Open YassinAbdelrahman opened 4 years ago
Each cell is also a goal state, they can hand the building attached to them in any of them. Perhaps we could add another thing to nodes that is something like
center_of_goal_states
. For what exactly do you need the center?
This is all probably just for the attacker that wants to block enemies from reaching goal states:
For the action of blocking enemies we need the goal state the enemy uses the most, so the goal state with the most enemies around it. If we do not calculate that from the centre of a group of goal states but for every single goal state block in a 5x5 area I think it would cost too much time (since there are probably 13 goal blocks in 1 group of goal blocks).
But on the other hand, calculating the most crowded goal state for every single one of them will improve the position of our attacker agent, since it will stand in the best possible goal state with the most enemies around it.
When a goal state consists of more than one block (see picture) the graph sees this as multiple goal states. There needs to be a function to find the center of this block of goal states.