Growing nearer to the agent choosing greedily. Estimated time that all tests are passing and agent (should) be working (well or not, I don’t yet know): Tuesday 5th February.
A point of interest arose when writing the findPieceRewards function (a function which finds the reward associated with each position of the current piece to be used [...]

Hell of a frustrator today. Went to the lab to collect the rotation print outs to find out how the rotation behaves. After much humming and harring, I resolved to create a paper Tetris model.
Once I had cut out the little pieces, I placed them at their coordinates as given by the tests and spun [...]

Just a progress log.
Filled in the contourScan method. Tested and working well. Could do with some more tests though…
Added a MathUtils class for various mathematical functions such as rotating a point and outward rounding (unused at the moment).
And created (and tested) a method for finding the index of a certain point in the observation.
6/13 tests [...]

This game seems like a fun little project to make in my spare time.
I’ve come up with an idea for the name: ‘Tectonic’. To do with the moving plates and such.
That’s about all I got for new news on it…

Abstracted the agent_step function into two functions: storeReward and takeAction. These are also used by agent_end and agent_start respectively. The tests have been written for both of these too.
Problem: Piece tracking
When chasing the goal, actions will be assigned to the piece that may collide with other blocks in the field. There is no easy way [...]