Before the alpha released I had fixed some unit AI so the units starts attacking when they are idling. Also fixed so when a unit is under attacking, all “friends” that’s close to the unit also starts attacking.
I have fixed so you split up the pathfinding search over many frames and made some optimizations. I have also fixed the pathfinding problem that was in the released alpha.
//Markus
Have had much to do the recent weeks, making the AI a fun opponent to play against. There have been lots of tweaking and tuning variables to make it behave in a good way.
So today, for the first time I let the AI play and finish the game. I wanted to check all the behaviours it showed during a match, how it prioritized different goals and how it ended a game. Of course it made a couple of strange decisions and behaved very weird at times, but it managed to play the whole game which was very satisfying to see.
The main things to fix are to make the agents complete the goal they had before they for example encountered an enemy they had to dispatch. It works sometimes, but not every time. Sometimes they tend to stand still until something triggers them.
Also their behaviour when being very much stronger than the rest of the opponents should be much more aggressive.
Oh, and I’m happy to announce that the yellow team won :)
/Bo
Haven’t got a movie this time, but I have worked with path smoothing for the units and trying the fix the collision problem. Have also fixed some minor pathfinding bugs. Hopefully there will be a movie next time.
//Markus
All components of the defensive behaviour has now been implemented. It’s not perfect yet and need lots of testing, but I will have time to deal with that later.
One quite big thing left is to make a proper priority system for zones under attack. How many units should be deployed to stop enemies taking over one of the zones one team has? Wich of the zones have the highest strategic value?
I’m thinking of using the already implemented priomap that the team uses for take-over-prios to determine the strategic value. By doing it this way the strategic prio will change with the teams alignment. Also the prio will affect the size of the counter-team trying to prevent an hostile-take-over. This counter-team will consist of a part of the aggressive units. A part determined by the main-teams defensiveness.
I’ve also planned to implement all unit-upgrades into the AI before tuesday next week.
I’ve got three upgrade priorities done, but there are much more to implement. But it’s quite easy work, so I shouldn’t fall behind schedule.
//Bo
This week I’ll be working with the defensive behaviour of the team AI. I’ve made it so that it can fortify zones that are valuable and prioritized by the team. The zones first to be fortified are those which are at the edges of the teams’ current prio-map, and eventually maybe the money-zones too. A simple defensive behaviour for the individual troops is also up and running.
I’ll also program some kind of budget plan for the teams, so that they put aside some of their money for upgrades and some for fortifications etc. This budget will of course depend on what alignment the team has at the moment.
Next thing after those things is to create a behaviour for the team if one or more of its zones is being taken. The reaction should be affected on how defensive or aggressive the team is at that time, and what priority that zone has for the team.
Also earlier this week I tweaked the fuzzy-logic so that it works much better. I’ll probably have to change it again some time in the future. But I’m happy with how it works right now.
//Bo