Author Topic: Operations Research in Simutrans  (Read 1396 times)

0 Members and 1 Guest are viewing this topic.

Offline zhaop

Operations Research in Simutrans
« on: December 02, 2014, 12:37:51 AM »
Hi all!

Long time I haven't played. Just finished courses on Optimization and Operations management (today), and I got quite excited about some of the topics covered, like vehicle routing, minimum spanning tree, minimum cost flow network, etc.

How could all these models help us play Simutrans in a more optimal way? What tools could be made that help us make decisions on how we build better networks and schedule things better and churn out more profit?

As a first problem, I'm trying to see if there's a way to apply network flow models to find the most optimal shape of a passenger transport network for a given map with given towns. I'd be curious to see in what cases hub-based networks would be optimal than "decentralized" networks (for example).

And loads of other things that could be modeled and optimized... :)

(Edited for clarity)
« Last Edit: December 03, 2014, 06:04:59 PM by zhaopatrick »
I'm working on an AI for Simutrans that can make lots of money, and I'm currently learning the squirrel API for AI players.
https://github.com/zhaop/simutrans/tree/marmot

Offline Banksie_82

Re: Operations Research in Simutrans
« Reply #1 on: December 05, 2014, 04:12:36 AM »
I suspect most of what you have raised goes over the head of the average Simutrans player, even most of the game developers. I touched on logistics very briefly at university, but even I wouldn’t be able to contribute much to the conversation nowadays.

I’d say that Simutrans – Standard would be too simplistic to apply real world theoretical models and expect to notice a significant difference in results.

Also, I think Experimental, while far more ‘real world’ then Standard, would also be lacking in critical elements of real world environments, particularly where some of these models operate.

In saying that however, hub based v’s decentralised is a topic of discussion that pops up from time to time on this forum. Obviously though, it’s with respect to the limits and parameters of the game.

I would be very interested to hear of any experiments you run and the results you obtain.