Walt
Q1:
Cluster or Decoy is needed for PD interplay. If Cluster is chosen, Interceptor is needed to avoid PD being overwhelmed without recourse.
Actually, Interceptor should be a switch on the launcher, not a missile type. HE at least could target missiles and explode in proximity. Pseudocode for the missile should look something like "if a hostile missile has just left 3/4 of my explosion range, explode." Cluster could use similar logic to split and have the submunitions pick different targets. This would let ships bulk up their PD against missile heavy opponents without spamming ridiculous amounts of dedicated PD that leaves them with no resources to fight non-missile opponents.
EMP has been tried in mods and balance seems to be touchy. With significant AoE it's too powerful against shields. Without significant AoE it's practically useless.
Incendiary is probably safe, but it would rely on the swing around property of the current missile targeting since it'll only be good against exposed rear engines or as a minisher against already breached and inadequately compartmentalized victims. It would be useless against shrouded or internal engines, but boost thrusters are impractical to shroud.
Nuclear would cause issues. Making missiles too potent forces everyone to pay a PD tax.
Decoy, Torpedo, and SRM all fall out of my answer to Q2.
Drone is not on the list, but might be possible. A drone is slightly slower, runs collision avoidance on enemies as well as allies, and has a small, short ranged laser which it uses until it's lifetime runs out.
Q2:
Yes. I would consider the parts to be 1 payload, 1 guidance, and 2 fuel. They are each assumed to have 1 unit of mass and the missile is assumed to have 4 units of thrust.
A missile must have at least 1 fuel. This would be a short range non-explosive torpedo and not very useful. It would accelerate at four times the standard rate and burn for half as long. This is probably useless.
A missile with 1 fuel and 1 guidance would accelerate twice as fast as a standard missile and burn for half as long. It would do only minor impact damage. This is one possible version of the Decoy.
A missile with 2 fuel and 1 guidance would accelerate 33% faster than a standard missile with the same burn time and do only minor impact damage. This is the other possible version of the Decoy.
A missile with 2 fuel and 2 payload would be one version of the Torpedo. Acceleration and lifetime would be as the standard missile and damage would be doubled if inverse square falloff is used or remain the same with radius increased by 41%.
A missile with 1 fuel and 2 payload would accelerate 33% faster than a standard missile with half the burn time. This is another possible configuration of Torpedo.
A missile with 1 fuel and 1 payload would accelerate twice as fast as standard and hit as standard. This is a third possible Torpedo.
A missile with 1 fuel, 1 guidance, and 2 payload would be the SRM. It would have the same acceleration and half the lifetime of a standard missile.
EMP and Incendiary should use 1 payload. Drone is not viable if it uses more than 1 unless missile sizes larger than 4 are possible. Cluster could be built on the SRM or Torpedo variants with 2 payload.
If there are lots of small missiles I would make available light launchers that are smaller or mount more tubes shortways and can only accomodate the smaller missile types.
Q3:
I would give PD 1d2 damage to give missiles some variability in survival. Missiles with payloads should have 2 HP. Decoys with no payload should have 4 HP. Submunitions should have 1 HP unless the number of submunitions is small. If nukes are included against my advice they should still have 2 HP and have to be protected with decoys or submunitions.
Q4: Yes. Failure to do so would put short ranged or unguided missiles out of balance.
Q5: Missiles already effectively circle to the rear of the target before homing. The difficulty is to get them to not do so.
Q6:
A missile only has so much fuel so extending range makes no sense. Sensor arrays being required for homing would nullify a lot of existing ships, some of which could not be modified to include sensor arrays for lack of space. This would reduce the variety of early systems in bounty mode.
Each launcher having a limited number of fire control links makes perfect sense. Sensors offering more links also makes perfect sense. Control links should be shared across the whole ship rather than the sensor lifting the cap per launcher so a ship with very many launchers and factories may need more than one sensor to achieve its full rate of fire. Torpedoes not using cluster payloads wouldn't use control links.
Possibly launchers should even have fewer than 3 control links with command centers having some control links such that a control room with 2 launchers would have 6 total links.