import gym
from testing import test_actor_class
from actor import ModifiedGeneticNNActor as MGNNA
from genetics import top_selection
if __name__ == '__main__':
env = gym.make('HalfCheetah-v2')
print("Observation space's shape: ", str(env.observation_space.low.shape))
print("Action space's shape: ", str(env.action_space.low.shape))
test_actor_class(
MGNNA,
env,
savefile='HalfCheetah-v2-GNNAM.txt',
population_size=100,
actor_args={'hidden_layers': [3, 3]},
evolve_args={
'generations': 101,
'simulation_reps': 5, # Changed
'max_steps': 1000,
'selection': lambda p: top_selection(p, cutoff=0.20),
'keep_parents_alive': True,
'p_mutation': 0.2,
'mutation_scale': 0.5,
'render_gens': 20,
'savenum': 1,
'allow_parallel': True
},
render_args={
'fps': 60,
'max_steps': 3000
})
import gym
from testing import test_actor_class
from actor import ModifiedGeneticNNActor as MGNNA
from genetics import top_selection
if __name__ == '__main__':
env = gym.make('HalfCheetah-v2')
print("Observation space's shape: ", str(env.observation_space.low.shape))
print("Action space's shape: ", str(env.action_space.low.shape))
test_actor_class(
MGNNA,
env,
savefile='HalfCheetah-v2-GNNAM.txt',
population_size=100,
actor_args={'hidden_layers': [3, 3]},
evolve_args={
'generations': 101,
'simulation_reps': 1,
'max_steps': 1000,
'selection': lambda p: top_selection(p, cutoff=0.60), # Changed
'keep_parents_alive': True,
'p_mutation': 0.2,
'mutation_scale': 0.5,
'render_gens': 20,
'savenum': 1,
'allow_parallel': True
},
render_args={
'fps': 60,
'max_steps': 3000
})
