-
Notifications
You must be signed in to change notification settings - Fork 1
/
Diary\Plan
17 lines (17 loc) · 1.58 KB
/
Diary\Plan
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
25.01.2021
creation date of the diary
current status of project: all but one benchmark is solvable, only straight tunnel continues to elude us
further, modifications(removal of col dept N in wait, frd, dodge and nmc) of the code (21.01.2021) vastly improved runtimes on all but 2 benchmarks
next objective: expand into the A domain
-create new benchmarks/modify old ones to include packing stations
add more robots
-new benches, modify code to adapt the wait flag
quick summary of events so far: managed to merge plans without collision, struggled to merge plans with collisions (reason: missing variable for tracking col depth as
unique identifier), added said modifier after Christmas holidays, added proper dodge commands, modified dodge commands for edge cases (robot finishes in regular path,
robot finishes in path of a dodge return), modified encoding to remove col depth from wait(), frd(), dodge() and nmc() and started taking calculation times
(04.02.2021) finished a plan creater for the A-Domain (without quantities)
adjusted bench 1,4,6,7 to include packing stations.
bench 6 will result in an interessting problem during the merger, cause robot with shelf goes over shelf without robot
bench 7 will result in robot needing to dodge, who has no spaces to move to
(11.02.2021) solved problem for bench 7, started no shelf collision, between one robot, thats carrying a shelf and a standing shelf.
(17.02.2021)nearly finished shelf collision, but having a small timelag for bench 6, it's one timestep thats not getting used, still under investigation