Target aquiring

i can''t give you any efficiency tips, but your idea sounds good to me. be sure to adjust the priorities for if a ship is damaged, or attacking an important target, though. what i mean is, there is no reason to leave off attacking a nearly destroyed ship to get a new target, when you could kill that ship in another few seconds and be rid of an enemy; or when the current target is attacking the mothership, it should remain high priority even if a transport comes by.

There are two questions that need to be answered here:

1) How will an NPC ship choose another ship as its target; and,
2) How will you implement this.

The latter depends very much on the former.

Ask yourself what you would do given a certain scenario. One technique is merely to encode the rules that you use to make a decision directly into the computer (often using a script system with a finite state machine). Other techniques might consider complex relationships between ships and evaluate them mathematically to determine one of possibly many ''current scenarios'', with each scenario assigned a response action.

Personally, I would think that a fuzzy state machine would give you a reasonable evaluation and state transition framework. There''s plenty of info on the web on state machines and fuzzy logic.

Good luck,

Timkin

I''m going for quickness and simplicity. I plan to have many ships active at one time, so they must be able to decide very rapidly.

I faced a similar situation some time ago. Basically, what I did was, for each ship, go through all the ships (or just the ships in the immediate area) and give each one (that is not on the same team as it) a value based around a few variables:

Distance.
Angle difference between our ship’s current direction and the direction between our ship and the ship in consideration.
Health.
Priority.

And more, depending on how smart you want your AI to be in this decision. Then chose the ship with the highest (or lowest, depending on how you code it) value as your target.

Now each of these variables needs to be scaled appropriately to fit into proportion with the rest, giving different variables more importance than others.

Alternatively, you might like to simply select the closest ship as the target.

Then, on the other hand, you may wish to go over the top and weigh the entire situation out, choosing targets that will permit you a strategic advantage. For instance, you may like to take into account the position of all friendly ships and the position of all enemy ships, and then chose a target that will allow you to stay close to your allies, while avoiding targets that are bunched in large groups of enemies.
Another neat trick would be to take a look at what each of our allied ships are targeting, and then make a decision based upon that. For instance, you might like to avoid targets that are already being dealt with in order to balance out the battle, or you might like to focus on such targets to promote group-killing tactics.
You could look at what each of the enemy ships are targeting and the move to help those allied ships under the largest threat.
Here is a neat one: you could have your ships jump on the opportunity to intercept and cut of an enemy craft that is being tailed by an ally. This could be done quite simply by taking the angle difference between the direction from the ally chasing and our ship and the direction from the ally chasing to the enemy ship, then adding it too the angle difference between the direction the enemy is flying and the direction from the enemy ship to us, then checking the result to see if it is below, say, 60, and if so we know that we can intercept this ship nicely and therefore make it our target. That is assuming my on-the-spot maths shills are correct. If you were to use a formula such as this you''d want to tie it in with the standard target selection routines etc.

But, then again, I am no pro at this sort of thing so don''t read this thinking that this is standard stuff. It is merely my suggestions - what I might do in a situation like such.