Progress for week 11 (2019)

Fra Robin

Revisjon per 14. mar 2019 kl. 09:02 av Mwaandah (Diskusjon | bidrag)
(diff) ← Eldre versjon | Nåværende versjon (diff) | Nyere versjon → (diff)
Gå til: navigasjon, søk

Innhold

Malin Aandahl

Budget

  • 3. fitness function:
    • Must find a way so it is updated before every time the fitness function is run. Either try to use edit protobuf things or find another way?
  • Run experiments for all the environments(at least for the first two fitness functions). Can run the config-files in exp2?

Example:

      world_file=exp2/world3.world
      restore_directory=world3
      output_directory=output
      enable_light_sensor=0
      enable_touch_sensor=0
      population_size=20
      num_children=20
      num_generations=100
      evaluation_time=50
      evaluation_threshold=20
      restart_interval=2
      disable_passive_joints=1
      enable_pose_logging=0
      #num_evolutions=5 #rather specify this on command line if not default
      #start_run=0  #rather specify this on command line if not default

Accounting

  • Almost done with fitness function for bbox(estimated today)
  • Started three experiments.

Anders Rønningstad

Budget

  • Find a better function that decides which airport to go to.
  • Expand so everything works with airports spread out.
  • Write about implementation:
    • Explain general code setup.
    • Explain simulation strategy.

David Kolden

Budget

  • Run more experiments.
  • Generate more plots.
  • Refine some code.
  • Write about experiments.

Wonho Lee

Accounting

  • Started reading on 'Tagotae' approach for locomotive controller
  • In progress for setting up system for Revolve
  • Received 'Intelligence as adaptive behaviour' by RD. Beer from library
Personlige verktøy