How Civ V will handle workers ?

[gunter]

Maybe it has already been discussed but I would like to know if you have any news on how workers will be handled. I am wondering if it will be still necessary to have dozens of them here and there and if the auto-improve feature has been improved or they are still supposed to be quite " ignorant " on their improvement auto-jobs. ( for instance auto build roads when roads have now a price to maintain them )

The last but not the last I have always wished an option to poison and/or gas them all via a kind of virus weapon. Could you please confirm me that no such option will be available ?

 

[VikingMaekel]

The last but not the last I have always wished an option to poison and/or gas them all via a kind of virus weapon. Could you please confirm me that no such option will be available ?

What? Why would you want that option?

 

[gunter]

What? Why would you want that option?

That's only a game ...... therefore I would appreciate a lot seeing them turning green due to my secret bio weapon. Simply that.

Don't believe a nuclear explosion is more gentle anyway........

 

[VikingMaekel]

I'm not comparing to RL, I just don't see the need for a virus weapon that only targets workers.
I fail to see the need for virus/bio/chemical WMD's that do target military units as well. A nuke is more than enough to portray the power of WMD's.

 

[gunter]

That would be not only worker intended of course but they are always " on the field " so they are an exposed target and moreover with no defense. An easy target for bio arrows.

Anyway I am more worried about the automation of workers cause if you order them " build road a go go " I assume they will do that without thinking about the relative cost.....

What do you think ?

 

[craig123]

Correct me if I'm wrong, but won't automated workers be similar to how the AI uses workers (i.e. the same algorithm)? If they just spammed roads then that would cripple the AI.

 

[gunter]

Correct me if I'm wrong, but won't automated workers be similar to how the AI uses workers (i.e. the same algorithm)? If they just spammed roads then that would cripple the AI.

Somehow yes, how can the AI know how many roads are necessary without wasting money for non needed roads ?

If the AI can't do that properly , do we have to micromanage all the workers' stuff ??

 

[Drawmeus]

Correct me if I'm wrong, but won't automated workers be similar to how the AI uses workers (i.e. the same algorithm)? If they just spammed roads then that would cripple the AI.

I would think so, though it might be a modification of that algorithm. Either way, I can't imagine they'd release without an AI that can handle the new roads design or release with worker automation that always screws you over.

 

[gunter]

Automated workers have always been pathetic into previous civs but now roads must be maintained like railways etc... so workers can drain your funds within a blink of an eye building crap here and there.

I'm so scared about that now I have to quit my job and go home immediately

 

[Ahriman]

I would be amazed if automated workers would build roads anywhere except linking cities.

Somehow yes, how can the AI know how many roads are necessary without wasting money for non needed roads ?

Easily. Calculate the shortest route needed to connect a city to the trade network, build a road there. Then do this again. Then again, as long as there are any cities unconnected.
Don't build roads on any other tile.

 

[Zhahz]

Since your population is going to grow slower (I think), you'll have fewer but sexier cities (in general), tiles your cities can work will unroll slower (as you expand via culture or purchase, in general), maintenance will actually mean something (less unit spam of all types), only 1 worker per tile (no more stacking workers to speed improvements either), etc - due to all of the gameplay improvement/changes, you will likely have a smaller number of workers.

 

[SeismoGraf]

Easily. Calculate the shortest route needed to connect a city to the trade network, build a road there. Then do this again. Then again, as long as there are any cities unconnected.
Don't build roads on any other tile.

This is one good and straightforward solution for the AI. Finding the solution with the minimum amount of roads for the whole empire and adapting the existing road network to this optimum, each time a new city is added, would be more difficult and probably not worth the effort (workers busy with optimizing road network; CPU/programmer/player time). Judging where to put redundant roads to increase civilian/troop mobility and make the trade network more resilient in case of war, this will be the real art.

 

[Ahriman]

Judging where to put redundant roads to increase civilian/troop mobility and make the trade network more resilient in case of war, this will be the real art.

My guess is that the AI will not do this. Which is probably fine. This will be hard for even a human player to judge, because it requires planning, which AIs are not good at.

Similarly, it will not pre-emptively start building a road towards where it plans to found a city, so that the road is nearly done by the time you actually build it. But again, no big loss. These are the kind of things that make the human more efficient.

 

[Giant Dwarf]

I would be amazed if automated workers would build roads anywhere except linking cities.


Easily. Calculate the shortest route needed to connect a city to the trade network, build a road there. Then do this again. Then again, as long as there are any cities unconnected.
Don't build roads on any other tile.

don't forget that you can also use a harbor to connect a city to the capital. You save $ on road maintenance at the expense of movement. I'm hopeful the AI is programmed to do this.

 

[SeismoGraf]

My guess is that the AI will not do this. Which is probably fine. This will be hard for even a human player to judge, because it requires planning, which AIs are not good at.

Similarly, it will not pre-emptively start building a road towards where it plans to found a city, so that the road is nearly done by the time you actually build it. But again, no big loss. These are the kind of things that make the human more efficient.

It would be surprising if the AI did, indeed. Very much "fun AI", e.g. by pillaging one road tile, cutting several cities from trade network and disrupt their only paved reinforcement path; surround and besiege their freshly founded city; attract and divert AI troops to a remote locations etc.

When 2 cities are already connected via a river (is this still true?) or port to the trade network, roads become a pure mobility consideration. I wonder how the AI will handle that.

Sometimes I see an idealized picture of an empire with 7 cities, the capital in the center, 6 cities around it in a star-like arrangement. From the capital 6 shortest-path roads/RR to each city, then each of these 6 connected via a roughly circular, peripherical highway. Hmm. CivFanatics sometimes have weird dreams...

edit: @Giant Dwarf: "programmed to do this" like in save $ or not deciding at the expense of movement?

 

[gruther4]

Easily. Calculate the shortest route needed to connect a city to the trade network, build a road there. Then do this again. Then again, as long as there are any cities unconnected. Don't build roads on any other tile.

This is a very logical and intuitive solution, but it would not result in optimal road placement. It would get very close most of the time, though.

I understand that I will probably need to defend that statement... I'll do it with an example. Imagine 6 cities arranged in a hexagon. The optimal solution to connect them all would be a ring with one side missing, for a total distance of 5x[distance]. However, if you start with one city, and then add the city directly opposite to it, the shortest distance between the "new" city and the network will be through the middle. If you put a path through the middle, the minimum total distance becomes 6x[distance]. If you changed the order of "selecting" cities, it might find the optimal solution, but it really depends on what order you add them.

I'm sure a simpler example to illustrate this can be found, but that is what came to mind.

A truly optimal road solution algorithm for a given set of cities (or planned cities) would be an interesting problem... does anyone care to figure it out?

 

[Ahriman]

but it would not result in optimal road placement. It would get very close most of the time

I agree. But I think it is simple and good enough. I doubt they're going to devise a complex optimal road building algorithm. Though they might.

The problem really comes when you have a lot of ties.

 

[SeismoGraf]

A truly optimal road solution algorithm for a given set of cities (or planned cities) would be an interesting problem... does anyone care to figure it out?

I think this would be a really tough problem, harder than for example the well-known travelling salesman problem. A merger of locally optimal solutions for clusters of cities would usually not give an optimal solution for all cities. Each tile with a road on it can be thought as a separate node. Certainly good approximations exist with reasonable run times. Do you have a special interest for the topic?

 

[SeismoGraf]

Why sure, I have this P=NP solution to the travelling salesman problem on my desk in front of me, would you like me to PM it to you?

Who would say no? Either get the 1000k Millenium Prize and the fame or do something else with it. Although I recently read that there is a promising proof candidate from an Indian mathematician indicating the opposite of your claim...

 

[andrewlt]

Why sure, I have this P=NP solution to the travelling salesman problem on my desk in front of me, would you like me to PM it to you?

Let's wait while they go through that paper to see if the proof is real.