Remoting: Rework Manager Targeting

[rougeayrn]

User Story:

As a Remote Manager, I want to abandon remotes that aren't making a profit, either due to harassment or just because the room sucks.

Acceptance Criteria:

• [ ] Write 'goal' function that determines and modifies the remote quantity goal
  • [ ] Consider CPU Availability
  • [ ] Ensure CPU Estimation is as accurate as possible (standard deviation?)
  • [ ] Consider Spawn Availability (Spawntime available compared to Avg. Spawntime Required)
• [ ] Write 'potentialRemotes' function
  • [ ] Identify Nearby Rooms List
    • [ ] Within Range Limit 4? Maybe 5?
    • [ ] Path-ability Distance, Just Terrain, < 375 Round Trip
  • [ ] Filter Rooms... (Maintaining both cast out and kept)
    • [ ] Not Path-able? Reconsider in 10k.
    • [ ] Not Safe? 10k.
    • [ ] Used by Ally? 10k.
    • [ ] Used by Self? 10k.
    • [ ] Prior Failed Remote? 50k.
    • [ ] Prior Run Off Remote? 50k.
  • [ ] Analyze Potential Profitability
    • [ ] + Amount Generatable
    • [ ] - Spawn Costs
    • [ ] - Structure Costs
    • [ ] - Energy on floor Loss
    • [ ] - Energy not transported or used
  • [ ] Order by Profitability
• [ ] Write 'activeRemotes' to...
  • [ ] Update actual profitability
  • [ ] Check for Fulfillment
  • [ ] Check for Harassment
  • [ ] Add remotes up to goal
  • [ ] Replace failing remotes
  • [ ] Replace remotes where actual is below the next best potential
• [ ] Data Structure in claimed room's memory
  • [ ] remoteSites: { [roomName]: stuff }
  • [ ] remoteConditions: { [condition]: info }
  • [ ] potRemoteSites: [roomName, potProf ][ ]
  • [ ] rejectedRemoteSites: [roomName, tickToReconsider ][ ]

OLD Acceptance Criteria:

• [x] Rebuild allowedNumberOfRemotes
  • [x] Rename to goal, because it might not be met and its a shorter name.
  • [x] Build Basic spawn time consideration
  • [x] Build Basic CPU consideration
  • [x] Raise and lower remote limit based on CPU and spawn time.
  • [x] Add limiters to keep the system within limits.

    ---

• [x] Build a cached structure for Remote Manager to store calculated potential profitability.
• [x] Calculate potential remote profitability.
  • [x] + Max energy generated
  • [x] - Container Maintenance cost
  • [x] - Creep cost
• [x] Calculate actual remote profitability.
  • [x] + Max energy generated
  • [x] - Container Maintenance cost
  • [x] - Creep Cost
  • [x] - Energy on floor
• [ ] Handle remembering what remotes have been abandoned so we don't repeatedly try the same failing remote.
• [ ] Modify calculations for easy expandability as more creeps or features are added.

  ---

• [ ] Choose remotes based on profitability.
• [x] Allow Sector center room to be chosen, but not non-center SK rooms.
• [ ] Prioritize based on profitability (potential when first chosen, actual past that)
• [ ] Abandon unprofitable remotes.
• [ ] Modify remote creeps to retarget if current target is invalid.

  ---

• [ ] Modify CPU consideration to include standard deviation consideration.
• [ ] Improve RemoteManager.goal() to actually calculate average expense for spawntime and cpu.

  ---

• [x] Add Game_Extended to the global scope given we can't prototype Game.
• [x] Add Game_Extended.ownedRooms.
• [x] Add Typeguards.isCenterRoom.
• [x] Add Utility.getRoomNamesInRange.
• [x] Add getBody() to CreepRole.
• [x] Add creep.bodyArray.

[rougeayrn]

Consider a precursor heuristic.

[rougeayrn]

What if we did a precursor that considers the important parts and didn't do a complete calc of profitability? Think about it.