic, overall congestion is somewhat worse and more parallel roadways are being heavily used.  Car usage has doubled, and mass transit use has risen several fold.

The following isn't a problem, it's just something interesting.

The biggest wow is certainly pedestrian use.  It's not that it's out of line in relation with mass transit in the graphs.  The phenomenon takes place in the route queries.  It's not that I don't want Sims commuting to the next tile.  In my most dense city tiles, Sims are going through neighbor connections by walking.  Not by the hundreds, not by the thousands, but by tens of thousands (the biggest is almost 40,000 on one avenue).  Many do walk directly from their homes.  That's not all.  It's that many of them were previously on buses, got off at a stop somewhat close to the neighbor connection and walked the rest of the way.  Yes, there are stops right next to the connections.  They aren't transferring off en masse there, though some are.  Also, the Sims aren't "shortcutting" by walking in and out of bus stops along the way.  In virtually all other simulators I've used, Sims have not had this behavior, they would have stayed on MT.  The walking is definitely happens with congested connections.  In city tiles that don't have bus stops, there are only a few dozen walkers to neighbor connections from houses right next to it. 

I think it's definitely is due to features you have described about your simulator.


The City & County of Honolulu, a Mayor Diary based on Honolulu, Hawai'i.

mark's memory address - I've created a blog!

z

#111
Thousands of Sims crossing the border by walking - that's fascinating.  I've never seen or heard of anything like it.  But looking at the simulator's internal numbers, I can see why it's happening.  At maximum congestion, the buses are traveling slower than the Sims can walk.  (This certainly happens in real life sometimes.)  So the Sims get off the bus and walk simply because it's faster.  I think I understand why this happens only near borders.  The border is considered a job destination, and maximum walking speed is far below maximum bus speed.  Only when the Sims get relatively near the border is it clear to the simulator that walking is the fastest way there, and so that's where they get off the bus.  That's what I think is happening.

As for the fallout from removing the Residential Halver, unfortunately, this is to be expected.  But it can be recovered from by creating more jobs where necessary and also by beefing up the transportation infrastructure and possibly using a higher capacity version of the simulator as well.  If you do this, the end result should be a better functioning region than you had when using the Residential Halver - it may just take a while.  And when your congestion drops below the maximum, your Sims should stay on the bus for those border crossings.

b22rian

hi.. Ive been following the thread with a lot of interest lately..
That is quite interesting what sumwonyuno had to say about the pedestrians and buses..

Actually I will share with you one of the things i enjoy most about sim Z, somewhat related to it..
But one of the best changes for me in using sim z is that bus usage is now down to a more realistic levels
in most cities about down to 8: 1 ratio to cars.. Which i think is still prolly higher than most people have
using sim Z..But before the number of buses i had roaming the streets and roads was a bit out of hand..
I think the usage was far too high.. Now some of this may have been because I had gotten a bit carried
away .. using too many bus stops..But without doing anything on my own Sim Z has gotten the bus usage down
to a more sensible number i feel.. I also love how buses now count toward congestion levels as they should..
This is the other change for me i think is great..

Thanks, Brian

sumwonyuno

I had been wondering for a while when you had stated in your beta explanation post about a relationship between transit use and capacity level (using Ultra to simulate low transit use).  It makes complete sense now.  I'll try bumping it up to High capacity; at the moment, I doubt I need to use Ultra.


The City & County of Honolulu, a Mayor Diary based on Honolulu, Hawai'i.

mark's memory address - I've created a blog!

b22rian

Quote from: sumwonyuno on December 25, 2008, 01:29:27 PM
I had been wondering for a while when you had stated in your beta explanation post about a relationship between transit use and capacity level (using Ultra to simulate low transit use).  It makes complete sense now.  I'll try bumping it up to High capacity; at the moment, I doubt I need to use Ultra.

yup. you need to find that setting that is challenging to you, but not so hard you feel its impossible or frustrating..
Keep us posted how your doing.. its nice to hear from you..,

Thanks, Brian

z

sumwonyuno, I've been thinking more about your city, and I think I can explain a little more about what's going on in a way that may be helpful for deciding what to do next.

I think as you and Brian have both gathered, the reason you don't see pedestrian situations like the one you've described in other simulators is due to the fact that buses count as ordinary traffic in Simulator Z.  For any other simulator, the natural thing to do would be to pile Sims into buses - all 40,000 of them - because that doesn't increase congestion at all.

But Simulator Z acts differently, as you know.  What has to be going on for 40,000 pedestrians to be streaming across the border?  First of all, your avenue has to be at maximum congestion.  This means that at the Medium setting you're using, it has to carry at least 250% of its nominal capacity, which is 4000.  (With this sort of border situation, I'm assuming that essentially all the traffic is moving in only one direction for each commute period.)  So 250% of 4000 is 10,000, which means that there are at least 10,000 car and bus passengers traveling on the avenue.  Add to this the 40,000 who got off the bus (since the number walking directly from their houses is negligible), and you get an initial 50,000 Sims trying to use one side of that avenue.  That's 1250% of the avenue's nominal capacity.  No wonder you've got problems!  No real road can hold much more than about 300% of its nominal capacity.  So the simulator looks at this situation, where the Sims are all originally using buses and cars, and tries to figure out what it can do to reduce the congestion.  It can't do anything about the Sims in cars (presumably other routes nearby are also clogged), so they're stuck there.  But the Sims on buses can change to any other form of mass transit.  Unfortunately, there are no rails anywhere nearby.  But as I mentioned before, walking is now faster than riding the bus, and pedestrians don't contribute to congestion.  (Normally this makes sense, as there are few enough of them to stay on the sidewalk.)  So the simulator has the Sims get off the buses and walk, thereby both reducing congestion and speeding up the Sims' trips.

Understanding this, you can see why using a higher capacity version of the simulator and doing nothing else will not solve the problem completely.  For example, with the High version of the simulator, avenues have a capacity of 6000 Sims per side, which means that 250% congestion occurs at 15,000 instead of 10,000.  So switching to the High version of the simulator will probably reduce your border pedestrian traffic only from 40,000 to 35,000.  The Ultra simulator, with its capacity of 12,000 per avenue tile, reaches maximum congestion at 30,000, which means you'll still have about 20,000 Sims walking to the border.  That's why in my earlier message, I recommended "beefing up the transportation infrastructure and possibly using a higher capacity version of the simulator as well."  A single rail line in the right place, even with the Medium simulator, will do wonders for this problem.  If you wanted to be very discreet about this and not change your landscape, you could put a single subway line under your avenue.  In the Medium version simulator, it would have a capacity of 15,000 Sims.  But even at maximum congestion, the subway speed is still faster than that of an uncongested bus line (though much slower than an uncongested subway).  So all your would-be pedestrians would now take the subway, and even though they'd be crammed in like sardines, it would still be the fastest way to go.

Hope this is helpful...

sumwonyuno

#116
I have said it was just something interesting, not a problem.  Once in the neighboring city tile, the Sims do get back onto buses and onto their destinations.

Nevertheless, I think I've found a missing piece of this puzzle that explains it all.

[Salt Lake Boulevard]


Scratch what I said earlier, I've chosen Ultra to have the least congestion as possible, which according to our hypothesis, in turn should encourage car use and limit bus and pedestrian use.  I was shocked to find that pedestrian behavior was roughly the same.  Both congestion and traffic views had changed to barely have anything above yellow (only a short slice of a bus route in the middle of highrises, 50+ tiles away from the connection, was red).  I moved on, since the Sims were indeed transferring to buses in the neighbor city tile.

[Pali Highway]


Next, I went into another tile that previously had no bus service.  It did have a handful of connection walkers.  I added a few bus stops and it stabilized at 3000 Sims walking to the (only) neighbor connection.  I did save the city tile and I went into my Downtown to place bus stops to "catch" those connection walkers.  I looked at the edges of my Downtown to find any more instances. 

[Downtown edge]


There is the King & Beretania Street Couplet.  Beretania is one-way westbound, while King is one-way eastbound.  I was wondering why had the Sims abandoned their buses in order to go to the next city tile.  Wouldn't it be faster for the Sims to keep taking the bus on King Street when it was less congested?  Then it hit me.  Beretania was closer to homes.

The general layout for many of my linearly-laid city tiles have major east-west arteries.  Residential dominates towards the mountains (north) while commercial and industry take up whatever is left toward the shore (south).  So, westbound lanes were closer to homes, including sidewalks.  According to the simulator, walking the westbound half of streets was much more convenient.

Evidence can be seen in the first picture.  I had queried each highrise on Salt Lake Boulevard (the avenue + oneways) and they all chose walking.  Parallel side streets were "fine".  First, if they all had chosen cars, it would have caused additional congestion on Salt Lake Blvd.  Whether they chose cars or buses, they would have to make a U-turn at the next stop light at the west in order to go east, further increasing commute times.  The simulator had deemed walking as the most optimal.

The Pali Hwy picture should have also hinted at this behavior.  The 1145 pedestrians on the northbound lanes heading south to the connection did this because this sidewalk was closer to their homes.  Most of the pedestrians on the other side were sent off walking because if they had stayed on the bus, it would have added congestion to the southbound lanes.

In the last picture, those 17000 pedestrians would have really jammed up King Street if they had been placed on buses.


The City & County of Honolulu, a Mayor Diary based on Honolulu, Hawai'i.

mark's memory address - I've created a blog!

z

I think you're getting the hang of how the simulator works, and I agree, the Ultra version seems most appropriate for your cities, as it is often the best version for cities with little or no rapid transit.  Having some congestion ("barely above yellow") seems just about right for rush hour in cities such as these.

The one mystery that was left here is those 36,454 pedestrians in the first picture.  I think I can explain this.  It has to do with the fact that different parts of the simulator work at different rates.  The first time the simulator is run, it notices the extra capacity, and your congestion map changes significantly to reflect that.  But as I mentioned in an earlier post, actual changes in commuting patterns usually take years to complete, due to the anti-herding effect.  So over a period of years, that pedestrian figure should gradually drop.

Meanwhile, there's just one other point I'd like to clear up:

QuoteI've chosen Ultra to have the least congestion as possible, which according to our hypothesis, in turn should encourage car use and limit bus and pedestrian use.

You're partially correct here, but what upgrading to a higher capacity simulator does is highly dependent on how much congestion there is in a city as well as what transportation types are available.  In your cities, there's no mass transit that's faster than cars, so car usage should already be at its maximum, and moving to the Ultra simulator shouldn't change it.  The same is largely true with buses, with the exception that the reduced congestion will allow thousands of pedestrians to stay on buses instead of walking.  So to summarize, car usage should be about the same, bus usage should increase somewhat, and pedestrian mileage should decrease.

I really appreciate all the detail you've posted about your cities here, as it's quite different from anything I've personally tried, both in layout and in transportation usage.  You've helped to verify that the simulator works properly in a very wide variety of circumstances.  Thanks! :)

sumwonyuno

#118
I think I will settle on High for my region.  Medium and Low have too low capacity.  Ultra just has too much capacity.  With roads at 12000 for 100% usage, commuters have undesirable traffic patterns (e.g. they're all comfortable going to a road's neighbor connection, rather than a parallel highway's because it's shorter and has so much capacity).  There is a lack of congestion :P and bus ridership, which are the opposite of what I intend to simulate.

Now, I do have a question.  I assume RTMT is compatible with this, since you are on the team.  Which capacity level should I be using?

[edited to correct the names of the simulators]


The City & County of Honolulu, a Mayor Diary based on Honolulu, Hawai'i.

mark's memory address - I've created a blog!

z

#119
QuoteUltra just has too much capacity.

Well, I didn't call it Ultra for nothing.  $%Grinno$%  What you say makes sense, though; my intention was that Ultra should only be needed for very high density cities with little mass transport, and so I'm glad that High works for you, since your cities don't appear to be all that dense.  (I think what you are calling Hard is High and what you are calling Easy is Low, although these are the opposite of the actual correlations.)  I noticed your connection between congestion and bus ridership; this makes sense, as in RL, the more congested the roads get, the more likely people are to take public transit.

As for RTMT, you should install it using any capacity; then you should use the RTMT file included with the Simulator Z distribution to replace the equivalent file in the standard RTMT distribution.  This new file is designed for regions that ever anticipate having simulator capacities higher than Low, and has been updated to better support Simulator Z, though it's designed to work with any traffic simulator.  It will be one of the two standard capacity files in the next RTMT release.  Although it's easy to switch simulators back and forth, the RTMT files containing capacities are a different matter (as are all other transit stations).  Although you can change the RTMT capacity file in mid-game, this doesn't change the capacity of existing stations, only new ones.