This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
# Hell is Game Theory Folk Theorems | |
# https://www.lesswrong.com/posts/d2HvpKWQ2XGNsHr8s/hell-is-game-theory-folk-theorems | |
import random | |
import math | |
def minimize(prev_list, prev_temp): | |
return 30 | |
def maximize(prev_list, prev_temp): | |
return 100 | |
def equilibrium(prev_list, prev_temp): | |
if min(prev_list) < 99 and prev_temp != 100: return 100 | |
return 99 | |
def randomize(prev_list, prev_temp): | |
return random.choice(prev_list) | |
def average(prev_list, prev_temp): | |
return int(prev_temp) | |
STRATS = (minimize, maximize, equilibrium, equilibrium, randomize, average) | |
class Agent: | |
def __init__(self): | |
self.strategy = random.choice(STRATS) | |
self.total_reward = 0 | |
self.buffer = {v: 0 for v in STRATS} | |
def execute(self, prev_list, prev_temp): | |
return self.strategy(prev_list, prev_temp) | |
def reward(self, v): | |
self.total_reward += v | |
self.buffer[self.strategy] += v / 1000 | |
if random.random() < 1/8: | |
strats = list(self.buffer.keys()) | |
weights = [math.exp(rew) for rew in self.buffer.values()] | |
self.strategy = random.choices(strats, weights=weights)[0] | |
def __str__(self): | |
return f'[Agent | strategy:{self.strategy.__name__}\t| total: {self.total_reward:.2f}]' | |
def str(self, v): | |
return f'[Agent | strategy:{self.strategy.__name__}\t| total: {self.total_reward:.2f} | prev: {v}]' | |
class Field: | |
def __init__(self, N=10): | |
self.agents = [Agent() for _ in range(N)] | |
self.prev = [30 for _ in range(N)] | |
self.pt = 30 | |
def simulate(self): | |
actions = [agent.execute(self.prev, self.pt) for agent in self.agents] | |
self.prev = actions[:] | |
self.pt = sum(self.prev) / len(self.prev) | |
for agent in self.agents: agent.reward(50 - self.pt) | |
# print(self) | |
def __str__(self): | |
return f'\nPrev temp: {self.pt}\n' + '\n'.join(' ' + s.str(n) for s,n in zip(self.agents, self.prev)) | |
try: | |
f = Field() | |
print(f) | |
for i in range(1000): | |
for _ in range(100): f.simulate() | |
print(f'\nStep {i+1}00') | |
print(f) | |
except: | |
import traceback | |
traceback.print_exc() |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment