The International Simutrans Forum

 

Author Topic: Sim City (2013)  (Read 26432 times)

0 Members and 1 Guest are viewing this topic.

Offline jamespetts gb

  • Simutrans-Extended project coordinator
  • Devotee
  • *
  • Posts: 18565
  • Cake baker
    • Bridgewater-Brunel
  • Languages: EN
Re: Sim City (2013)
« Reply #70 on: October 05, 2013, 01:29:36 PM »
An interesting update on this game here. They are "exploring" the possibility of an offline mode, but, of more interest to those with an eye to comparison of basic simulation dynamics, they have decided that it is not possible to support larger cities because the Glass Box engine is such that making them any larger will result in most users being unable to load, let alone play, the games.

This is interesting because it shows that current computer hardware is a long way away from what is necessary for agent based simulation on anything other than a very small scale to be workable. It was previously not entirely clear why cities were limited to 1x1km - it is now clear that the reason is the performance load of the Glass Box engine.

That is rather reassuring for developers of Simutrans, I think, using a more conventional and somewhat more abstract model, that our system is the right way of doing things for some time to come.

Offline Sarlock

  • Devotee
  • *
  • Posts: 1340
  • Languages: EN
Re: Sim City (2013)
« Reply #71 on: October 05, 2013, 05:22:09 PM »
There is a reason they have to use supercomputers to do large agent-based simulations.  If a 1kmx1km system is a serious tax on computing hardware, an exponential increase to even just 2x2 or 4x4 could bring the CPU to a halt.

With many systems, a basic simulation of batches of agents performs equally well with much less load on the CPU.  As long as you have the right models in place, it is extremely lifelike.

A couple more generations of computing power increases and Carl will be able to do a to-scale simulation of GB......

Offline asaphxiix

  • *
  • Posts: 723
Re: Sim City (2013)
« Reply #72 on: October 05, 2013, 07:34:37 PM »
I dream of one day playing simutrans on google maps.

So how does the simutrans engine cope so well with the task? Is it necessarily a trade-off with the graphics and way engineering features which ST excels less at, or would it be possible to implement the engine of simutrans with a more graphics and way engineering oriented type of commercial games?

Offline jamespetts gb

  • Simutrans-Extended project coordinator
  • Devotee
  • *
  • Posts: 18565
  • Cake baker
    • Bridgewater-Brunel
  • Languages: EN
Re: Sim City (2013)
« Reply #73 on: October 05, 2013, 07:38:09 PM »
The main difference, I think, is that we do not keep track of each individual person living in each individual house, but generate passengers and freight on the fly as and when necessary. Also, the Glass Box engine uses agents even for things, such as electricity and sewage, that are not even agent agent based in real life, which is silly.

Offline Sarlock

  • Devotee
  • *
  • Posts: 1340
  • Languages: EN
Re: Sim City (2013)
« Reply #74 on: October 05, 2013, 08:59:44 PM »
Indeed, you can easily overdo it with using agents for things that are easily simulated on a bulk level.  Take electricity demand for instance: the way Simutrans does it is fine, calculate load based on number of buildings in the city and as long as the city is provided this electricity to any place in the city grid, we consider the entire city supplied with electricity.  Does it need to be more complicated than that?  Absolutely not (unless we were building an electricity grid simulator!).

If we have 10 residential buildings with 50 citizens each (total population = 500) and 10 job sites on a map, we could create 500 agents and let them each figure out how to get to work.  Or, with a lot less computing power, we can just simulate that effect by using a few well written assumptions and calculations and come to basically the same result.

Offline An_dz

  • Web Admin
  • Administrator
  • *
  • Posts: 2841
  • D'oh
    • by An_dz
  • Languages: pt, en, it, (de)
Re: Sim City (2013)
« Reply #75 on: October 06, 2013, 05:03:32 AM »
SimCity is already cracked and it runs offline, there's a lot of pirates saying it worked ok. Anyway most said the game is bad.

There's a mod that lets you build outside the city limits too.

Offline IgorEliezer br

  • Devotee
  • Administrator
  • *
  • Posts: 4083
  • Cake recipes are cool... REALLY!
    • Igor Eliezer Architect and Urban Planner/Arquiteto e Urbanista
  • Languages: PT, EN, AutoLISP, Python
Re: Sim City (2013)
« Reply #76 on: February 07, 2014, 06:20:41 PM »
For those who like to watch SimCity letsplays: SimCity: Cities of Tomorrow

Offline Isaac.Eiland-Hall us

  • Benevolent Dictator
  • Administrator
  • *
  • Posts: 3649
  • PanamaCityPC.com/support/
    • Facebook Profile
  • Languages: EN
Re: Sim City (2013)
« Reply #77 on: February 09, 2014, 07:41:48 AM »
SimCity remains dead to me. :)

Offline vorlon

  • *
  • Posts: 62
  • Languages: FI, EN
Re: Sim City (2013)
« Reply #78 on: February 09, 2014, 03:56:28 PM »
SimCity remains dead to me. :)

I tend to agree, but I can't dismiss the irony in this line of thought: simcity and other such games that appeal to a broad audience were once considered a threat by the "hardcore"-gamers. And now, 10 years after Simcity 4 was released, the people who are worried about the death of serious gaming are the simcity players.  ;)

Offline Isaac.Eiland-Hall us

  • Benevolent Dictator
  • Administrator
  • *
  • Posts: 3649
  • PanamaCityPC.com/support/
    • Facebook Profile
  • Languages: EN
Re: Sim City (2013)
« Reply #79 on: February 10, 2014, 09:34:14 PM »
I think the advent of Minecraft proves that sandbox open-style games (like Simutrans! Hurrah!) are really a serious genre with serious players like all the other genres.

Offline jamespetts gb

  • Simutrans-Extended project coordinator
  • Devotee
  • *
  • Posts: 18565
  • Cake baker
    • Bridgewater-Brunel
  • Languages: EN
Re: Sim City (2013)
« Reply #80 on: February 10, 2014, 10:28:00 PM »
I think the advent of Minecraft proves that sandbox open-style games (like Simutrans! Hurrah!) are really a serious genre with serious players like all the other genres.

Indeed! I do love the idea of an ongoing, highly intricate and realistic enormous multi-player simulation of transport of all sorts of the sort that we are beginning to have with the Bridgewater-Brunel server games with Experimental. We just need to improve the economics to incentivise more realistic play, to model more relevant dynamics and to interlink various aspects of what we do simulate and we shall have a thing of beauty and joy forever.