import numpy as np
[docs]class epsGreedyPolicy:
def __init__(self, agent, randomAgent):
'''[summary]
[description]
Arguments:
agent {[type]} -- [description]
randomAgent {[type]} -- [description]
'''
self.agent = agent
self.randomAgent = randomAgent
return
[docs] def act(self, states, eps):
'''[summary]
[description]
Arguments:
states {[type]} -- [description]
eps {[type]} -- [description]
Returns:
[type] -- [description]
'''
actions = []
for state in states:
if np.random.rand() <= eps:
actions.append( self.randomAgent.act( state ) )
else:
actions.append( self.agent.act( state ) )
return actions