Menu

LEX File Exchange
EA Support Files
SC4 Wikipedia
Network Addon Mod
Dependencies
Chat
Welcome, Guest. Please login or register.
Did you miss your activation email?

October 26, 2021, 03:57:49 AM

Login with username, password and session length

Author Topic: NAM Unified Traffic Simulator Development and Theory  (Read 211987 times)

0 Members and 1 Guest are viewing this topic.

Offline z

  • BSC Team
  • Forums Legend
  • *
  • Posts: 4889
  • Total likes: 52
  • Reputation: 26
    • RTMT World Headquarters
  • CL: Detailman!
Re: Traffic Simulator Z Development
« Reply #360 on: November 25, 2009, 02:33:25 PM »
[EDIT]:  All right, I'll be sending both of your copies sometime this afternoon.

That sounds good to me.  I think there's a wealth of information here, and the more people working on it, the better.

Offline sumwonyuno

  • Forums Senator
  • *
  • Posts: 736
  • Total likes: 6
  • Reputation: 4
  • what does this do?
Re: Traffic Simulator Z Development
« Reply #361 on: November 25, 2009, 04:38:43 PM »
All right, the DVDs are sent and they should reach your respective homes sometime after Thanksgiving (December?).  If anything's wrong with them (blank, unreadable, broken, MIA).. well we'll see in a few days.  You guys are welcome to do whatever you want with the region (except say it's your own!  $%Grinno$%).  In a couple city tiles I had to do things to get a certain visual effect.  You may have to demolish some plopped buildings here and there, and activate cheats.  However, those things won't probably affect the results of the testing.

Now, variables overloading the same parameter?!  I've never heard such a thing...  &sly

So, for a lower pathfinding heuristic (better pathfinding), the closest (appropriate and available) job isn't so attractive; the higher the pathfinding heuristic (less desired pathfinding), the closest (appropriate and available) jobs is very much attractive.  Hmm, you guys may want to check East Capitalis.  A large chunk of commuters aren't going west to Downtown, but are going to suburban city tiles or going east to nowhere.


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

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

Offline z

  • BSC Team
  • Forums Legend
  • *
  • Posts: 4889
  • Total likes: 52
  • Reputation: 26
    • RTMT World Headquarters
  • CL: Detailman!
Re: Traffic Simulator Z Development
« Reply #362 on: November 27, 2009, 12:32:16 AM »
I've recently been looking more closely into the requirements for RHW and NWM compatibility for traffic simulators.  In general, I think that the current settings of Simulator Z are fine here, but there's one area in which I'd like to make a somewhat significant change, and I'd like to get some feedback before I do.

One of the requirements for NWM compatibility is that the capacities and speeds per tile of roads, one-way roads, and avenues all be the same.  Not knowing how NWM was implemented, I simply have gone along with these requirements up until now.  But taking a closer way that one-way roads are used, both by themselves and in NWM, I want to question that particular restriction, and possibly change its implementation in Simulator Z.

In NWM, one-way roads are used only as part of the one-way road extensions; they are never used as part of two-way avenues.  So any changes to one-way roads would affect only the one-way thoroughfares in NWM.  So the question arises:  Do these really need to be the same speed and capacity of roads and avenues?  If not, what should their speed and capacity be?

In the original Maxis simulator, the speed of one-way roads was identical to that of regular roads, but their capacity was doubled.  These ratios hold true in Simulators C, D, and E as well, where all network capacities are multiples of the original Maxis capacities, and the speeds are the same.  There is a great advantage to using one-way roads in these simulators.  In Simulators A, B, and Z, the capacities and speeds of one-way roads are the same as regular roads.  Nevertheless, due to the asymmetry between morning and evening commutes over various routes, and the fact that congestion is calculated on a whole-day basis, there is a slight advantage to using one-way roads in SC4.  My observations have shown that that advantage will range somewhere from 0% to 20%.  In practice, this is barely enough to notice, and I have received numerous complaints that one-way roads are now almost useless, and not worth the trouble of building.  To be honest, my experience supports this point of view.

What does traffic engineering say about the speeds and capacities of one-way roads?  The few sources I have been able to find agree on a single number:  that the capacity and effective speed of one-way roads is in general two-thirds higher than that of equivalent two-way roads.

That's a big difference from the current state of things.  How does this compare to what we can observe?  First, a basic principle of traffic engineering (and actually of physics) is that for comparable roadways, capacity is proportional to speed.  If you double the speed of traffic on a road, then the road's capacity - the number of cars passing over a given spot per unit of time - also doubles.

Next, let's look at some RL examples.  For example, here in beautiful Eugene, Oregon, which is a small city of about 150,000, there are a number of one-way roads and also wide one-way roads that travel across town.  There are 11th and 13th Avenues, which are two-lane, one-way roads, and then there are 6th and 7th Avenues, which are four-lane, one-way roads.  Speeds and capacities on these roads are basically the same.  Especially near the downtown area, there are stoplights at virtually every block, although these thin out the farther away from downtown you get.  And like most major one-way roads, the stoplights are timed to the speed limit.  On any one of these roads, once I pass an intersection with a green light, I can set my car's cruise control to 30 mph and ride straight through to the end of the one-way road, right through all the stoplights, unless the road is heavily congested.

Contrast this with the two-way, two-lane roads that go through the busier part of town, which is where the one-way roads are.  There is no way to sync stoplights for two-way roads, so where there are a lot of them, you have to stop frequently.  At other intersections, there are stop signs.  And even for major two-way roads without stop signs, there is generally one lane in each direction (as in SC4), and traffic turning both on and off the road tends to slow the road down considerably.  For example, traffic on Willamette, which is one of the main roads in Eugene, tends to travel much slower than traffic on the one-way roads I mentioned, even over those sections where there are no stoplights.  Overall, it's easy to see in this town how the two-thirds speed and capacity rule applies.

What about big cities?  An excellent example is the biggest city in the U.S., New York - specifically Manhattan.  There's a saying in New York something to the effect that, "You can go uptown, and you can go downtown, but you can't go crosstown."  At first, this may seem a little strange, as Manhattan is a narrow island, and crosstown is the shortest trip you can make.  But Manhattan has many big, wide one-way avenues going uptown and downtown, while the major crosstown avenues are two-way, even though they have the same number of lanes as the one-way roads.  It is not unusual for the wide one-way avenues to carry traffic at several times the speed of the two-way avenues (technically they're called streets), even though the nominal speed limits are the same.  And the East Side avenues are efficient enough at carrying traffic that unless you're traveling more than a few dozen blocks, it's really no faster to go over a few blocks to FDR Drive, a major highway that parallels the avenues.

Some may say that one-way NWMs shouldn't have a significant speed or capacity advantage over avenues.  But in the real world, they do.  Look at Manhattan again.  On the Upper West Side, Broadway, which is a four-lane, two-way avenue, parallels Amsterdam and Columbus, which are wide one-way avenues.  Anyone who wants to travel more than a few blocks by car will take Amsterdam or Columbus, as they are much faster than Broadway, even though the speed limit is the same.  So for all these reasons, I don't think that raising the speed and capacity of one-way roads by a realistic amount makes a traffic simulator not compatible with the NWM.  To the contrary, I think such a change makes it more compatible.

What about all the side streets in Manhattan that are one-way?  They don't have the advantage of timed lights, and their speed and capacity is about the same as normal streets.  Recently, a proposal has been made for one-way streets in SC4.  I think this is great idea, as they would have the same capacity and speed as regular streets, and would provide a nice alternative to higher-capacity one-way roads.

Speaking of alternatives, boosting the speed and capacity of one-way roads would have yet another benefit.  A long-standing complaint in SC4 is that all roads and avenues have the same speed limit.  If one-way roads have a higher, more realistic speed limit, then players can immediately construct their own higher-speed avenues.  Eventually, higher-speed forms of NWM could be made this way.  I believe that higher-speed two-way roads could also be made using the one-way network; someone with more knowledge of this process can confirm or deny this.  Assuming they could, they would serve as an excellent intermediate between standards roads and RHW-2.  And unlike RHW-2, residences and businesses could be built right up against them.

I mentioned earlier that one-way roads are considered to give a two-thirds boost to capacity and speed, and I also mentioned that the current setup for Simulators A, B, and Z gives a boost of from 0% to 20%.  I am proposing to increase the speed and capacity of one-way roads to be 50% higher than that of regular roads, which should make them come out about right.  As I mentioned, this will immediately make faster hand-made avenues possible, while leaving regular avenues unchanged.  What do people think about this proposal?  If there is enough popular response, I can post a full release of Simulator Z with these changes in this thread within the next couple of days.

Offline b22rian

  • RTMT Team
  • Forums Guru
  • *
  • Posts: 3434
  • Total likes: 664
  • Reputation: 13
  • CL:
    ...let us ch@t...
Re: Traffic Simulator Z Development
« Reply #363 on: November 27, 2009, 05:51:14 AM »
OK,

     Thanks a lot Steve for looking into this .. also your earlier posting on " nearest Destination Attractiveness"
parameter was flat out amazing  &apls ..And you are really unravelling the mysteries which surround understand
the traffic sim !

     I like what you said here, and ive thought about OWR's in the same ways you have at times..
But I was thinking the 50 % speed bonus to them over roads was just a bit too high.. I would say
the speed bonus should fall some where between 20 - 50 % , but not sure of an exact figure...

Id like to hear what others want with this ?

Brian

Offline gottago

Re: Traffic Simulator Z Development
« Reply #364 on: November 27, 2009, 06:31:32 AM »
I'll admit I've been lurking on the traffic sim development threads for a while now and following the discussions--great work going on here.

Z, re: your post on OWR,  I'd heartily endorse the 50% higher speed and am frankly surprised that there is no current advantage to them, as there obviously should be, and as you detail from RL examples and stats. And 50% seems entirely reasonable considering the advantages they offer in terms of the greatly increased capacity (or to borrow a term, "throughput") they offer, which you note must be defined in terms of speed.

Edit to add: this might be a good subject for posting a general poll for wider feedback.
« Last Edit: November 27, 2009, 06:38:59 AM by gottago »

Offline RippleJet

Re: Traffic Simulator Z Development
« Reply #365 on: November 27, 2009, 07:39:57 AM »
I am proposing to increase the speed and capacity of one-way roads to be 50% higher than that of regular roads, which should make them come out about right.

Every time the network speeds have been tweaked, they've been increased... ::)
How about this time lowering the speeds of roads and avenues instead?

On the other hand, I would also want to hear from Alex if and how this would affect the NWM... ;)

Offline ldog

Re: Traffic Simulator Z Development
« Reply #366 on: November 27, 2009, 10:01:21 AM »
Every time the network speeds have been tweaked, they've been increased... ::)
How about this time lowering the speeds of roads and avenues instead?
Faster! Faster! Faster! We all wanna go faster!

I'd just change the cap and leave the speeds alone. The lessened congestion will provide enough speed advantage.

50% on the cap sounds about right. Since I've been testing mostly with stock caps lately, I do think the 100% increase Maxis used is a wee bit much.

Offline RippleJet

Re: Traffic Simulator Z Development
« Reply #367 on: November 27, 2009, 11:09:57 AM »
Faster! Faster! Faster! We all wanna go faster!

 :D  Yeah, I know... ::)

Read this post and the rest of that thread... :)

Offline z

  • BSC Team
  • Forums Legend
  • *
  • Posts: 4889
  • Total likes: 52
  • Reputation: 26
    • RTMT World Headquarters
  • CL: Detailman!
Re: Traffic Simulator Z Development
« Reply #368 on: November 27, 2009, 03:11:25 PM »
Every time the network speeds have been tweaked, they've been increased... ::)

Not true.  The most recent version of Simulator Z reduced rail speeds by 20% to 30%, with the biggest reductions coming in the heaviest used rails - subways, el rail, and commuter rail.  All of these are now significantly lower than the original Maxis speeds.  The monorail is now equal to the original Maxis speed; I can't really drop it any lower since it serves HSR and BTM as well.  Also, roads and avenues are currently 10 kph lower than in Simulator A.  There are no plans to raise any other speeds.

I'd just change the cap and leave the speeds alone. The lessened congestion will provide enough speed advantage.

Speed and capacity go together for similar networks.  Not only is this a basic law of traffic engineering and physics, as mentioned in my original post, but it's reflected in the highways and RHW as well.

Quote
How about this time lowering the speeds of roads and avenues instead?

See the reference to your post below...

Quote
On the other hand, I would also want to hear from Alex if and how this would affect the NWM... ;)

I would as well.  That's one of the reasons I put in a detailed analysis of his home town.  But this proposal has been made in light of information I received from him recently about the workings of NWM.  Alex has already said that he thought that some increase in the one-way road capacities and speeds would be OK for NWM; the question at this point is how much.  I have tried to propose a realistic amount, and there will be plenty of time to discuss it and test it.

Read this post and the rest of that thread... :)

Yes, I've been thinking a lot about that post since it was made.  Basically, one of the points you make is that you get better mass transit usage with a wider spread of vehicle speeds, which is true.  But Simulator Z was starting to suffer from too much MT usage; this was a result of using Perfect Pathfinding.  I was getting complaints about roads being almost deserted, and not enough customers for businesses.  Hence the speed reduction, and people seem uniformly happy with the results.  Speeds are now pretty realistic (at least in terms of top speed) across the board.  I think that this was Maxis' original intent, but it had to spread the speeds out when it raised the PH to .09.  The changes I'm proposing to one-way roads simply bring realistic behavior to the one part of speeds and capacity where it has been sorely lacking.  Extensive testing has shown that the rest of the current speed structure works well for the simulator.  Lowering road and avenue speeds, as you suggested above, would bring back the car usage and business customer problems.
« Last Edit: November 27, 2009, 03:31:34 PM by z »

Offline ldog

Re: Traffic Simulator Z Development
« Reply #369 on: November 27, 2009, 05:34:55 PM »
Speed and capacity go together for similar networks.  Not only is this a basic law of traffic engineering and physics, as mentioned in my original post, but it's reflected in the highways and RHW as well.

And if this were reality that would be true. One big difference between the traffic simulator and reality; each run of the traffic simulator is a snapshot of a single point in time. So raising the speed only raises the amount of cars passing through because it makes the route more attractive to the pathfinder. Congestion has a pretty big effect on speed. Network capacity has a pretty big effect on congestion. So while the effect of changing the capacity is not 1:1 or even anything close, raising or lowering capacity from a gameplay view has a similar effect to raising or lowering speed; provided there is enough traffic to make a difference.

The important thing is that the (pair of) owr needs some kind of advantage over the (pair of) regular road or there is no point in using it. On the other hand a pair of  owr should not >= to an ave, but I think there is room to play around with in between. And of course they cannot be so superior to a road that there is no longer any point in using the road either (original streets were a good example of worthlessness).

So do what you want, it is going to work fine either way.

Offline Tarkus

  • Administrator
  • Forums Legend
  • *
  • Posts: 11708
  • Total likes: 5956
  • Reputation: 75
  • NAM Team Tankadillo
    • NAM HQ
  • CL: Dr. PuzzlePiece
Re: Traffic Simulator Z Development
« Reply #370 on: November 27, 2009, 09:54:44 PM »
But this proposal has been made in light of information I received from him recently about the workings of NWM.  Alex has already said that he thought that some increase in the one-way road capacities and speeds would be OK for NWM; the question at this point is how much.  I have tried to propose a realistic amount, and there will be plenty of time to discuss it and test it.

As I recall from the early discussions on "NWM compatibility", the reason we went for the equalization was necessitated as a side-effect of the more critical Road/Avenue equalization.  Without it, the "fake OWR-based 2-way Avenues" that some people have built would gain an advantage over the Avenue and the multi-tile 2-way NWM networks (an idea which you've addressed later in your post, which I'll touch on as well).  Personally, I think 50% may be a bit much, and somewhere in the 10-25% range is perhaps more appropriate.  That would still work out to about a 50% increase in capacity on 2-tile one-way networks.  It's too bad we can't set Intersection and Turn Capacity values separately by network, as most of the benefit of OWRs seems to come at intersections, due to fewer conflict points and the effect of the green wave (which is definitely in full force on the 11th/13th couplet in Eugene).

Speaking of alternatives, boosting the speed and capacity of one-way roads would have yet another benefit.  A long-standing complaint in SC4 is that all roads and avenues have the same speed limit.  If one-way roads have a higher, more realistic speed limit, then players can immediately construct their own higher-speed avenues.  Eventually, higher-speed forms of NWM could be made this way.  I believe that higher-speed two-way roads could also be made using the one-way network; someone with more knowledge of this process can confirm or deny this.  Assuming they could, they would serve as an excellent intermediate between standards roads and RHW-2.  And unlike RHW-2, residences and businesses could be built right up against them.

I mentioned earlier that one-way roads are considered to give a two-thirds boost to capacity and speed, and I also mentioned that the current setup for Simulators A, B, and Z gives a boost of from 0% to 20%.  I am proposing to increase the speed and capacity of one-way roads to be 50% higher than that of regular roads, which should make them come out about right.  As I mentioned, this will immediately make faster hand-made avenues possible, while leaving regular avenues unchanged.  What do people think about this proposal?  If there is enough popular response, I can post a full release of Simulator Z with these changes in this thread within the next couple of days.

Unfortunately, the problem with the One-Way Road network is that Maxis hardcoded a virtually irreversible "tidal flow" effect in terms of how that network handles path files.  The path file for the orthogonal OWR (0x09004b00) is identical to the path for orthogonal Road (0x00004b00), and at first glance, it looks like it is pathed for 2-way traffic.  However, when an OWR is built, the game effectively forces the "wrong way path" to go the "correct" way.  There are some rare occasions where a special "crossover path" can counter the flow (something I discovered by accident), but that effect wouldn't help here.  In essence, while it is possible to create a 1-way network out of a 2-way network (see the RHW), it does not appear to be possible to create a 2-way network out of a 1-way network.

Further, the way the OWR network relates its directionality to the RULs, when an OWR's direction is reversed, the network textures (or models) are not rotated, making it impossible to tell which direction an OWR is going from the RUL end.  This is part of the reason the traffic signal T21s on the OWR network are so skimpy (that and the lack of functional stop points). 

In all honesty, I think it'd be easier to turn the Avenue into a 1-tile network than to get the OWR network to accept 2-way traffic. ::)  Though that would open a whole other can of worms.  (And as Vince recently discovered, you can do some pretty wild stuff with median flags.)


Offline z

  • BSC Team
  • Forums Legend
  • *
  • Posts: 4889
  • Total likes: 52
  • Reputation: 26
    • RTMT World Headquarters
  • CL: Detailman!
Re: Traffic Simulator Z Development
« Reply #371 on: November 28, 2009, 12:31:56 AM »
I appreciate everybody's input here.  It sounds like there is a consensus that one-way roads could use some beefing up; it's just a question of how much is appropriate.  In the real world, they make a big difference in traffic flow, although drivers may need to take slightly longer routes as a result.  I think it would be good to see the same general results in SC4, as long as we can do so without causing unwanted side effects.

I'd like to address a couple of the points that Alex raised:

Quote
  It's too bad we can't set Intersection and Turn Capacity values separately by network, as most of the benefit of OWRs seems to come at intersections, due to fewer conflict points and the effect of the green wave (which is definitely in full force on the 11th/13th couplet in Eugene).

I agree completely.  It's because we can't simulate either of these effects properly in SC4 that increasing the capacity and speed seems appropriate, as the same effect is accomplished through different means.

Quote
In essence, while it is possible to create a 1-way network out of a 2-way network (see the RHW), it does not appear to be possible to create a 2-way network out of a 1-way network.

That's too bad; I was hoping this could be done.

Quote
In all honesty, I think it'd be easier to turn the Avenue into a 1-tile network than to get the OWR network to accept 2-way traffic. ::)  Though that would open a whole other can of worms.

It sounds like it would be best to pass on that.  I think we've got enough worms for now...  ::)

Personally, I think 50% may be a bit much, and somewhere in the 10-25% range is perhaps more appropriate.  That would still work out to about a 50% increase in capacity on 2-tile one-way networks.

From my own observations, as well as those I received from other users (which were all in the form of complaints about the current state of one-way roads), I think this overestimates the intrinsic advantage of one-way roads over two-way roads in SC4 by a fair amount.  I'd like to make a slightly different proposal here.  Instead of a 50% increase, let's try a 40% increase in an open beta.  On one hand, we can be rather sure that the results will not exceed the two-thirds premium that is standard in the real world, and what we should see should usually be less.  Yet 40% is certainly high enough to make a significant difference from the current state.

As far as hand-built avenues go, these would have a 40% premium over regular avenues as well as NWM two-way avenues.  The intrinsic difference between one-way and two-way roads does not come into play here.  Having a four-lane avenue with a speed limit 40% higher than a standard avenue is certainly quite common, especially in suburban areas; I think it would be a useful and welcome addition to SC4.  On the other hand, very wide avenues such as those in the coming NWM are typically found in more built-up parts of the city where there is enough traffic to require them, and in these areas the speed limits tend to be the same as roads.

So that's my proposal; we can try it out in beta form and see how it works.  What do people think about this?

Offline b22rian

  • RTMT Team
  • Forums Guru
  • *
  • Posts: 3434
  • Total likes: 664
  • Reputation: 13
  • CL:
    ...let us ch@t...
Re: Traffic Simulator Z Development
« Reply #372 on: November 28, 2009, 05:33:09 AM »


From my own observations, as well as those I received from other users (which were all in the form of complaints about the current state of one-way roads), I think this overestimates the intrinsic advantage of one-way roads over two-way roads in SC4 by a fair amount.  I'd like to make a slightly different proposal here.  Instead of a 50% increase, let's try a 40% increase in an open beta.  On one hand, we can be rather sure that the results will not exceed the two-thirds premium that is standard in the real world, and what we should see should usually be less.  Yet 40% is certainly high enough to make a significant difference from the current state.

As far as hand-built avenues go, these would have a 40% premium over regular avenues as well as NWM two-way avenues.  The intrinsic difference between one-way and two-way roads does not come into play here.  Having a four-lane avenue with a speed limit 40% higher than a standard avenue is certainly quite common, especially in suburban areas; I think it would be a useful and welcome addition to SC4.  On the other hand, very wide avenues such as those in the coming NWM are typically found in more built-up parts of the city where there is enough traffic to require them, and in these areas the speed limits tend to be the same as roads.

So that's my proposal; we can try it out in beta form and see how it works.  What do people think about this?


    Sure Steve , it sounds like a great idea to me to do a Beta first.. That way hopefully you can get plenty of
feedback from users as to how they like it and it affects their transit systems.. Just off the top of my head it
sounds maybe a little on the high side to me ( course maybe as its not been tested yet..) but I think I could live
with 40 %, if thats what it ends up being in the end.. I know i said a big range 20-50 %.. But after reading
 alex's post above.. now im thinking more on the lower end of that range.. Of course we can just as easily end
up testing other values such as 30 % and 20 % later on also, and im sure we will.

Thanks, Brian

Offline RippleJet

Re: Traffic Simulator Z Development
« Reply #373 on: November 28, 2009, 07:15:37 AM »
I'd be careful with overshooting too much, and would thus suggest no more than 30% in the first test.
It's always easier to increase it in a second beta test (if needed) than to lower it... ::)

30% would be roughly the same speed advantage Maxis gave to avenues over roads and owr's.
With a PH lower than 0.09, the advantage would anyways be bigger than that provided by Maxis for avenues.

Offline tamorr

Re: Traffic Simulator Z Development
« Reply #374 on: November 28, 2009, 12:10:32 PM »
  I personally welcome the increase in advantage, even though I am not able to do any testing, I would prefer there be said advantage, if only for OWRs... As that is a network down in such a way to counter the effects of roads becoming too much to handle. Well the is my take on it....


I would have to agree with RippleJet in that it is better to raise a value in small incriments than it is to lower said value. Makes things a bit more like the smoothing into what can be donate of the value that is being looked for. Ease into that which is sought out. Step by step we'll get there with a basis to be set... So I'd say starting at 30% would be best and increasing from there as said. :)
  "It is wiser to think about your actions before doing them, but be warned One must act quickly before another takes action for you."
  "Knowledge may be Power, but it is how you use that Knowledge that makes One Powerful."
  "I am a Philosopher, Punnist, Poet, and Rambler so keep in mind I think ahead and backwards to point where communication is sometimes not completely understood, even if Enlish is my primary language, it doesn't mean I know it well N proper."
  "Always do your best to acheive your goals and Dreams one at a time."
"Patience is a virtue."

Offline z

  • BSC Team
  • Forums Legend
  • *
  • Posts: 4889
  • Total likes: 52
  • Reputation: 26
    • RTMT World Headquarters
  • CL: Detailman!
Re: Traffic Simulator Z Development
« Reply #375 on: November 28, 2009, 04:27:09 PM »
As people generally know, I like to have as accurate a simulation as possible.  (I'm understanding my CML of "Detailman!" more all the time.)  At the same time, making users happy is at least as important.  So in light of recent feedback here, I decided to do some more research on one-way road capacities and effective speeds.  On one hand, I found quotes such as the following from the article Should Cities Convert One-Way Streets to Two Way?

Quote
"You need seven lanes of a two-way arterial to achieve the same capacity as four lanes of a one-way couplet," says transportation planner Michael Cunneen.

This would imply a 75% capacity and effective speed advantage for one-way roads, even higher than the 67% I had seen cited elsewhere.

On the other hand, further research showed lower figures quoted.  In retrospect, it is easy to see why.  Not all one-way roads have synchronized traffic lights, and if the traffic lights and intersections are far enough apart on two-way roads, the difference between the two types of roads narrows.  So the 30% increase being discussed recently, when added to the intrinsic difference between one-way and two-way roads in SC4, is within the realm of realism, although it's a bit on the low side.  The 30% figure does make it even easier to justify alternative, high-speed avenues, though.  Whereas the standard avenue in Simulator Z has a speed of 50 kph (31 mph), a hand-built avenue using one-way roads would have a speed of 65 kph (40 mph), which is a very common speed for avenues outside city centers, at least in the U.S.

The way capacities are set in Simulator Z, it works out a bit better to increase them by a third, as it still leaves the simulator with nice round numbers.  My experience is that speed and capacity differences of less than 10% are not really noticeable by the user, and do not have a noticeable impact on the game.  And current Simulator Z capacities are not exactly proportional from one capacity level to another, so this would be nothing new.  To give a specific example, the road capacity for Simulator Z (Low) is 2400; increasing it by 30% gives 3120, while increasing it by a third gives 3200.  The difference of 80 between these two numbers is small enough to get lost in the noise.

So barring any major objections, those are the numbers I will use for the beta release.  The 30% number is certainly high enough to make a significant difference, but I wouldn't go any lower than that, at least for this beta release.  Comments are welcome.

Offline delta9

Re: Traffic Simulator Z Development
« Reply #376 on: November 28, 2009, 08:52:44 PM »
Well, as a mathhead (that's an 'a', not an 'e' ;)) I'd say go with the 1/3 increase.  Rational multiples are always good  :D

I definitely agree that the OWRs should have the advantage for the reasons stated, for realism and gameplay's sake (one of the rare instances in which those two things overlap ::)), and agree completely with the resultant real/'fake' avenue speed difference.  For example, many of the well-traveled avenues in the north part of Austin such as Anderson Lane and Burnet Road are set at 30mph/~50kph (and are generally faced by CS$ to use SC4 terms ;D), while Lakeline Boulevard, the arterial avenue that serves a significant portion of residents in my northwestern suburb of Cedar Park, is set at 40-45mph/~65-70kph (and a red congestion view in Z Ultra :D).  There are plenty more examples of this around town and in other cities.  It would be a useful change, and would give me a reason to finally use Marrast's dividers that have just been festering in my plugins for so long $%Grinno$%

Of course, there's the whole scale thing that I think RippleJet/Tage (if I may) brought up, and if you, as you mentioned, developed a traffic sim around that, I'd be down for testing and whatever help I can provide (haven't done any hands-on modding yet but I have a good general understanding of the game's simulators and a bit of a background in programming).  The scale of the game has always been a nagging problem to me, from having 20+ people commute from a mansion to work, to having 40+ foot wide residential streets, 50 foot wide roads, 50 foot high embankments and bridges, etc...  If that's an idea you or anyone else would like to pursue further, you can rest assured that at least one person would use it ()stsfd()

EDIT: The scale issue wasn't brought up in here, but rather in this thread, as linked earlier.  But besides, I think the question of overall scale in the game should be based upon building footprint more than anything, and that's an entirely different topic, perhaps one to be pursued.  Back to traffic simulators...
« Last Edit: November 28, 2009, 09:12:40 PM by delta9 »

Offline z

  • BSC Team
  • Forums Legend
  • *
  • Posts: 4889
  • Total likes: 52
  • Reputation: 26
    • RTMT World Headquarters
  • CL: Detailman!
Release of Simulator Z V1.3 Beta 1
« Reply #377 on: November 29, 2009, 01:27:02 AM »
I've attached the Beta 1 release of Simulator Z v1.3 at the end of this post.  It contains the settings for one-way roads discussed in my previous post.  This is not a final release, so it can always be changed, but based on the discussion here, it sounds very likely that these will be the final numbers.  I've also included Simulator Z Classic, which is now a standard part of the Simulator Z release.  As per an earlier discussion with Alex, I've taken his suggestion for selecting a slightly higher avenue capacity and speed for this one version of the simulator, so that it fulfills the goal I had of more closely matching the original Maxis simulator.  Standard avenue capacities and speeds are unchanged in the other versions of Simulator Z.

Along with the revised version of Simulator Z Classic, for the first time I'm enclosing a Traffic Volume Data View designed specifically for Simulator Z Classic.  While constructing this new Traffic Volume Data View, I found a bug in the standard Traffic Volume Data View, which is used with the standard Maxis simulator; I've enclosed a corrected version of this data view for those who are interested.  Final versions of all of these files should be released with the next NAM.
« Last Edit: November 29, 2009, 01:31:59 AM by z »

Offline Shadow Assassin

  • NAM Team
  • Forums Guru
  • *
  • Posts: 3232
  • Total likes: 12
  • Reputation: 25
  • Needs more cowbell.
  • CL: The Grid Reaper
Re: Traffic Simulator Z Development
« Reply #378 on: November 29, 2009, 03:29:17 PM »
Hey, z, sorry for the late reply...

This looks suspiciously like the type of problem that sumwonyuno has been experiencing, in that it could be a case of Nearest Destination Attractiveness being set too low.  Could you please try changing this parameter (or Pathfinding Heuristic, as it is more often named) from .003 to .009, run your city for a couple of years, and see if the route changes?  If not, could you exit your city without saving, run Simulator A on it for a couple of years, and see if that changes the route?

I haven't had time to test this problem yet - but it does not seem to affect overall running of the city.

Having read Tage's reply and your response, pictures would be helpful for this problem.  Also, following the steps I outlined for #1 would be very helpful here too.

I see that Tage has answered this one, and I agree with his answer.

I haven't gone into that city tile since I asked about it - but the problem solved itself after upgrading a segment of street to road the last time I played the tile. Now only buses still remain (which are all coming from houses bus stops within the city tile now, rather than the rail station, as previously).

How long did you let the game run after upgraded the street to a road?  It can sometimes take between five to eight years for a major change to be fully propagated, even though traffic will start appearing on the road relatively quickly.

The game ran for about two years, but as time progressed the changes ultimately were successful. Now I have to upgrade that road to an avenue...

Anyway, just an enquiry: when I had two "towns", on both sides of a city tile, rolling Non-Job-Zots occured in some parts of town when I upgraded the ferry pontoon on the south side to a car ferry, and created a new car ferry on the north side. These non-job-zots went away when I created a road bridge further upstream. Abandonment was due to "commute time". The average commute time dropped from about 60 to less than 30, then the simulator fixed everything after five years or so of game time. Is this normal behaviour for when such a large infrastructure upgrade is undertaken?

It's just an enquiry - there is no problem otherwise.


Also - I'm having trouble seeing where existing bus stations are on the data view... I recall a mod of tropod's that showed bus stations on the relevant volume data view as bright green compared to the usual blue of the default view... is something like this available on the modified volume views? It would be exceedingly handy for making sure that areas aren't doubled-up, particularly with how well RTMT blends in with normal street furniture, just like it should in real life.
New Horizons Productions
Berethor ♦ beskhu3epnm ♦ blade2k5 ♦ dedgren ♦ dmscopio ♦ Ennedi
emilin ♦ Heblem ♦ jplumbley ♦ moganite ♦ M4346 ♦ papab2000
Shadow Assassin ♦ Tarkus ♦ wouanagaine
See my uploads on the LEX!

Offline z

  • BSC Team
  • Forums Legend
  • *
  • Posts: 4889
  • Total likes: 52
  • Reputation: 26
    • RTMT World Headquarters
  • CL: Detailman!
Release of Simulator Z V1.3 Beta 2
« Reply #379 on: November 30, 2009, 01:38:06 AM »
The Streets of Sim City

[First, thanks for your post, Shadow Assassin - I'll get back with a reply soon.]

So far, testing of the one-way roads has been giving results about what I expected - the changes have a definite positive impact, although I think they're a bit on the low side.  I will be very interested to hear from other people what they think of these changes, and whether they should be adjusted further.  But overall, I think that this is a positive development in adding flexibility to existing networks.

Along those lines, I started looking more closely at the street network.  Maxis always had an ambivalent attitude toward streets.  On one hand, they were highly integrated into the system; the game even creates them automatically when you zone large enough areas.  And if you turn all your streets into roads, you get complaints from the game.  Yet in the original game, the capacity of a street was a paltry 100.  This increased significantly in the 2x and 5x simulators, and Simulators A and B continued this increase by making street capacities a minimum of 1000, even for their Hard versions.  Until now, Simulator Z has had street capacities roughly comparable to Simulators A and B.

There's been a bit of a dilemma here in that we want to have streets have sufficient capacity to handle necessary traffic without getting unnecessarily congested, but at the same time we don't want to make them so attractive to traffic that they're used as major routes.  Some of the experiments I've recently performing and have documented in this thread point to a way where both of these goals can be accomplished without conflict.  With perfect pathfinding, such as used by Simulator Z, network speed becomes one of the biggest factors in determining which routes Sims will take; in many cases, speed becomes the determining factor.  Therefore, by maintaining an appropriate difference between the speeds of roads and streets, Sims will choose roads whenever it makes sense.  This being the case, it's possible to boost the capacity of streets without having a negative impact on the game.  Instead, doing so will simply lessen the congestion of streets, and greatly lessen the promptings of the advisor to upgrade them to roads.  However, simply increasing the capacity of streets will not in and of itself create more street traffic.

This raises the obvious question:  What's the proper capacity for streets?  What I keep running into as I tune the simulator is that ideally, capacity is proportional to speed.  Although as Lenny points out, the traffic simulator does not reflect reality in many ways, I have found that using this real world relationship does produce good results in the simulator.  So I've created new street capacities for Simulator Z by taking the road capacity and multiplying it by the ratio of street speed to road speed.  Additionally, I've made one change in street speeds; I've lowered the street speed for cars from 35 kph to 30 kph.  (Road speed is 50 kph.)  This was done for two reasons:  1) To bring it more in line with other street speeds compared to road speeds; and 2) to help ensure that the speeding premium, which may be as high as 30%, does not draw through traffic from adjoining roads.  As Sims have effectively unlimited commute time in Simulator Z, and we're trying to discourage them from using streets excessively anyway, this should have no negative effect on game performance.

Making street capacities proportional to road capacities makes a big difference, as there was very little relationship between them before.  The range of increases was from 50% for Simulator Z (Low) to 300% for Simulator Z (Ultra).  Simulator Z Classic was unchanged, as its purpose is to replicate the capacities of the original Maxis simulator.  But the changes in the other simulator versions should make it much easier to construct cities that use roads, avenues, and highways as main thoroughfares, and streets for everything else.  Usage of features such as SAM becomes much easier without running into problems from other parts of the game.  So please give this version of Simulator Z a try, and let me know what you think.  The attached package contains the same files as the Beta 1 version of Simulator Z v1.3, and includes the same changes to one-way roads.