• 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.

b22rian

I wanted to make a few further comments on "realism" with the traffic plug ins and what they mean to me...

In general I agree with "z"'s statement that his traffic plug  does seem to to mirror some of the current larger
american cities  like chicago... where as the reality is they use more cars than mass transit...

If your idea of "realism" is more toward "idealism"  meaning what should we be striving for  in terms of our
increased usage of mass transit and what city planners should be up to  ? In terms of transit goals in larger cities
than i think sim's A and B do this very well.. And also , and  (Jp can speak this this better than I can) The A and B
sim's may actually show a better realistic representation of what goes on in larger european cities, which as I
understand it to be a bit more conscious of using mass transit than their american counter part larger cities..
use their cars and highways more..

In the end though this is still a game.. and you still have to fit your traffic system into whatever traffic mods..
you chose to installl.. So like i said before if its RHW version 21 with 6 and 8 lane super highways  andyou want to
pack with tons of traffic  ? Than sim Z is going to force more traffic on to thoie roads. with all other factors being
equal.. if your goals are for a more elaborate and sophisticated mass transit system in your city , than i think
sims' A and B are going to give you more mass transit usage.. my test results show this to be true..

but isnt  itgreat we have all these wonderful options... ?  ;D

Thank you Z, Mott and J Plumbley

Brian

z

The Alpha 4 release is a major improvement on previous releases of Simulator Z.  The main changes are as follows:

Much improved support for monorails.  After receiving Orion79's bug report about low monorail usage, I discovered that there were many properties in the simulator that affected monorail usage that by all rights shouldn't, and that didn't affect any other form of transport.  By adjusting these properties to various extents, I was able to take monorail usage that was almost nonexistent in my cities and raise it typically into the thousands.  Comparing it against Simulator B, I found Simulator Z's monorail usage to be significantly higher in every case, sometimes by a factor of 100 or more.

Overall improved simulator performance.  The changes needed to fix the monorail bug serendipitously made the whole simulator run better.  (Thanks, Orion!)  Average commute time declined, and many of the data graphs showed a small but noticeable improvement.  Sims had an easier time finding jobs; pathfinding was effectively working better.  Traffic and congestion on all networks decreased.

Support for even more cars.  I tweaked the distribution between cars and mass transit to favor cars a little more, without increasing any road capacities.  The last time I had tried this, problems with traffic and pathfinding started to develop.  But this time, after the improvements described in the previous paragraph, everything went fine.  (Note:  According to many sources, including some quoted by jplumbley, 88% of commuting in the U.S. is done by car, with just 5% overall by mass transit.  In Europe, the car figure is 78%, but is quickly approaching the U.S. figure.  But mass transit lovers, fear not!  There's a bonus for you below.)

Buses now contribute to traffic just like other vehicles.  This is a bit of a tough one, because as jplumbley noted, SC4 buses don't act like real buses at all.  A good analogy would be that every bus stop has a pool of single-passenger motorcycles, and a Sim can get on one and ride it on any route he or she wants as long as the Sim leaves it at another bus stop.  So that's nothing like a real bus, but it certainly is traffic.  And as jplumbley noted, a "busfull" of passengers in SC4 would cause much more traffic than a real life bus.  I tried making making buses contribute to traffic a while ago, and it turned into a mess; the roads got way too clogged.  But once I saw how much better the A4 simulator was working, I gave it another shot.  This time it worked great!  The simulator was much better about routing cars and buses so they didn't get in each other's way, and it wasn't even necessary to raise road capacities, even though bus traffic tends to be greater than car traffic.

As for the fact that SC4's definition of "bus" is unrealistic, the same is certainly true of its trains, which like everything else in SC4, are actually simulated as single-passenger vehicles.  Yet the game's default is to count train traffic.  My change to buses now simply completes the counting of all vehicles as traffic.  But getting this to work without raising network capacities was not easy.

More realistic subway costs.  Subways are ridiculously underpriced in SC4; if they were that cheap in the real world, most cities would have subway networks like Manhattan's.  Unfortunately, it does not appear to be possible to change the cost of a subway tile, as it appears to be hardwired in the executable.  However, the monthly cost per tile can be changed.  So think of it this way:  SC4 subways are paid for by bonds, and interest has to be paid on them every month.  They're interest-only bonds, so they have to be paid forever.  However, demolishing a subway tile obviously removes its monthly cost; perhaps someone can figure out a good cover story for that.

I think that at one point, I figured that to be accurate, subway costs should be increased by a factor of 50.  But I can't just do that in this simulator, as I'd bankrupt a lot of cities quickly.  So I increased the monthly cost by a factor of four.  This is just enough to discourage laying subway lines all over the place, but you can still do that if you have lots of simoleans.  This fits in with the other changes in this simulator; mass transit in general works much better, so lots of subways are not really necessary (though you can still build them if you want).

More accurate numbers for the Commute Time Graph.  This simply refers to the numbers on the side of the graph.  As discussed earlier in this thread, it's impossible to make them accurate for all cities, but the new numbers should be more accurate for most cities.  The shape of the graph, which is the most important thing about it, is unchanged.

All the rail capacities have been lowered slightly.  These were formerly set at 80,000; they are now set at 62,500.  There seemed to be sufficient headroom to make this drop; please let me know how this works out.

A mass transit version of the simulator.  More cars may be more realistic for most cities, but not all (as jplumbley will not let me forget  $%Grinno$%), so I've included a version of the simulator that uses the original distribution of cars and mass transit, which is also in use in Simulators A and B.  Road capacities are about the same as the Easy version of Simulators A and B, although street capacities are equivalent to the Hard version, and highway capacities are higher, for reasons of realism described earlier in this thread.  Other than these differences, the two versions of Simulator Z are identical.  And although the road capacities of the mass transit version are similar to Simulators A and B, the behavior of the simulator as a whole is very different, and is basically the same as the standard version.  Installation instructions are in the ReadMe.

There is now a Cleanitol file for everything.  This was not in the original plans for the alpha version, but someone just kept asking for it...  In any case, please remember that this is an alpha release, and you should know what you're doing if you're going to use it.

Finally, I fixed a problem with the integration of this simulator with existing Clean Air Act mods; details are in the ReadMe.

So that's it for now.  I'm quite pleased as to how this simulator is working out, and I hope to go into beta testing soon - maybe even next release, if this one works out well.  In the mean time, please try this one out and post feedback.  Thanks!

EDIT:  This version of the traffic simulator has been superseded.  You can get the current version here.

RippleJet

Quote from: z on November 07, 2008, 12:24:05 AM
Unfortunately, it does not appear to be possible to change the cost of a subway tile, as it appears to be hardwired in the executable.

The subway laying cost is given by the property Simolean Cost Per Tile in the MiscNetwork exemplar Subway Placement Tuning Parameters (TGI 0x6534284A, 0x084344E0, 0x0A367CA7).

It comes with a value of §150.

z

Ah... I was looking under Plop Cost in Subways under Misc Catalog.  But of course that's just the catalog cost!  No wonder it didn't work.

But I'm thinking that I might want to stick with adjusting the monthly cost, which is right in the Traffic Simulator exemplar, as otherwise I'd have to copy the whole Subway Placement Tuning Parameters exemplar, and there seems to be plenty in there that could conflict with various mods.  Any opinions on this?

0rion79

#84
Object - Monorail usage with Z's Alpha 4th

Z, I have tested the same city, Goldshares, from which I've posted the images and I've seen a drastical increasing of monorail usage. At last, I have been forced to use stations with a capacity greater than 5.000 and even in the normal lines, I've seen an increment from 6% (the previous value) even to 98% (2000 passengers).
Still, I have also been able to see that Sims like more the subway.

If mathematic is not an opinion, we know that subway trains goes up to 150 km/h in your mod, while monorail up to 250 km/h, so it is MUCH faster than subway and still Sims like subway more.

Even if the improvement is great, I think that there is still something "wrong", even if I want to test monorail in another city of mine with a different shape, due to a possible bias.
Since being a psychologist makes me a scientis too, I want to make some experiments to verify my opinions :)

The reason for which I believe that Goldshares may lead to a bias is that, even if the monorail goes parallel to subway and so the path length should be more or less the same, shops and offices are only in the city center, while on the outside of the city there are only rewards and civic facilities as chools and hospitals, that simply employ much less people than commercial offices and services.
For this reason, the Sims in my city have much less reasons to use monorail than subway, while in my next city, that is a "fixing" of a previous one that I've called Metropolis Alpha, I've still used the shape of the ring but in the inner part there are residents, while in the outer part of the ring there are high-density offices, so now monorail should really have a meaning for being used.

If something else comes up, please tell me what you think. In the meantime, I want to make another experiment, but I will tell you later. In the mean time, please give me your opinion about my hypothesis. C u!

---------
EDIT
--------

I have made my experiment.
I have modified another city of mine that was merely used to test my modified train stations where I modified the transit switch points to prevent pedestrians from using them as shortcuts.
This time, I did as following: I've built no road from residential to industrial area, forcing all sims to use mass transit and filling the residential and industrial areas with busses, to allow Sims to reach their working places without walking too much.
Note that the residential areas are zoned symmetricaly so that it has been possible to build a rail line, an underground line and a monorail line each adjacent each other, in the middle of the city, preventing bias from residents' position or different zone density.
Then, I've tested the effects with and without a parking garage nearby to monorail station.

So, these are the results:
         with PG      Without PG
Train:     2               163
Subway: 15              34
Monorail: 252            179

As you can see, the monorail is now recognized as the fastest way to reach the working place and wealthy Sims that want to use the car will choose the train, that is a bit slower than the monorail, only because very often train stations have an included parking garage.
I also have tried to built first a street and then a road and then an avenue that links the residential area to the industrial one and there has been no significant change, except that some Sims (60) has begun to drive directly to work when a road has been built, but I'm using Z's mod (the one that DOES NOT end with MT).

Instead, there has been a radical change when I have upgraded the avenue with the highway: then everybody has begun to drive and/or take buss to work, using the avenue and ignoring the other mass transit systems.

Z, this is a bit strange, don't you think? Monorail should still be faster than highway but maybe the last trait, when Sims are forced to drop off the monorail and use the bus to reach their true destination, has a strong influence on the commute strategy, even because I've placed bus stop directly in front of each terminal, to prevent biases due to too long walking trips.

I think that, *if* you have forced busses to move at 15 km/h in streets (maybe I remember wrong), this is having a too heavy impact on busses usage. IMHO, a speed from 25 to 35 km/h would fit much better and still be realistic.

As other news, I'm having congestion problems with avenue roundabouts. Is it normal? Is there any way to increase their capacity?
I have had to drastically increase Morifari's bus blockers and truck blockers too, that I use a lot in my cities to regulate traffic.

MassHelper

#85
Z, you put the cleanitol file for this one (Yay!)...

With all due respect, Z, thank you

:) Mass

Note: What if we have the Census Repository Building #3? Because the file depends on the Simulator A of the NAM..... (I hope I don't have to destroy it..) One last thing, would it be possible if you can post the new capacities for the transit paths....?
SC4 Modders' Assistant and Adviser

RippleJet

#86
Quote from: MassHelper on November 07, 2008, 06:49:01 AM
Note: What if we have the Census Repository Building #3? Because the file depends on the Simulator A of the NAM..... (I hope I don't have to destroy it..)

No, we can't have you destroy it... $%Grinno$%

Enclosed is a zip file with two new versions of the Census Repository Facility, Version 3.
You need to manually replace the one in your plugins folder with the appropriate one.
And you will probably have to replop it, for the transit capacity to upgrade.



Name of the fileCompatible with Traffic Plugins
 
RW_3x1_Census Repository Facility_6000.SC4LotNetworkAddonMod_Traffic_Plugin_Z_MT.dat
 
RW_3x1_Census Repository Facility_12000.SC4Lot    NetworkAddonMod_Traffic_Plugin_Z.dat

0rion79

Too complicated. I have built it in almost all of my cities and don't want to change it.
Also, I've seen that now ti is bigger than 1x1, and instead I like it that small.... :p

MassHelper

#88
Thanks Tage.... ;)

:) Mass
SC4 Modders' Assistant and Adviser

MassHelper

#89
Quote from: z on November 05, 2008, 05:29:28 PM
For bus and subway stations, I would recommend using RTMT with the CAM capacity settings.  They have many advantages over the in-game stations, as well as having much higher capacity.

What he said, and has said before, is "What was required in that scenario was mixing Residential and Commercial towers to promote people walking to work..."  This is fine if this is what you want to do, and yes, it will allow Simulators A and B to support cities with very high populations.  But most large cities (e.g., Chicago) are not built this way; a common model is to have a large downtown area surrounded by residential areas, many of which are a large distance away and have long commutes.  It was for these common situations that Simulator Z was built.

BTW, Z, do u mean promote walking (and this is not good if it's mixed with NAM Simulators, is it?) as in the option that can be installed along with CAM?
SC4 Modders' Assistant and Adviser

z

Quote from: MassHelper on November 07, 2008, 12:48:49 PM
BTW, Z, do u mean promote walking (and this is not good if it's mixed with NAM Simulators, is it?) as in the option that can be installed along with CAM?

"Promote walking" is part of one of the CAM simulators, which should not be installed along with Simulator Z or any of the NAM simulators.  I'll have to add them to the Cleanitol list.  Meanwhile, since you've been able to run Simulators A and B, this would mean you don't have any of the CAM simulators installed, as they override the ones in the main NAM folder due to their location.

MassHelper

Of course.... but you previously mentioned about adding promote walking in order to use NAM simulator A right.....

:) Mass
SC4 Modders' Assistant and Adviser

z

#92
Quote from: 0rion79 on November 07, 2008, 04:21:35 AM
Object - Monorail usage with Z's Alpha 4th

I also have tried to built first a street and then a road and then an avenue that links the residential area to the industrial one and there has been no significant change, except that some Sims (60) has begun to drive directly to work when a road has been built, but I'm using Z's mod (the one that DOES NOT end with MT).

Instead, there has been a radical change when I have upgraded the avenue with the highway: then everybody has begun to drive and/or take buss to work, using the avenue and ignoring the other mass transit systems.

Z, this is a bit strange, don't you think? Monorail should still be faster than highway but maybe the last trait, when Sims are forced to drop off the monorail and use the bus to reach their true destination, has a strong influence on the commute strategy, even because I've placed bus stop directly in front of each terminal, to prevent biases due to too long walking trips.

Do you still have a parking garage at the origin of the monorail?

In general, this behavior isn't too surprising, as in this simulator, as in real life, there is a heavy bias toward driving.  The avenue was so much slower than the monorail that the Sims gladly took the monorail instead.  But when you upgraded to a highway, the speed difference was much less, and the Sims couldn't resist the pull of the open road.  Also, transferring to and from the monorail takes time, and then there's the bus at at least one end.  The next version of the simulator makes the Sims more sensitive to time differences in paths, so this behavior should be less dramatic.  Also, I think that if you made your transportation corridor longer, at some point you would start to see Sims deserting the highway and moving back to the monorail.  Finally, if you want, I could tell a little bit more if you posted some pictures of your setup.

EDIT:  I just went back and did the math here, and it looks like the Sims are probably taking the shortest route after all.  In Simulator Z, a monorail takes about one minute to cross a large tile; a car on a highway takes a little over two minutes.  So that's basically a one-minute difference.  But if your travel corridor is, say, a quarter of a large tile (64 squares), then the difference comes down to 15 seconds.  Now even if your Sims park at a garage right next to the monorail, they're typically going to have to walk a couple of squares or so to get into the monorail station, and that takes up a number of seconds right there.  Going from the monorail station to the bus station at the other end takes some more seconds, even if the bus stop is right at the door.  And then at the conclusion of their trip, the Sims typically have to walk a number of squares from the bus stop to their job; many more seconds are used there.  So soon that 15 second advantage becomes a deficit, and the Sims are better off taking the highway.

As for the buses on the highway, some Sims don't own cars, or don't like to use them, so they're going to take rapid transit.  Taking the bus on the highway saves precious seconds transfering to and from the monorail station.  Whether it's actually faster, it's hard to say; throughout this whole process, I've discovered that monorails are just plain weird.  I know how to double their usage in cities by changing a single property in the simulator, but that makes the simulator not work properly with RHW-8's, so I can't do it.  Meanwhile, subways aren't much faster than highways, so when you take access and transfer time into account, they're slower.

But notice what happens when you had an avenue present before the highway.  The 15 second difference (in this example) becomes closer to 45 seconds, which gives the Sims plenty of time to do their transfers, and so they take the monorail in this situation.

Quote
I think that, *if* you have forced busses to move at 15 km/h in streets (maybe I remember wrong), this is having a too heavy impact on busses usage. IMHO, a speed from 25 to 35 km/h would fit much better and still be realistic.

Yes, they are at 15 kph (which is three times their speed in Simulator B); this is to discourage buses from using streets in residential neighborhoods as main routes.  Generally, there should be enough roads in a neighborhood that streets don't need to be used much by buses.  Again, pictures of your setup would help.  I'd also be interested in hearing from other people about what they think about the issue of buses and streets.

EDIT:  Given what I said in my previous edit, I don't think this is your problem.

Quote
As other news, I'm having congestion problems with avenue roundabouts. Is it normal? Is there any way to increase their capacity?
I have had to drastically increase Morifari's bus blockers and truck blockers too, that I use a lot in my cities to regulate traffic.

I don't use roundabouts a lot, but I haven't seen unusual congestion around the ones I do use.  Avenue roundabouts have the same capacity as avenues, and this can't be increased withouth increasing the capacity of all types of roads.  Once again, a picture of the area would be very helpful, especially a congestion view.  Or if you could tell me how many and what kind of roads are going into the roundabout, and what their traffic volume is, that would help some.  A combination of these two would be the most helpful.

The next version of the simulator (which I'm currently working on) is really good at clearing up congestion by rerouting Sims, so that may help you here.  Even the version you have should do that much better than earlier versions; try letting the game run for a while and see if that helps.

EDIT:  The next version of the simulator also specifically has an adjustment for avenues to lower their congestion; in my testing, this is working well.

b22rian

Quote from: z on November 11, 2008, 10:09:56 PM
Do you still have a parking garage at the origin of the monorail?

In general, this behavior isn't too surprising, as in this simulator, as in real life, there is a heavy bias toward driving.  The avenue was so much slower than the monorail that the Sims gladly took the monorail instead.  But when you upgraded to a highway, the speed difference was much less, and the Sims couldn't resist the pull of the open road.  Also, transferring to and from the monorail takes time, and then there's the bus at at least one end.  The next version of the simulator makes the Sims more sensitive to time differences in paths, so this behavior should be less dramatic.  Also, I think that if you made your transportation corridor longer, at some point you would start to see Sims deserting the highway and moving back to the monorail.  Finally, if you want, I could tell a little bit more if you posted some pictures of your setup.



    Actually there were a couple things that came up in this posting.. that to me remind me why i love
this game so much because how it closely mirrors real life in terms of traffic situations...

First of all and "z" alluded to this in and out of the posting, but I will try to be more specific...
But when sims first move in they basically move in as 1 of 3 catagories ...
i think these groups are selected randomly by the game and their isnt a whole lot you can do about it..
 
    1) sims that prefer their cars..
    2) sims that prefer mass transit
    3) sims that will take the shortest and fastest ? routes... using either cars or MT or a combination..

we all know there will always be people in real life where you could put mass transit right under their noses and
they still will drive their cars.. Also some people will use mass transit always because they may not have cars,
their cars are being repaired maybe, or gas prices are too high for them to drive..
Anyways, to me this all mirrors real lfe quite closely and its always been something ive loved being in the game..

   The other thing i saw come up which reminded me about real life, because we have a few of these here are
the "round abouts" in the game.. In real life its my understanding that those are always a pain in the rear end to drivers .. and always quite congested... Course in the game they look nice and are cool to build.. but yes to me
they actually mirror real life round abouts quite well...

anyways, i do obviously understand some people's frustrations about these things in the game , and as so many
times it just comes down to how you see the game really in general...

Thanks, Brian

z

#94
The Beta 1 version of Simulator Z is now completed, has undergone extensive testing, and is ready for general release.  A copy of this release is attached at the bottom of this post.  In the Alpha 4 release, I mentioned that there had been major improvements to the simulator since the Alpha 3 release.  The Beta 1 release contains equally major improvements compared to the Alpha 4 release.  At this point, the feature set is complete, and no major changes are anticipated between now and the final release.  Furthermore, the quality and reliability of this release is quite high; in all of the alpha releases, only two problems were found, one of which was an incompatibility with an unreleased NAM feature (which has since been fixed).  As a result, I expect few, if any, problems with this release, and it may very well turn out to be the final release.

Whereas when I first started working on this simulator, I saw it fitting only into a particular niche, the collective improvements I referred to above have made it applicable to all cities, towns, and rural areas.  I am now comfortable with calling it a next generation traffic simulator, and the title of this thread has been changed to reflect this.  Highlights of some of my most important tests are shown below.  The improvements to this release include the following:


  • An even better pathfinder.  I have discovered properties in the simulator that unexpectedly contribute to better pathfinding, and I have tuned them to get results that are as close to optimal as possible.  Use of these properties allows the adaptive heuristics approach that I refer to below.  It also greatly decreases congestion in all situations.
  • Four capacity levels of the simulator.  The better pathfinding allowed me to drop capacities well below what I was using before, and still get equivalent or even better results.  The four versions of the simulator are named Low, Medium, High, and Ultra, reflecting their network capacities.  The road capacities of the Low, Medium, and High versions of the simulator correspond very closely to the road capacities of the Hard, Medium, and Easy versions of Simulators A and B.  However, Simulator Z's versions can each be used in a wider variety of circumstances than those of Simulator A or B, as is shown below.  Here are the network capacities of Simulator Z's versions, followed by those of Simulators A and B for comparison.

    [tabular type=2 caption="Simulator Z"]
    [row] [head]Network[/head] [head]Low[/head] [head]Medium[/head] [head]High[/head] [head]Ultra[/head] [/row]
    [row] [data]Road[/data] [data]2400[/data] [data]4000[/data] [data]6000[/data] [data]12,000[/data] [/row]
    [row] [data]Rail[/data] [data]10,000[/data] [data]16,000[/data] [data]30,000[/data] [data]65,200[/data] [/row]
    [row] [data]Highway[/data] [data]9000[/data] [data]15,000[/data] [data]22,500[/data] [data]45,000[/data] [/row]
    [row] [data]Street[/data] [data]1000[/data] [data]1200[/data] [data]1600[/data] [data]2000[/data] [/row]
    [row] [data]Avenue[/data] [data]2400[/data] [data]4000[/data] [data]6000[/data] [data]12,000[/data] [/row]
    [row] [data]Subway[/data] [data]10,000[/data] [data]16,000[/data] [data]30,000[/data] [data]65,200[/data] [/row]
    [row] [data]Elevated Rail[/data] [data]10,000[/data] [data]16,000[/data] [data]30,000[/data] [data]65,200[/data] [/row]
    [row] [data]Monorail[/data] [data]10,000[/data] [data]16,000[/data] [data]30,000[/data] [data]65,200[/data] [/row]
    [row] [data]One-way Road[/data] [data]2400[/data] [data]4000[/data] [data]6000[/data] [data]12,000[/data] [/row]
    [row] [data]RHW[/data] [data]6000[/data] [data]10,000[/data] [data]15,000[/data] [data]30,000[/data] [/row]
    [row] [data]Ground Highway[/data] [data]9000[/data] [data]15,000[/data] [data]22,500[/data] [data]45,000[/data] [/row]
    [/tabular]

    [tabular type=2 caption="Simulators A and B"]
    [row] [head]Network[/head] [head]Hard[/head] [head]Medium[/head] [head]Easy[/head] [/row]
    [row] [data]Road[/data] [data]2500[/data] [data]3750[/data] [data]5600[/data] [/row]
    [row] [data]Rail[/data] [data]6000[/data] [data]9000[/data] [data]13,500[/data] [/row]
    [row] [data]Highway[/data] [data]4800[/data] [data]7200[/data] [data]10,800[/data] [/row]
    [row] [data]Street[/data] [data]1000[/data] [data]1500[/data] [data]2250[/data] [/row]
    [row] [data]Avenue[/data] [data]2500[/data] [data]3750[/data] [data]5600[/data] [/row]
    [row] [data]Subway[/data] [data]6000[/data] [data]9000[/data] [data]13,500[/data] [/row]
    [row] [data]Elevated Rail[/data] [data]6000[/data] [data]9000[/data] [data]13,500[/data] [/row]
    [row] [data]Monorail[/data] [data]6000[/data] [data]9000[/data] [data]13,500[/data] [/row]
    [row] [data]One-way Road[/data] [data]2500[/data] [data]3750[/data] [data]5600[/data] [/row]
    [row] [data]RHW[/data] [data]4800[/data] [data]7200[/data] [data]10,800[/data] [/row]
    [row] [data]Ground Highway[/data] [data]4800[/data] [data]7200[/data] [data]10,800[/data] [/row]
    [/tabular]

    The question naturally arises, How did I come up with my numbers?  There are three parts to this answer:  Experimentation, research, and compatibility with existing simulators.  As I discuss below, the fact that populations in a typical city in SC4 are many times higher than a similar city in the real world make real world traffic numbers largely irrelevant.  Instead, it's necessary to pick numbers that make the city look real, and this is what I have done.  The street and road figures for Simulators A and B work rather well for their respective difficulty levels, and I chose similar numbers because my experiments confirmed these were quite reasonable, and compatibility with existing simulators makes things easier for the user.  But you'll notice my highway numbers are much higher than those of Simulators A and B.  Although the numbers themselves cannot be taken from real life, the relationship between road capacity per lane and highway capacity per line can be, as this relationship is always fixed within a small region.  What is this relationship?  Here's a quote from the editor of TollroadsNews.com, from the article Expressways to Boulevards—A Bad Idea:

    QuoteBoulevards inevitably perform worse than expressways in carrying through traffic. Their capacity is about 60% lower, and they have more than double the fatality rate. Whereas an expressway lane carries up to 2500 vehicles/lane/hour, a surface arterial of which a boulevard is a variant carries only about 1,000 to 1,100 vehicles/lane/hour. That is because speeds are lower and because of time stopped at signals, or negotiating intersections, as opposed to the grade-separated nature of expressways.

    So this means that a typical highway lane will carry 2.5 times the volume of a road or avenue line.  Therefore, I obtained the RHW capacity by multiplying the road capacity by 2.5.  With the Maxis highways, it's a little more complicated, since they have three lanes per tile compared to a road's two lanes.  This means that for these highways, the 2.5 must be multiplied by 1.5, giving a total multiplicative factor of 3.75.  So the Maxis highways have 3.75 times the capacity of a road.  It's interesting to compare this to the original Maxis simulator where roads have a capacity of 1000 per tile, and highways have a capacity of 4000 per tile - the ratio is very close to what I calculated.

    But what about the rail numbers?  These are definitely higher than for Simulators A and B.  Yet they should be.  Look at what the nature of congestion is for car traffic versus something like subway traffic.  With car traffic at rush hour, the more cars you put on the road, the slower everybody goes.  This is reflected in the Congestion vs. Speed curve.  Yet for subways, the more people you have traveling by subway, the more crowded the subway cars get.  But they don't go any slower.  If you have huge crowds of people waiting for a subway train, it may take a few seconds longer to get on and off at stops.  But while the subway train is traveling, which it spends most of its time doing, it always travels at the same speed, regardless of the time of day, which is totally unlike cars.  So the effect of traffic volume on speed is quite different with rails than with road traffic.  This is reflected in the relative capacities of the two types of transportation.  For example, the new Second Avenue Subway in New York City is designed for a daily capacity of 600,000 people.  Meanwhile, directly above the subway line is Second Avenue itself, a four lane, one way avenue.  According to the quote above, this avenue will carry about 1000 cars/lane/hour, or 4000 cars per hour.  But in rush hour, it's going to carry a lot less, as traffic barely creeps along in Manhattan during rush hour.  So the subways (as well as the other rails) can carry vastly more traffic than the roads, especially during rush hour.  This is not reflected in Simulators A or B.  And if anything, my rail numbers should be even higher than they are to be realistic.  But once again, realism has to be balanced with what works for SC4, which means that these numbers should be low enough to give some challenge.  I have selected the numbers I use through experimentation, with the goal being that rails get congested slightly slower than road traffic.  Experiments with Simulator A have shown that the opposite is true, which makes traffic in its cities less realistic.
  • Complete compatibility with with RHW 3.0.
  • Much improved intersection effects.  The vanilla Maxis simulator is designed to decrease network capacities around intersections, thereby increasing congestion on busy networks, and indirectly decreasing speed.  The original Maxis numbers were too high to simulate intersections with stoplights; at best, they might simulate a four-way stop.  In Simulator A, these numbers were raised much higher, so that in most cases they had very little effect at all.  Sims would zip through intersections, almost never even slowing down.  The coming introduction of NWM and wider RHW complicated this problem, as the intersection square effect has an effect on many of those roadways over their entire length.  So produce any significant intersection effect without creating problems for NWM and RHW 3.0, only the capacities of the squares one and two squares away from the intersection can be modified.  I have done this in a way that produces a significant slowdown in the squares surrounding an intersection on a busy road, but due to the abovementioned limitations, it is not possible to simulate even a short stoplight cycle.  Nevertheless, at busy intersections, you may now see squares immediately surrounding the intersection turn congested, sometimes showing red; this at least makes the Sims slow down for the intersection.  One square beyond the affected square may also become congested, but usually by a lesser amount, as the Sims begin slowing down for the intersection.
  • More flexibility in balancing traffic loads.  This is described in detail below.

Although a number of people have posted their experiences with Simulator Z in this thread, and some specific numbers have been posted, I have not posted the results of any specific tests myself.  I am doing so in this post first, to back up some of the more important claims I have made for Simulator Z, and second, to show the simulator is significantly different from Simulators A and B.  For all of these tests I am using the Beta 1 version of Simulator Z.  Beta 1 is much more flexible about routing Sims to their jobs; it uses the transportation infrastructure as efficiently as possible.  Specifically, it uses the new Load Stabilization Dynamics™ technique (LSD) to balance the transportation needs of Sims.  So although the simulator generally produces 2.5 to 6 times as much car traffic as other simulators (which is really nice if you plan to use the wider versions of RHW or NWM), in cities with very extensive mass transit systems Simulator Z has been observed producing fewer cars than Simulator B at a similar capacity.   As a result, there is no longer a separate "MT" version with Beta 1, as it is no longer necessary.  With the Sims using LSD for all their travel, their journeys seem to go by in a flash.  (This may also explain some of the accidents you see, although the accident rate in general is the same as before.)

For those knowledgeable about the simulator properties who are wondering what the heck Load Stabilization Dynamics is, it's just my name (chosen largely for its acronym) for an emergent property of the simulator that results from modifying a number of seemingly unrelated properties in a very special way.  Travel strategy percents were not changed, nor were speed limits raised; instead, the relationship between selected properties was changed.

One of the first questions that needs to be asked when we examine a traffic simulator is, What exactly are we trying to simulate?  Calculations based on reliable sources showing that a large tile with a population of 430,000 has the same population density as Manhattan, which is apparently the most densely populated place on the planet.  Other figures show that only 18% of Manhattan residents drive to work, while 72% take mass transit.  Having lived in Manhattan, I can believe those figures.  However, they don't take into account all the people who drive into Manhattan to work, which definitely boosts the percentage of people commuting by car, and therefore the number of cars on the streets.  (New York is the second most congested city in the U.S. in terms of road traffic, and that's New York City as a whole, so you can imagine what a mess Manhattan is.)

The problem here is that SC4 doesn't correspond very well to these figures.  Here's the Zones view of one of my Chicago cities, specifically the mostly residential Logan Square/Humboldt Park area:



Most of the residential area is zoned medium density, with a much smaller amount low density, and one tiny square in the lower left quadrant is high density.  This is exactly what the zoning in this part of Chicago is like, with the exception that I had to put in a few high-density commercial areas to provide enough jobs for now.  But the residential capacity is zoned as closely to reality as the game will allow, and the resulting city looks somewhat similar to this part of Chicago.  This is far from downtown; it is certainly no Manhattan.  Yet the game population is approximately one million, many times higher than it should be.  So something's wrong here.

In one post, Tarkus says the following:

QuoteThe game's counting of residential population is already over-inflated as it is, too, by a factor of about 2:1 or even 2.5:1.  David (dedgren) did a lengthy series of posts the game's scaling in terms of population and density as a part of his long-running Project Three Rivers Region which illustrate this concept very well.  He's in the process of moving the original posts from ST over here, but in the meanwhile, you can read about his thoughts on the matter starting on Page 31.

But an even more direct pointer to dedgren's conclusions is here.  I was unable to find Tarkus' ratios of 2:1 or 2.5:1 mentioned anywhere in either link.  However, at one point dedgren concludes (using Chicago as a model, no less!) that population density in SC4 is 10 to 20 times too high; in another example, he concludes it's 10 times too high.  This corresponds with my experience exactly.  Having built a half dozen tiles of Chicago exactly to scale, right down to the streets, with correct zoning, I estimate that if I ever finish it, the total population will be about 30 million - about 10 times too high.  So taking this into account, the aforementioned model of Manhattan-style transport would start applying to SC4 cities somewhere in the 4.3 million to 8.6 million range.

So what about everybody else?  According to the source jplumbley quotes, 88% of all travel in the U.S. is done by car.  This article from the Washington Post repeats the 88% figure, adding that in Europe, the figure is 78%, but "Europeans are gaining on us."  The article adds that public transit accounts for 5% of travel in the U.S.; in Europe, it is 16% and declining.  One might assume that in big cities, more rapid transit is used, but that's not always the case.  Although New York City as a whole is the biggest user of mass transit in the country at 53%, Denver has only 5% mass transit usage, and Fort Worth, Tulsa, and Oklahoma City are all around 1%.  And at the top end, after New York City, the second highest mass transit user is Washington, D.C. at 38%, and the numbers go down from there.  (Source:  SustainLane.com's 2008 US City Rankings.)

But who wants to build a city with SC4 that has only 5% or 1% mass transit usage?  Not many people, as such a city would have some of the best features of the game hardly used or not used at all.  So a realistic balance needs to be struck between realistic transportation usage and game enjoyment.  Ideally, the player should be able to build any type of city, and the game should respond in a realistic manner to the city's architecture.  I know that that is what the CAM and NAM teams (among others) have tried to do, and that is what I have tried to do with Simulator Z.

Now a good traffic simulator for this city should produce traffic patterns that are fairly realistic for the city I've shown above.  This city has virtually no subways, and one main elevated line running parallel to the blue diagonal (Milwaukee Ave.).  I've also restored the old Chicago Surface Lines, which are tram lines that run along the main avenues, each separated by exactly a mile.  Here's the Traffic Volume Graph for Simulator A (Easy) for when the city is in equilibrium:



Here's the traffic graph for Simulator Z, Beta 1:



The biggest difference, obviously, is the car volume, which is 3.75 as much in Simulator Z as in Simulator A.  The bus volume is about 25% higher in Simulator A, but this is a surprisingly small difference when you consider that buses count as traffic in Simulator Z.  The rail usage is identical in the two simulators, but elevated train and tram usage is 60% higher in Simulator Z.  The trams run at the speed of elevated trains (unrealistic, but not fixable), so the Sims are just being much smarter about taking the fastest method of mass transit in Simulator Z.  So Simulator Z has much more realistic car usage for an area like this, while simultaneously making much better use of mass transit.

So what about congestion?  I should mention that the current versions of Simulators A and B have a bug where congestion is not displayed properly; it is displayed as being much less than it actually is.  The congestion is still there; it's just not being displayed.  I have fixed the copy of Simulator A that I am using to eliminate this bug so that congestion is shown properly; no other properties of the simulator have been changed or affected.  I believe that this bug is going to be fixed in the next release of the NAM.

Here's the congestion map for Simulator A (Easy).  I ran it both with and without the congestion display fix; in this case, the results were identical:



Notice that there's almost no congestion at all in the whole city.  (The red square is the Newman Hospital, which has a road that needs to be fixed.)  There's not even any yellow.  Just about the only congestion is the passenger rail line near the bottom of the graph, which is solid red.  How bad is this congestion?  If you look at the game view directly above the graph, you see that the rail line has 45,101 for the morning commute.  And the evening commute is even a little worse - it's a little over 46,000.  So that's over 91,000 Sims on a line with a capacity of 13,500, which means that the line is running at 674% of capacity.  And it stays that way, too.  This is true even on the unmodified Simulator A, which is supposed to prevent things like this with the low numbers in its Congestion vs. Speed curve.

Here's the congestion graph for Simulator Z:



Notice that unlike Simulator A, there is slight congestion in some areas, which is just right for an urban residential area such as this.  The congestion also occurs largely around the high density areas, such as the lower left corner and just above the red square, along with a main commute corridor just to the left of the river near the bottom. This is basically what one would expect.  But the rail line that was so congested in Simulator A is not congested at all here.  This isn't just due to a difference of capacities - the actual use of this rail line in Simulator Z is only one third as much.  The Sims' commuting is just distributed much more evenly here, and is more realistic overall.

Finally, here is another view of the same area, taken with a larger picture so you can see where some of the actual congestion is.  Note that the larger picture shows congestion on routes in the commute corridor going from the residential area to the industrial area, which makes complete sense since this is rush hour.  (It's always rush hour!  :D)



Now I'll move on to the Near South Side of Chicago.  Here is the Zones view:



Once again, notice that the vast majority of residential zoning is medium density, although there is a fair amount more of high density zoning than in the previous city, both residential and commercial.  The city's population is 1.9 million.

The next few pictures compare Simulator A at its Hard, Medium, and Easy settings with Simulator Z at its Low, Medium, and High capacity settings.  I have used the term "Low" rather than "Hard" because the Low setting may often be used in rural areas, where it is not difficult at all; conversely, there's nothing easy about using the High setting in a city of several million.  I should also mention that the extensive testing confirmed something of which I had previously seen only occasional evidence:  Simulator Z is much faster than Simulator A or B.  In this particular city, the game runs more than twice as fast at high speed when Simulator Z is used, and the difference in speed grows as the population of the city grows.  The reason for this is that I have been able to eliminate some of the exponential properties of the pathfinder which are still present in Simulators A and B.  And since the traffic simulator accounts for only part of the time it takes to run the game, a doubling of total game speed means that the traffic simulator is running about three to four times faster.

First I'll start with the congestion view of Simulator A (Easy).  In each set of views, I've included monorail usage at the exact same point for comparison.  All simulations started from the same point, and as you'll see from the traffic volume graphs, they were run until equilibrium was reached, which for this city took five years.  Monorail usage here is 5361, which is quite reasonable for this particular line.  Overall, there's a moderate amount of congestion for this city which is not terribly dense by SC4 standards, even on the Easy setting.  Most of the congestion is on roads, while the yellow going across the bottom is on one of the rail lines.  This is reminiscent of the rail line congestion in Simulator A shown in the previous city, though not as severe.



Compare the above picture to the one below, which is Simulator Z on its High setting, which has approximately the same road capacities as Simulator A (Easy).  There is some congestion, but it is just a small fraction of what shows up in Simulator A.  The reason for this is that the pathfinder is much smarter in Simulator Z.  For those familiar with the simulator properties, the Pathfinding Heuristic is lower, but that's only a small part of what's responsible for the difference.



Now let's look at the Traffic Volume Graphs, which shed a lot more light on what's going on.  As with the Data Views, Simulator A (Easy) is shown first.  The query tool report of a car volume of 9037 on the avenue is for the benefit of a NAM team member who said he had never seen an avenue volume much above 5200; here it is, and with the standard Simulator A.



The graph below show Simulator Z; the monorail usage is only a little higher than that shown for the comparable version of Simulator A above.  But the graphs themselves are quite different; these are clearly very different simulators.  If you look at the above graph for Simulator A, ignoring the numbers, it has the various travel types aligned in approximately the same places that they were in the graph for Logan Square/Humboldt Park shown earlier; Simulator A uses approximately the same fixed proportions for travel types, even though the current city (the Near South Side) has twice the population of the previous city, and many more areas that are zoned high density.  But if you look at the graph for Simulator Z below, you see that not only is it quite different from the equivalent graph of Simulator A directly above, but it is also very different from the graph for Simulator Z for Logan Square/Humboldt Park.  In other words, by using adaptive heuristics, Simulator Z is able to take into account the different density distributions of the cities, and as a result it is much more able to route the Sims in a way that minimizes congestion while using the same capacity levels.  This explains the large difference in congestion shown in the two congestion views directly above, while the views and graphs for Simulator A show that it is unable to do this.



Now let's look at the above graphs in detail.  It's important to remember that these graphs show trips, and multiple trips may make up a single commute.  This is one reason pedestrian usage is so high - if you walk to the bus from your home and then from the bus to work, that counts as two pedestrian trips plus one bus trip.  So the numbers in these graphs can't be mapped to numbers of commutes in any way that results in complete accuracy.  However, by comparing the graphs for a single city at a given time from two different simulators, it's possible to get a very good idea of how the simulators are routing Sims.  And in fact, the results one gets from such an approach match up very well with the congestion views.

In posts about earlier versions of Simulator Z, I reported that car usage ranged from 2.5 to 6 times as much as in other simulators.  And in my Logan Square/Humboldt Park city above, the ratio is 3.75, which is well within this range.  But the beta version of Simulator Z is much more flexible, and will adjust car usage over a much wider range.  By comparison, car usage in Simulator A is rather inflexible.  For example, Simulator A's car usage in the Near South Side is essentially twice that of its car usage in Logan Square/Humboldt Park, reflecting the relative size of the cities.  By contrast, Simulator Z's car usage in the Near South Side is only 20% higher than in Logan Square/Humboldt Park.  This results in a ratio of car usage between Simulator Z and Simulator A of 1.36 in the Near South Side, well below the lower margin of 2.5 reported for previous versions of Simulator Z.  Why does this happen?  Although Simulator Z has a strong preference for car usage, this preference is overridden when it would cause unbalanced traffic flow resulting in congestion.  From the player's point of view, it appears that traffic is getting heavy enough that the Sims decide that they would rather take public transit.  As you can see from the congestion views, this flexibility does not exist in Simulator A.

There is one other significant difference between the two traffic graphs above.  If you compare them, you'll notice that bus and subway usage have pretty much switched places in the two graphs.  Simulator A relies far more on buses, while Simulator Z relies far more on subways.  Simulator Z's approach seems to make sense, in that subways are far faster than buses.  The success of this approach is verified in the congestion views above.

Now let's look at the same city run at Medium difficulty in both simulators.  Once again, road capacities are basically the same.  Here's the congestion view for Simulator A:



And here's the congestion view for Simulator Z:



The differences are much greater than in the previous case.  There is significantly more congestion in Simulator A, but in Simulator Z, the congestion increase is hardly noticeable.  In fact, it takes a bit of examination between the two Simulator Z views to see that there is actually a small increase in congestion in the Medium simulator.

Here are the corresponding Traffic Volume Graphs.  First, Simulator A:



And Simulator Z:



First, notice the monorail usage for the two simulators.  Simulator A reports no monorail usage at all, while Simulator Z reports a moderate usage of 3222.  So in Simulator A, even though network capacities have decreased and congestion has increased significantly, Sims have abandoned the monorail!  Why?  As mentioned in an earlier post, the monorail is a very tricky network, and requires special tuning over a number of simulator properties to work optimally.  Simulator A does not have this special tuning.

As for the graphs in general, notice how Simulator Z's Traffic Volume Graph is almost identical for Simulator Z above at the High capacity.  So even through these graphs are greatly different from Simulator Z's Traffic Volume Graph for Logan Square/Humboldt Park, it shows that once Simulator Z has figured out an optimal traffic distribution for a city, it sticks with it.  This makes sense, as anything significantly different would be, by definition, less than optimal.

The story is quite different for Simulator A, though.  The graph for Medium capacity looks very different from the graph for Easy capacity, shown earlier.  Car traffic throughout the city has dropped by 40%.  Bus traffic has dropped by 21%.  Since in Simulator A, buses don't contribute to traffic but are affected by it, this means that the significant increase in congestion from car traffic is slowing buses down enough to force Sims off them and to alternate transportation routes.  And where are they going?  To the subway, along with a lot of former car commuters.  Subway usage is up 292% compared to the Easy version of Simulator A, and is now much closer to that of Simulator Z.  Yet the fact that there is still much more congestion with Simulator A shows that distribution of travel types is only part of the picture.  The actual pathfinding in Simulator A is still not sufficient to come anywhere near the efficiency of Simulator Z, as shown in the congestion graphs.

Now let's run this city with the Hard version of Simulator A, and the Low capacity version of Simulator Z, which actually has a slightly lower road capacity than the Hard version of Simulator A.  But let's also throw in another congestion view for comparison, so we can see how Simulator A compares to the Maxis simulator.  For the first view, I have taken the Maxis simulator, but replaced its network capacities and maximum commute times with those of Simulator A (Hard):



Now here's the Simulator A congestion view:



Note that although congestion in the two above views occurs in somewhat different places, the overall amount of congestion is very similar.  The Maxis congestion view has slightly less congestion in the upper left quadrant and lower right quadrant, while it has more congestion in the lower left quadrant.  (The upper left quadrant is just as colorful in the two views, but in the Maxis view, there is much more yellow, compared to the red of Simulator A.)  The two views are quite similar in total levels of congestion, despite the fact that the top view is merely the plain Maxis traffic simulator using the network capacities and maximum commute times of Simulator A (Hard).

From the above tests, this surprising conclusion emerges:  When run with identical network capacities and maximum commute times, there is no significant difference between Simulator A and the vanilla Maxis simulator in terms of traffic congestion and congestion dispersion.

Now someone might say, "OK, to be fair, what does this city look like if you use the Maxis simulator with Simulator Z's network capacities and maximum commute times?"  This is certainly a fair question.  On one hand, with Simulator Z's higher rail capacities, which were justified near the beginning of this post, you would expect to see a reduction in congestion.  After all, it's only the network capacities that differentiate the different levels of Simulators A, B, and Z.  But if plugging these numbers in results in a congestion view which is essentially the same as Simulator Z's, then Simulator Z is clearly not doing any better than Simulator A when compared against the vanilla Maxis simulator.  So first let's plug in Simulator Z's numbers and see what we get.  We also make buses count as traffic, since Simulator Z's capacities are based on this being true.  Here's what we get:



Well, this is a different sort of mess.  If there's a reduction in overall congestion, it's not significant.  It's hard to say whether the overall congestion is more or less than the previous views, although it certainly seems to be in the same ballpark.  But this congestion is distributed over a much wider area, resulting in more yellow and less red.  The traffic flow is more balanced, although it is still at its worst in the areas zoned high density, as it should be.  This supports my contention that the rail numbers in Simulators A and B are too low, resulting in an unbalanced traffic flow.

But what about the congestion view using the full Simulator Z?  Here it is:



It's clear that this view contrasts greatly with the previous three congestion views.  So unlike Simulator A, Simulator Z significantly outperforms the vanilla Maxis simulator even when the latter is using identical network capacities and maximum commute times.  This should demonstrate conclusively that there are major strategies at work in Simulator Z that are not present in Simulator A.  But just to completely prove this point, let's take Simulator Z's network capacities and maximum commute times and stick them into Simulator A.  Once again, since Simulator Z's capacities are based on buses' counting as traffic, they count as traffic here.  Here is the result:



It's a little better than the standard Simulator A view (second in this series), but not much.  And it's still far more congested than the Simulator Z view directly above.

As a side note, I should point out that neither the Medium nor the Hard versions of Simulator A were designed or recommended to run on a city of almost two million.  So what I'm doing here is essentially a stress test - running these simulators in difficult situations to amplify effects that might otherwise be difficult to see.  On the other hand, Simulator Z at similar capacities does fine in this city.  So you may interpret that as you will.

Comparing the unaltered Simulator A and Z views above to the earlier, higher capacity ones, we see that the Simulator A (Hard) congestion view above shows considerably more congestion than Simulator A (Medium), with many yellow areas turning to red, and other areas showing congestion for the first time.  This, of course, is not surprising.  Simulator Z on Low capacity also clearly shows more congestion than Simulator Z (Medium).  However, what is striking is that Simulator Z on Low capacity, which is comparable to Simulator A (Hard), shows less congestion than Simulator A (Easy), which is displayed at the beginning of the Near South Side series.  It is hard to think of a more dramatic contrast between the two simulators than this.

Meanwhile, we can also see the difference between the two simulators in monorail usage on their lowest capacity levels.  Above the congestion view of Simulator Z, we see a monorail usage of 7821, which as before, is normal for this line.  In fact, it's the highest monorail volume we've seen for the monorail on Simulator Z, which makes sense, as the other networks are starting to get a bit crowded.  But directly below, right above the Simulator A graph, we see a monorail usage of 134.  Once again, the reason for this is that Simulator A is not tuned to use the monorail properly.

Now let's look at the graphs themselves.  First, as usual, Simulator A:



And now, Simulator Z:



Once again, the Simulator Z graph looks almost identical to its predecessors.  It's got a good traffic distribution, and it's not going to mess around with it.  And this time, Simulator A's Traffic Volume Graph looks very similar to its previous one.  Basically, Simulator A has done the best it can do.  But if you look at the congestion graphs above, you see that compared to Simulator Z, its best is not very good.

Now some people, when looking at the congestion view of Simulator Z at Low capacity, might think that it does its job too well.  If even in a city of nearly two million there's barely any congestion, where's the challenge?  The answer is simple.  This city was built before Simulator Z was developed, and it has a vast system of public transit.  Specifically, it has a very large system of underground subways.  That doesn't affect the previous comparisons, because both simulators had equal access to the subways.  But now let's turn the subways off.  The switch is around here somewhere... %confuso  There!  Now we rerun the simulators just the same as the last run, except there are no subways.  If you think traffic was bad last time, look at this.  First, Simulator A:



And now Simulator Z:



Overall, Simulator A has worse congestion than when the city had subways, but not uniformly.  For example, some of the big red patches in the upper left are gone.  Why should there be less congestion in some areas just because the subways are gone?  As for the monorail, its volume is now back up to 6308.  But with the subways out of commission, you'd expect much higher use than that for the monorail.  Indeed, in the Simulator Z view, monorail usage is up to 47,055.

Meanwhile, congestion in Simulator Z is substantially worse compared to previous runs.  If you look closely, you see it's still not quite as bad as that in Simulator A, but it's close.  Why did it lose its huge advantage over Simulator A?  The answer is one word:  Buses.  Like all other simulators, Simulator A uses buses leased from Simgularity Bus Lines.  These buses have the unique property that you can stuff as many Sims as you want into them, and never run out of space.  Yet at the same time, the buses are infinitesimally small, so they have no effect on traffic, although they are affected by it.  They're a little bit like black holes on wheels.  But Simulator Z uses standard city buses, which contribute to traffic.  And therein lies all the difference.  With the subways gone, Simulator A can stuff more and more Sims into buses without increasing traffic.  Even so, overall congestion increases significantly.  But Simulator Z is more realistic in this respect and has no such free option, so congestion increases even more.  Still, Simulator A's ability to stuff unlimited numbers of Sims into congestion-free buses is not enough to give its city overall less congestion than Simulator Z.

For a better comparison of the two simulators, let's level the playing field a bit and rerun the last simulation with Simulator A, except this time we'll make the Sims take regular city buses:



This view is directly comparable to the congestion view directly above it, which uses Simulator Z.  And of course, the one above that is Simulator A where buses don't contribute to traffic.  So it doesn't look like it's a good idea just to turn this feature on in Simulator A - I had to link it to too many other features in Simulator Z to make it work.  This is why you can't just look at properties individually - they all tie in together.   Simulator Z works as an integrated whole, and taking pieces of it just doesn't work.

Meanwhile, let's go back to that last congestion view using Simulator Z where the subways are turned off.  For those who like a challenge, Simulator Z certainly offers one here.  In reality, the Near South Side of Chicago has only one subway line running through it, so this view is more realistic in that respect.  But the simulator is running on its lowest capacity here.  For a situation like this, you'd want to run it on Medium, or maybe High to get realistic results.  And if you're trying to model a city like Denver, where only 5% of the population uses public transit, the Ultra simulator is the one to use.  The Ultra simulator is also useful if you're playing a city with a lot of intercity traffic, as you're dealing not only with your own city's traffic, but with traffic from other cities as well.

Does this mean that Simulator Z loses its advantage over Simulator A in cities without subways?  Not at all.  If you recall the first congestion views in this post, which were of Logan Square/Humboldt Park, you'll remember that Simulator Z showed a more realistic congestion pattern than Simulator A.  And as I mentioned at the time, this is in a city with essentially no subways.  So game realism is increased in all settings.  And if you gradually add subways to a congested city, the amount of congestion shown by the two simulators rapidly diverges, as the pictures of the Near South Side show.  To encourage more realistic subway use, and to make the game a little more challenging in a realistic way, I quadrupled the monthly cost of subway tiles back in the Alpha version.  So overall, the Sims are much smarter about how they travel, but the limits on their transportation systems (specifically buses and subways) are a bit more realistic.

So that's it.  I hope I've made my case that Simulator Z really is a next generation traffic simulator.  This post has concentrated on only a few topics, namely the ones that are easiest to display in static pictures and graphs.  But I think you can get a feel from this post that traffic is handled in a very different way from any other simulator, and this is very apparent during game play.  One of the things I hear most often from users of this simulator is that their traffic flow is more balanced and realistic.  I think the pictures and graphs in this post help show why this would be so.


A Request

What happens next is up to you, the players of this game.  First of all, if this sounds at all interesting, you can download the latest version of Simulator Z from the bottom of this post.  I would appreciate feedback from everyone who uses this simulator, including any problems you may have.  (You shouldn't have to worry too much about this; there have been no known bugs for the last two versions.)  And along with feedback, there's a very important question I would like to ask the community.  If this traffic simulator really has unique advantages, as I have tried to show, or even if it is simply preferred by many users, as comments throughout this thread show, do you think this simulator belongs as a standard part of the NAM, along with Simulators A and B?  I am not a member of the NAM team, and understandably, they are not going to include this simulator in the NAM just because I ask them to do so.  But if enough people make it clear that they'd like to see this simulator included in the NAM, I think it could make a real difference.  So if you'd like to see this as an option, please post in this thread, even if you haven't actually used the simulator.  Also, I consider the Traffic Volume Views enclosed with the simulator to be connected to it, so if you would like to see them in the NAM, please mention that as well.  Thanks!

Meanwhile, here's the simulator - Enjoy!

EDIT:  This version of the traffic simulator has been superseded.  You can get the current version here.

b22rian

Hello Sim 4 devotion community.. and good day to you !

After several weeks of testing and now using Traffic sim Z..

I have now made the switch over to it from formerly using traffic sim's B and A..

I would like to fully endorse its use by all for the following reasons...

the reasons are on a comparative type basis after many months of testing and using traffic sim's A and B

on the hard difficulty settings, from the time those sim's were released for use..


1) .. A more balanced traffic type distribution patterns through-out my cities..

2) ...a higher percentage of car usage compared to traffic sim A as i eagerly await the super- highways
       of the RHW 3.0 release..

3)... very strong usage of all my mass transit systems.. Including higher monorail usage to the tune of 30 K
      increase over sim A.. ( I have a complex and extensive monorail system in my largest city..)
{ One may ask here how can you have both a higher car and mass transit usage for a given population ?..
    a fair question.. but the reason I give for this is.. because in sim Z , ive noticed that sims will look for
    a find the most efficient routes to and from work..often times this involves mutliple forms of transit on
    a given trip..So where as previously in sim A a sim would put up with simply driving a car all the way
    into work being throttled by long traffic jams.. The now smarter sim in the traffic Z enviroment will
   take advantage of multiple forms of mass transit if necessary to reach his destination quicker.. Although
   to best take advantage of these types of things, one must have planned and constructed the proper
   transport options in their cities..i may add.}

4).. so as a result better pathfinding in sim Z..

5) .. there is less game lag using sim Z than other traffic sims..

6).. and this should come as no surprise to anyone..but better zone development across the  board
      for the 3 zone types and higher populations.. being that "traffic" is such an integral part of sc 4.
   

also, I would like to encourage the addition of Sim Z into the inclusion of the next Nam release as a viable
option to the other traffic sim's.. And I encourage all to at least give sim Z a try , and make their own
decision as to if they want to use it as their traffic sim.. I can only tell you that ive used many of the
traffic sim's for years now playing this game.. and i would not go back to using any of them now that I
am using traffic sim Z..

Id like to hear some feed back from others as to what they think ?

thanks , Brian

z

#96
I am pleased to announce the return to this thread of my post announcing the Beta 1 release of Simulator Z, along with a downloadable version of the simulator itself.  The post also includes the results of extensive tests (with LOTS of pictures), illustrating how different simulators perform in different situations.  I made the tests especially extensive to answer some of the questions that I thought might arise, and some of the results were quite unexpected.  This post contains the most extensive testing of my simulator of any post in this thread, and for that reason, as well as for the Beta 1 release announcement, I consider it to be one of the most important posts in this thread.  It's located two posts before this one, but as it is quite long, you may find it easier to reach it by clicking here.  Although there are many tests shown, they cover only the parts of the simulator that were easiest to display.  I think that Brian's post, directly above this one, gives a well-rounded picture of what a typical user can expect to see with this simulator.

No problems are known to exist in the new Beta version of the simulator, and barring any surprises, this is the version that would be submitted to the NAM Team, should they decide to evaluate it.  So for those interested in this simulator, happy reading of the restored post, and I think you'll like the new version of Simulator Z.

z

I have added Park and Ride versions to all four variations of Simulator Z, making a total of eight simulator versions.  The Park and Ride feature of Simulator Z is identical to the Park and Ride feature of Simulators A and B.  What this means is that with these new versions, cars are not allowed to carry Sims all the way to their destinations (even if their destinations have those beautiful little built-in parking garages in back).  Instead, to make maximum use of this feature, you need to construct adequate parking space near mass transit stations in residential areas.  (Some stations, such as the Maxis train station, already have parking included.)  I have found RalphaelNinja's Ninja Boulevard Station and Ninja Boulevard Kiosk to be very useful in general, and particularly useful for Park and Ride, as they both contain large underground garages.  I have enclosed modified lot files for both of these stations; the modifications bring the monthly cost down to be more in line with similar stations, and they also double the capacity, bringing them more in line with RTMT (and making them especially useful for Park and Ride).  Since the exquisite design of these stations must have cost a lot of money, I left the plop cost alone.  Also, since I have enclosed only the lot files, if you don't have these stations already, you'll need to download them from the above links.

Park and Ride has a couple of other tricks that you might want to try.  If you use Park and Ride but don't build any extra parking (and don't use stations that have parking incorporated), you basically force your Sims to use mass transit.  In this case, you would want to have a very extensive mass transit system.  And if you use Park and ride with no public parking and no mass transit, you end up forcing your Sims to walk to work.  (It's basically Park and Ride with no place to park and nothing to ride.)

The package I've enclosed here is identical to the previous release, except for the addition of the Park and Ride versions and the revised Ninja Boulevard station lots.  If you have the previous release and are not interested in these additions, there's no need to download the current release, as the standard simulator versions have not changed at all.  But for those who want Park and Ride, now it's here.

sumwonyuno

Hello z,

I too am making creating a region based on a real world city.  I started using this simulator a few weeks ago.  I chose the Low setting, and I have tested it from the suburbs through to major employment areas.  This beta version is the best traffic simulator I've used yet (not saying that it's perfect, but it has strong advantages over other simulators).  The first and most obvious thing I did notice is the game (unpaused) runs much faster than the previous alphas and Sim A or B.  On top of that, more often than not, the route query seemed to be updated much quicker (a few weeks to a few months, rather than several years).

I do agree with your reasoning for increasing capacities.  I would like to add that since SimCity only simulates rush-hour traffic, and that the traffic is simulated as such as if it were occurring all at once, I do not believe treating network capacities as if they were per-hour is the correct action to take.  But I do admit there has to be the "challenge" element (isn't replicating a metro area enough?) and the simulator has to fit a wide range of uses.  I'm glad you have made the Low/.../Ultra versions to address these.


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

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

sumwonyuno

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.


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

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