Workers not getting to work has no consequences!

[Molano]

I can't promise anything, we are currently very busy with the patch, but new mods made by us are not impossible There's one super hard mode mod on the workshop, has anyone tried that out? I haven't had the time yet, but making mods based on the pre-made mods is not too hard, so you could try some tweaks and see if the game is then more to your liking.

All in due time, the game is awesome already and you guys did a great job on it! Thanks for awnsering! I'm trying out some mods now .

 

[EvilTom]

I might have phrased my last post badly.

Could an option be added to allow players with more powerful computers to increase the amount of cims being simulated over the 65k limit?

I'm wondering if that even if your computer could handle all 1 million sims on the road at once the timescales involved would still clog up the roads. It's not just about performance, it's about the game balance.

Plus... what size if your largest city? Has anyone seen anything about 300k? Or even just anyone hit the 1 million yet?

 

[JimmiG]

At the very least there should be a quick check to verify that there is an unbroken path from workers homes to their place of work, no matter how congested and convoluted. That would make silly cities with completely disconnected neighborhoods impossible.

 

[Brazilian Joe]

How was the 65K limit reached? Is it because of raw CPU power, memory usage/bandwidth, both, or something else?

Can you share some hard numbers with us?

Like, if you have two pathfinding threads, or better, one per logical core (including HT cores) could you double or quadruple the 65K travelers?

Do you already employ a staggered processing of travellers, e.g:

a) Assuming the simulation runs at 100Hz, you'd want each agent to be able to change its mind only 10% of the time
b) Therefore, the population is sliced in 10 groups, and each group is processed in a round-robin fashion
c) Pathfinding is a bit different, so the slicing may be done according to distance until next intersection for possible path changes.

If this kind of slicing/staggering of calculations is done, and the pathfinding is improved with multiple worker threads, could you raise the limit?

Also: Any workplace (commerce/office/industry) with less than 50% worker positions should change status to "not running: too few employees" and lose money/stock on that day because of it. Employees would lose happyness.
Also: Any agent which is unable to reach its destination and is teleported back to its origin should lose happiness because of it. If I try to purchase the latest video game COGS (Colossal Order Game System) but get stuck in traffic to the point of being teleported back home, I should become sad/angry about it. Same if I cannot get to work.

 

[Azirphaeli]

Alright CO, how about this as a compromise:

If an agent is assigned the "go to work" pathfinding and that fails (he can't get to work within a certain time period), then the job becomes abandoned (the industry losing that employee). This would cause the employee unhappiness and the job location unhappiness. Next, the civ will try to find a job at a different place, and then a new civ from that old place of employment will try to get to work before it starts hiring again.. (this would be like a check for the player, as perhaps an icon can float over this building indicating a potential infrastructure problem. "a car with a slash through it"?)

Agents not currently simulated with pathfinding and trying to get to work will be unaffected. This way, if there's a problem, it has an effect. It's not immediately catastrophic, but since that effected job will keep calling agents into to work, if there's no roads (exploiting) or just major traffic issues (it takes too long) there's unhappiness, the building spawns more traffic, and the player is alerted of the problem. If he leaves it alone for too long, the building will lose it's workers and be abandoned.

Just a suggestion on a reasonable way to deal with this, without destroying the gameplay but without leaving potential exploits and making employment more meaningful.

Super Edit: This will also act as a safeguard for distance. If agents have jobs that are too far away from their home for them to realistically get to, they will lose that job and switch to a close company if a job is available. This will help optimize pathfinding in a way, as agents will be gravitating towards closer jobs.

 

[Raph]

My problem is still that with population of around 130k only max. the half of people can be outside. Above 650k less then 10%. It seems to be unrealistic.
Assuming, that a city with so big population is also very big, the streets also become dead. So we get a very large metropolis with sparse or empty streets.

Realistically, wouldn't most people be inside most of the time (working, sleeping, most of free time spent at home for most people)? On average I spend maybe 2 hours commuting, and perhaps one or two hours doing other things outside of home (probably less). Four hours mean 1/6th of every 24h outside/in transit. 65k people * 6 is 390k people, which is when, if CIMs are like me, you'd first start seeing thinning out on the streets. Seems pretty OK to me, and even at 1M you'd have a decent amount (40% density compared to at 390K) of people outside.

How was the 65K limit reached? Is it because of raw CPU power, memory usage/bandwidth, both, or something else?

A guess would be that the limit is actually 65 536 agents, 2^16, since computers are binary and all.

 

[magitsu]

I don't believe anybody would ever do that, unless they had a very small city. It would be just too time-consuming.

There's also already a very popular mod that does auto demolishing of abandoned and burned down buildings.
But that would need to be killed fast since I too think that it is detrimental view to approach the issue of vacancy. High enough price for bulldozing buildings should drop that fast.

 

[trigorin]

My problem is still that with population of around 130k only max. the half of people can be outside. Above 650k less then 10%. It seems to be unrealistic.
Assuming, that a city with so big population is also very big, the streets also become dead. So we get a very large metropolis with sparse or empty streets.

Ok, I've made a quick calculation. If you unlock all the nine map slots, you get a max. population density of 1806/km2 which means 7222/slot. Maybe not so bad
But I still find the changing of relation to population unrealistic.

 

[Inge Jones]

Our concern is that 65k of citizens will not be enough to supply active workers to each industry buildings within reasonable intervals.

It's the "within reasonable intervals" that is open to negotiation. The problem is that day counter bar, which is so near to meaningless that you might as well scrap it and extend the scope of "reasonable intervals" to fit what is realistic to animate. So think of the day as being longer, in order to give more chance of the entire workforce (those who can reach a workplace) to actually reach that workplace. Then those who couldn't or didn't, count as absenteeism and treat accordingly. I am not sure if the Cims are triggered to start a journey at random, or whether there is an iteration through them all, so they all get a fair turn at starting a journey. If the latter, then count a day as being a complete journey-trigger cycle, and wait until the end of that journey trigger cycle to assess which workers turned up and which didn't - and calculate employment and industry effect from there.

 

[ROMMEL_HSQ]

Any chance to make that 65k limit modable?

 

[bucgene]

Our concern is that 65k of citizens will not be enough to supply active workers to each industry buildings within reasonable intervals. We have worked a lot on this feature and there sure is some room for improvement, but quite a many things have been tried and thought of. But keep the suggestions coming, we are very open to tweaking features, but just want to make sure the game will not break

Why not set a 'staffed status' timer to the building, when a worker arrive it reset the timer for all the industry buildings around the workplace, not just 1 building.

In this way. You don't need to worry not enough agent to reach the building if city gets too big. And also solve the 'building don't need worker' issue.

It is not a 'perfect' solution, but I guess its a good compromise.

 

[Brazilian Joe]

Workers should have a "3 strikes" approach, or maybe two strikes (1 bit).
A worker not going to its job gets one warning. If it fails to get to work twice in a row it's fired. Like others said, getting the worker fired may help making shorter trips.

 

[Epicity]

At the very least there should be a quick check to verify that there is an unbroken path from workers homes to their place of work, no matter how congested and convoluted. That would make silly cities with completely disconnected neighborhoods impossible.

They can't do that. It'll mean all workers will have to check all industries paths and that's too heavy a load for computers.

The best CO can do is allow modding of the cap and workers to act like resources.

 

[Epicity]

Bulldozing abandoned buildings should have a highish price. But industries moving into abandoned should take time so it balances out.

 

[Old Joe]

i assume this 65000 number is written somewhere in the code, and should be editable. at least to tweak between something like number of agents on streets, effective number of workers getting to workplaces cap, and cpu performance

 

[Alfa55]

@CO: If a cim has the status 'confused'. Is it an easy fix to link that to the workplace (via the cim) so that it can negativity effect the workplace? (eg. closure if more than 3 workers are 'confused' and stay home)

 

[Inge Jones]

i assume this 65000 number is written somewhere in the code, and should be editable. at least to tweak between something like number of agents on streets, effective number of workers getting to workplaces cap, and cpu performance

It's somehow beside the point. There is something about using the Cims' movements for mere illustrative purposes that is offputting to me. I am willing to make a huge compromise on timeframe but I cannot come to terms with a Cim keeping their job and being productive for the company when they don't even get to the place *ever*.

 

[heresandypandy]

Our concern is that 65k of citizens will not be enough to supply active workers to each industry buildings within reasonable intervals

You don't need the game to actually simulate the arrival of citizens to fix the main problem being discussed in the thread, though. All you need to do is make industries fail to generate or operate if there isn't a transport connection between I & R... At least that way you're actually maintaining the illusion.

 

[darkelf1000]

It's somehow beside the point. There is something about using the Cims' movements for mere illustrative purposes that is offputting to me. I am willing to make a huge compromise on timeframe but I cannot come to terms with a Cim keeping their job and being productive for the company when they don't even get to the place *ever*.

+1
i want it to be up to me to make sure my roads are good enough that people can and MUST get to work or else.

 

[Alfa55]

It's somehow beside the point. There is something about using the Cims' movements for mere illustrative purposes that is offputting to me. I am willing to make a huge compromise on timeframe but I cannot come to terms with a Cim keeping their job and being productive for the company when they don't even get to the place *ever*.

What you want is not realistic. The 65k limit is not for fun but based on technical limitations. The cim's are in your city. Just not on the roads all the time. And also not at the workplace all the time because of the 65k-thing. It's more realistic then to use statistical input (number of residences, level of eduction, ..) for that sort of thing instead of a agent-based simulation. And that's exactly what they have done.