First, let me preface this idea by saying: I don't think there is no place for micro in this game. However, I do feel that one of the constraining factors of the first game is that late-game coordinated strikes can be hell to organize and worse still to try and micro-manage for maximum efficiency.
The Admiralty
The Admiralty (or Fleet Command) is an upgrade for frigates or capital ships unlocked as part of fleet logistics upgrades, and is meant to allow the player to delegate fleet construction and tactical decisions to the AI, leaving the player to handle strategic decisions (or intervene tactically, if they so desire).
The Admiralty upgrade has two main functions: Automating new ship procurement, and allowing the player to determine what tactics their fleets use in combat. A ship which receives the Admiralty upgrade is a flagship (suggesting there can exist any number of flagships).
Procurement
After construction, the player specifies the type of fleet they want, the budget (as a percent of income or credit value per second), and the shipyards at the admiralty’s disposal, and the AI begins automatically building a fleet to match those specifications.
Available fleet types may vary based on race or research level, but could be expected to include main battle groups, carrier task forces, and planetary bombardment fleets. The fleet type itself would be defined in an external resource file as composition conditions and a sequence of build steps: These might take the form of explicit build commands (5 PsiHeavyCruiser), percentages that the game will attempt to bring the fleet in line with (i.e. .15 PsiLightFrigate), and technologies to be researched. The definition would also include “stopping points”, points at which a fleet can be considered complete. If the player has set procurement to auto-cast, then the admiralty will continue building as long as possible (until the file ends or there is no more fleet cap), but stopping points allow incremental upgrades on the player’s timeline and protect a player from unintentionally losing all their money building 200 Krosov Siege Frigates.
If the Admiralty needs to build or research something which costs more than its budget (say a fleet definition calls for a capital ship but the Admiralty only has a budget of 15 credits / second), then it will take as much as it is allowed to, stockpiling the funds until it can afford the item in question. If the player needs that money, they can reclaim it from the Admiralty with no further consequence than delaying whatever the Admiralty was saving up for.
One possible complexity is a player building multiple admiralties, specifically determining how they interact. The simplest answer is that they can easily coexist but do not normally interact: Each ship an admiralty builds is automatically assigned to a fleet corresponding to that admiralty, so if two admiralties are both given use of a shipyard there will not be a problem determining which admiralty commissioned any specific vessel.
Tactics
The ability of the admiralty to change a fleet’s tactical behavior is predicated on the idea that for each fleet there exists one controlling AI in the game’s code. Subsequently, a change in tactics corresponds with either the loading of a new AI or a change in the current AI’s parameters.
Fleet AIs would ideally be externalized like fleet definitions (though unlike the definitions, AIs would be code rather than configuration files). An example tactic is raiding:
1. The fleet should avoid enemy defenses and ships
2. The fleet should retreat if engaged by an enemy which poses a serious threat
3. The fleet should attack logistical structures in the following order:
a. Trade ports
b. Shipyards
c. Culture centers
d. Research centers
e. Others
4. The fleet should not attack the planet
5. The fleet should leave when all targets are destroyed or when an enemy fleet at least half its size enters the planet’s gravity well.
To take advantage of this behavior, a player would set the fleet’s tactics, then order the fleet to attack a planet. At that point, the player is free to focus their attentions elsewhere and let the AI conduct the attack (although micromanagement is obviously still a viable tactic).
Edited because copying from MS Word was a huge mistake.