with open(fn, 'r') as infile:

pop_data = json.load(infile)

max_fit = float('-inf')

elite_individual = None

for i, individual_data in enumerate(pop_data):

if individual_data['_fit'] > max_fit:

max_fit = individual_data['_fit']

elite_individual = lgp.Behavior()

elite_individual.s_lgp.prog = individual_data['s_lgp']

elite_individual.a_lgp.prog = individual_data['a_lgp']

# Get behavior

behavior_function = agent_behavior_base.behavior

# Open log file

logfile = open('average_baseline.txt', 'w')

logfile.write('gen, survive, avgF\n')

print('gen, survive, avgF')

for gen in range(200):

sim_count = 10

# Average over sim_count simulations

fit_sum = 0

survive_sum = 0

for i in range(sim_count):

# Create and initialize environment

env = environment.Environment(behavior_function, dt = .1, sim_time = 0,

time_out = 50, hash_map_grid_size = 40, width = 800, height = 600,

show_bins = False)

env.create_perimeter_walls(location = 'inside', thickness = 5) # Walls

create.create_test_terrain(env) # Terrain

create.create_random_swarm(env, count = 10, radius = 20) # Swarm

sim_time, survivor_count, total_agent_health, predator_health = env.run_sim_no_gfx()

fitness = total_agent_health - predator_health

fit_sum += fitness

survive_sum += survivor_count

log_string = '{:>3d}, {:>4.1f}, {:>7.1f}'.format(gen, survive_sum/sim_count, fit_sum/sim_count)

logfile.write('{:}\n'.format(log_string))

print(log_string)