9 Commits

Include renames

Author	SHA1	Message	Date
kono	ad07c1da8f	remove DinoSampling stuff	2020-04-05 13:10:13 +02:00
kono	a08b8160a3	add new results of needed timestamps in total	2020-04-04 17:14:12 +02:00
kono	328fc85214	modify q Learning to sample results and update R script	2020-03-28 12:35:33 +01:00
kono	ee1d62842d	split Antworld into episodic and continuous task ... - add new simple state for jumping dino, to see if convergence is guarenteed with with state representation - changed reward structure for ant game	2020-03-15 16:58:53 +01:00
kono	64355e0b93	add javadoc	2019-12-27 00:50:59 +01:00
kono	5a4e380faf	add dino jumping environment, deterministic/reproducable behaviour and save-and-load feature ... - add feature to save and load learning progress (Q-Table) and current episode count - episode end is now purely decided by environment instead of monte carlo algo capping it on 10 actions - using linkedHashMap on all locations to ensure deterministic behaviour - fixed major RNG issue to reproduce algorithmic behaviour - clearing rewardHistory, to only save the last 10k rewards - added google dino jump environment	2019-12-22 23:33:56 +01:00
kono	34e7e3fdd6	distinguish learning and episodic learning, enable fast-learning without drawing every step to reduce lag ... - repainting every step on no time delay will certainly freeze the app, so "fast-learning" will disable it, only refreshing current episode label - Added new abstract class "Episodic Learning". Maybe just use an interface instead?! Important because TD learning is not episodic, needs another way to represent the rewards received (maybe mean of last X rewards or sth) - Opening two JFrames, one with learning infos and one with environment	2019-12-21 00:23:09 +01:00
kono	c11cc2c3f2	add two simple scroll panes to represent environment and ant brain	2019-12-09 01:09:34 +01:00
kono	87f435c65a	add basic core structure and first parts of antGame implementation	2019-12-07 22:05:11 +01:00