def ex5_4():
minor = 4
version = "5_4"
case = get_case6ww()
expts = 10
in_cloud = False
roleouts = 30
episodes = 5 # samples per learning step
results = run_experiments(expts, get_reinforce_experiment, case, roleouts,
episodes, in_cloud, minor)
save_results(results, "REINFORCE", version)
roleouts = 30
episodes = 5 # samples per learning step
results = run_experiments(expts, get_enac_experiment, case, roleouts,
episodes, in_cloud, minor)
save_results(results, "ENAC", version)
def ex5_1():
minor = 1
version = "5_1"
case = get_case6ww()
expts = 8
in_cloud = False
# roleouts = 300
# episodes = 1 # samples per learning step
#
# results = run_experiments(expts, get_re_experiment, case, roleouts,
# episodes, in_cloud, minor)
# save_results(results, "RothErev", version)
#
#
# results = run_experiments(expts, get_q_experiment, case, roleouts,
# episodes, in_cloud, minor)
# save_results(results, "Q", version)
roleouts = 30
episodes = 5 # samples per learning step
results = run_experiments(expts, get_reinforce_experiment, case, roleouts,
episodes, in_cloud, minor)
save_results(results, "REINFORCE", version)
roleouts = 30
episodes = 5 # samples per learning step
results = run_experiments(expts, get_enac_experiment, case, roleouts,
episodes, in_cloud, minor)
save_results(results, "ENAC", version)
__author__ = 'Richard Lincoln, [email protected]' """ This example demonstrates how to compute Nash equilibria. """ import numpy from scipy.io import mmwrite from pyreto import SmartMarket, DISCRIMINATIVE from pyreto.discrete import MarketEnvironment, ProfitTask from common import setup_logging, get_case6ww setup_logging() case = get_case6ww() gens = case.generators#[0:2] #passive = case.generators[2:3] ng = len(gens) mup = [0.0, 10.0, 20.0, 30.0] nm = len(mup) def nash2d(): r = [numpy.zeros((nm, nm)), numpy.zeros((nm, nm))]# * 2#ng #r = numpy.zeros((nm, nm, 2)) #r = numpy.zeros([ng] + ([nm] * ng)) mkt = SmartMarket(case, priceCap=999.0, decommit=False, auctionType=DISCRIMINATIVE
