Tag: honour project progress

Haven’t updated my progress in a while. Anyway, I just thought I’d consolidate my thoughts online so I can remember what I need to do for my final report.

For the results, I’ll need to modify my graph colour scheme to be document friendly or simply show the graphs with gnuplot. I might do that as it’s already optimised for LaTeX.

The experimental performances have to be chanegd to match the style in the competition so I can have my own mini-version comparison competition.

WEKA has to be discussed, with it’s applications of being a static comparison as well as an incremental help.

I should hit up the other competitors again for a summary of their strategies or something if I can.

Got a test now, so gotta go!

Well I decided that this should be digitised. Be nice and easy to check off.

Priority	Details	Complete?
1	Allow agent-environment interaction	Yes
1	Allow specifiable number of rows and columns (columns < rows)	Yes
1	Allow modifiable piece distribution probabilities	Yes
1	Allow modifiable speed	Yes
1	Allow modifiable reward exponent	Yes
1	Have on-the-fly agent loading	Pretty much
1/3	Allow random seed to be set for the environment/agent	Yes/No
2	Show agent details for each step (via agent_message)	Yes
2	Performance graph (with variable x axis)	Yes
2	Show game details (lines made, total reward, episodes)	Yes
2	Savable agent states (saved policy)	No
3	Savable environment settings (and loadable + preset loading)	Yes
3	Experiment-mode (Queue multiple fields/agents for later comparison)	Yes
3	Optional GUI field (for speed)	Yes
4	Allow infinite field height (- Show top of field; Show number of rows beneath; Show total height)	No
4	Options for 1-piece and 2-piece Tetris (Show next piece)	No
4/4	Prefilled field – Drawn in with option to remove lines or hole-ify lines/Random field at y-height	No/No
4	Ensemble playing (multiple agents, possibly borrowing parameters)	No
7	Player vs. Computer	No
8	Shadow mode for seeing where a piece will land	No
9	Player vs. Player (with random inserting of almost lines)	No
10	Puzzle Kombat style fighting	No

I think I’ve finished Tetris Workshop (or at least the first version). There are still known problems (can only load agents called SmartAgent in the SmartTetris package), but everything else works the way it needs to.

I can set up an experiment, run it overnight and such and have the results saved to be examined later, able to be examined in either the total performance or the episodic performance.

Now I just need to do that. Run some experiments and record the results.