[note: my keyboard is acting up, especially the 'T' key. If something makes no sense, mentally add a T to see if i makes sense.]
Nice concept, great potential. There are some major and minor issues...
1) Path finding will sometimes skip nearby items and go to a farther one, sometimes even going up/down levels. Likewise, I can tag a block to be mined that is next to an idle miner, and yet it gets assigned to a miner that is ... in a galaxy far, far away.
2) The "job" system is still ineffective. For example, I just had a game where service bot A was trying to service bot B... Which was trying to service bot C... which was trying o service bot D... which was trying to service bot A! They'd just keep going in circles chasing each other around the system of corridors and "rooms" I had (I was mining out in a grid-like fashion).
3) The particles are "glitchey", they would sometimes show up while viewing higher or lower levels. Likewise, I've seen bots suddenly appear in the middle of solid ground. However, they happily moved to the nearest corridor/machine and then went on their way.
4) Once here's a few material reactors, the material count sometimes goes negative. this does not seem to affect operations, so it may only be a display bug (or possibly an integer under-/over-flow).
Suggestions:
1) Don't show all territory; Give bots a "sensor range" to uncover what's where. This gives players an incentive to go mine out areas and explore them, rather than simply dig straight to the high value ores.
2) Make maintenance cost money as well--Keep the player moving and mining! After having a sizable workforce, machinery base, and all research is done, the game just becomes an infinite money tree. Give players an incentive to go make more money to pay for maintenance, and keep bots moving and working.
3) Give jobs/tasks a time limit. If a bot cannot complete it's assigned job before the time limit expires, abort the job (and most likely put i back in he job queue for another bot to attempt). Also, if a bot "calls" for maintenance, have it STOP and WAIT for maintenance to be complete. Don't make maint bots keep trying to hit a moving target.
4) Add "elevator up", not just down. As the game grows, this could allow things like harder materials that require special miner bots, but in the meantime, players could route corridors under/over the area towards easier to mine areas.
5) Add the ability to sell bots and machines which are in the way or no longer needed (or even simply misplaced). You could even add a "operational lifetime" counter to reduce the sell cost--Immediately selling a misplaced item would recover full cost, while selling a generator after an hour (real-time) of use might only get back 10% of it's cost (scrap metal).
6) It seems idle bots still consume power, but make sure all bots continuously consume power (and possibly water), and have regular maint cost material and/or money as well. Working/moving bots drain it all faster, of course. This is another way to prevent maint bots from running endlessly chasing others--the maint bot will eventually break down and need maint on itself.
7) Let maint bots try to self-maintain; As long as they can get to he right machines to do so, they can replenish themselves--again, this helps avoid circular chasing and if main costs money, gives players incentive to keep moving and mining.
8) Add doors/gates. Give some control over sections of the mine and prevent bots from entering them until I want them to.
Other ideas (long term goals?):
1) Planning mode; Pause the game and let the player "plan" where to mine, and what machines to place where (including possibly mining out space needed for the machine). This can include adding notes/markers ("Found high value ores here! Check around for veins of more!"), or possibly "idle bot storage" and/or "maint areas".
2) Ability to move machines; Constructors can disassemble a machine one place and reassemble it another.
3) Specialized miners; Normal miners can dig out dirt, but other ores (and stone) not only take longer to mine, but give a decreased yield when mined. Specialized ore miners are the opposite--they can mine out high value ores more efficiently, but decreased efficiency for others.
4) Runs/veins of high value ore; I already do this mentally--When I mine out a high value ore, I pretend I cannot see the other ores around it, so I have to mine out he surrounding blocks to "expose" the valuable ores for mining. (In effect, veins are not always continuous, but may "skip" one block, thus the need to mine out the surrounding blocks in order to expose possible veins for "chasing", much as I'd do it in "real life".)
5) Job/task management; Be able to look at the tasks/job, both global and per-bot/machine, in order to cancel/reorder/reassign them.
6) Percentage based improvement research; Instead of just a dozen or so static researches, have something where each percentage improvement costs increasingly more (like how a typical "incremental" game does it). Thus, research is only limited by the computer's numerical limits.
7) Path finding; Use graph theory and nodes/edges to pre-calculate many paths. For example, if I mine out a "room" area connected to the start area by a corridor, there is no need to constantly path find between them; the path exiting the room area to the start area will always be the same so you can pre-calculate and save it for re-use. Thus a bot only needs to the nearest edge to connect to the desired area, then use the saved path, and on arrival, find the path to the specific target location. You still have to calculate two paths--the start and end paths--but since these are usually very short and quick, they don't use as much CPU as trying to calculate a full path every time. You can determine the edges and nodes using traffic and destination tracking: if 14 bots route along points (10, 17) to (8, 4) every "tick", those are high traffic points and the route between them is a candidate for pre-calculation. You can even track other metadata--"Where is the nearest collector from point (10,17)? Ah-ha! Near point (8,4)!" and so on.
8) Add overlays to show distributor range, and possibly collection zones for collectors. Maybe even service ranges of generators and pumps, and so on.
9) The game suggests you need to pay off your initial debt... Maybe keep track of that debt (plus accumulating interest), so players can pay it off... Or else! ;-)
After playing another game, here's some more notes and observations:
1) The z-order stacking seems to confuse the path finder even more, especially when the same machine is built at the same (x,y), but a different z.
2) It seems that as long as I tag a block to be mined on each level that *cannot* be reached at all, miner bots have less (but not entirely eliminated!) problems when it comes to finding jobs to do at different levels.
3) Possibly related, logistics bots in one game had a problem with elevators and collectors; In the game, I had collectors on each level in the same place, but the logistics bots would not use them--they tried to go up the elevators to the "home machine" to use. Once up top, however, they seemed to change their mind and go back down to one of the lower level collectors. But they'd just change their mind again and go back up top. They'd just bounce back and forth like that in an infinite loop.
4) I encountered the bug someone else mentioned where I'd place one machine, but it would build a different machine. In my case, it would sometimes build collectors in places I placed to build distributors.