[research/lotterysim] update crawlers, re-tune controllers for better acc, and stable apr

script/research/lotterysim/core/constants.py

@@ -39,9 +39,9 @@ PRIMARY_REWARD_TARGET = 0.33 # staked ratio
 # secondary controller assumes certain frequency of leaders per slot
 SECONDARY_LEAD_TARGET = 1 #number of lead per slot
 # maximum transaction size
-MAX_BLOCK_SIZE = 1000
+MAX_BLOCK_SIZE = 100
 # maximum transaction computational cost
-MAX_BLOCK_CC = 100
+MAX_BLOCK_CC = 10
 # fee controller computational capacity target
 FEE_TARGET = MAX_BLOCK_CC
 # max fee base value

script/research/lotterysim/discrete_instance_pi_headstart.py

@@ -16,7 +16,9 @@ if __name__ == "__main__":
     mu = PREMINT/NODES
     darkies = [Darkie(random.gauss(mu, mu/10), strategy=random_strategy(EPOCH_LENGTH)) for _ in range(NODES)]
     #dt = DarkfiTable(0, RUNNING_TIME, CONTROLLER_TYPE_DISCRETE, kp=-0.010399999999938556, ki=-0.0365999996461878, kd=0.03840000000000491,  r_kp=-2.53, r_ki=29.5, r_kd=53.77)
-    dt = DarkfiTable(PREMINT, RUNNING_TIME, CONTROLLER_TYPE_DISCRETE, kp=-0.010399999999938556, ki=-0.0365999996461878, kd=0.03840000000000491,  r_kp=-0.719, r_ki=1.6, r_kd=0.1, fee_kp=-0.068188, fee_ki=-0.000205)
+    #dt = DarkfiTable(PREMINT, RUNNING_TIME, CONTROLLER_TYPE_DISCRETE, kp=-0.010399999999938556, ki=-0.0365999996461878, kd=0.03840000000000491,  r_kp=-0.719, r_ki=1.6, r_kd=0.1, fee_kp=-0.068188, fee_ki=-0.000205)
+    #dt = DarkfiTable(PREMINT, RUNNING_TIME, CONTROLLER_TYPE_DISCRETE, kp=-0.010399999999938556, ki=-0.0365999996461878,  r_kp=0.229, r_ki=2.419, fee_kp=-0.068188, fee_ki=-0.000205)
+    dt = DarkfiTable(PREMINT, RUNNING_TIME, CONTROLLER_TYPE_DISCRETE, kp=0.0259, ki=-0.0319, r_kp=0.229, r_ki=2.419, fee_kp=-0.068188, fee_ki=-0.000205)
     for darkie in darkies:
         dt.add_darkie(darkie)
     acc, cc_acc, avg_apy, avg_reward, stake_ratio, avg_apr = dt.background(rand_running_time=False)

script/research/lotterysim/primary_discrete_auto_crawler_pi.py

@@ -6,7 +6,7 @@ from tqdm import tqdm
 import os
 from core.strategy import random_strategy
 
-AVG_LEN = 10
+AVG_LEN = 5
 
 KP_STEP=0.01
 KP_SEARCH=-0.63
@@ -14,7 +14,7 @@ KP_SEARCH=-0.63
 KI_STEP=0.01
 KI_SEARCH=3.35
 
-RUNNING_TIME=5000
+RUNNING_TIME=1000
 NODES = 1000
 
 SHIFTING = 0.05
@@ -30,7 +30,6 @@ KI='ki'
 KP_RANGE_MULTIPLIER = 2
 KI_RANGE_MULTIPLIER = 2
 
-
 highest_gain = (KP_SEARCH, KI_SEARCH)
 
 parser = ArgumentParser()
@@ -45,13 +44,13 @@ rand_running_time = args.rand_running_time
 debug = args.debug
 
 def experiment(controller_type=CONTROLLER_TYPE_DISCRETE, rkp=0, rki=0, distribution=[], hp=True):
-    dt = DarkfiTable(0, RUNNING_TIME, controller_type, kp=-0.010399999999938556, ki=-0.0365999996461878, kd=0, r_kp=rkp, r_ki=rki, r_kd=0)
     RND_NODES = random.randint(5, NODES) if randomize_nodes else NODES
+    dt = DarkfiTable(sum([distribution[i] for i in range(RND_NODES)]), RUNNING_TIME, controller_type, kp=-0.010399999999938556, ki=-0.0365999996461878, kd=0, r_kp=rkp, r_ki=rki, r_kd=0, fee_kp=-0.068188, fee_ki=-0.000205)
     for idx in range(0,RND_NODES):
         darkie = Darkie(distribution[idx], strategy=random_strategy(EPOCH_LENGTH))
         dt.add_darkie(darkie)
-    acc, apy, reward, stake_ratio, apr = dt.background(rand_running_time, hp)
-    return acc, apy, reward, stake_ratio, apr
+    acc, cc_acc, apy, reward, stake_ratio, apr = dt.background(rand_running_time, hp)
+    return acc, cc_acc, apy, reward, stake_ratio, apr
 
 def multi_trial_exp(kp, ki, distribution = [], hp=True):
     global highest_apr
@@ -61,21 +60,24 @@ def multi_trial_exp(kp, ki, distribution = [], hp=True):
     global lowest_apr2target_diff
     new_record=False
     accs = []
+    cc_accs = []
     aprs = []
     rewards = []
     stakes_ratios = []
     aprs = []
     for i in range(0, AVG_LEN):
-        acc, apy, reward, stake_ratio, apr = experiment(CONTROLLER_TYPE_DISCRETE, rkp=kp, rki=ki, distribution=distribution, hp=hp)
+        acc, cc_acc, apy, reward, stake_ratio, apr = experiment(CONTROLLER_TYPE_DISCRETE, rkp=kp, rki=ki, distribution=distribution, hp=hp)
         accs += [acc]
+        cc_accs += [cc_acc]
         rewards += [reward]
         aprs += [apr]
         stakes_ratios += [stake_ratio]
     avg_acc = float(sum(accs))/AVG_LEN
+    avg_cc_acc = float(sum(cc_accs))/AVG_LEN
     avg_reward = float(sum(rewards))/AVG_LEN
     avg_staked = float(sum(stakes_ratios))/AVG_LEN
     avg_apr = float(sum(aprs))/AVG_LEN
-    buff = 'avg(acc): {}, avg(apr): {}, avg(reward): {}, avg(stake ratio): {}, kp: {}, ki:{}, '.format(avg_acc, avg_apr, avg_reward, avg_staked, kp, ki)
+    buff = 'avg(acc): {}, avg(cc_accs): {}, avg(apr): {}, avg(reward): {}, avg(stake ratio): {}, kp: {}, ki:{}, '.format(avg_acc, avg_cc_acc, avg_apr, avg_reward, avg_staked, kp, ki)
     if avg_apr > 0:
         gain = (kp, ki)
         acc_gain = (avg_apr, gain)
@@ -116,8 +118,7 @@ def crawler(crawl, range_multiplier, step=0.1):
             step*=10
     np.random.shuffle(crawl_range)
     crawl_range = tqdm(crawl_range)
-    #distribution =  [random.gauss(ERC20DRK/NODES, ERC20DRK/NODES*0.1) for i in range(NODES)]
-    distribution = [0 for i in range(NODES)]
+    distribution = [random.gauss(ERC20DRK/NODES, ERC20DRK/NODES*0.1) for i in range(NODES)]
     for i in crawl_range:
         kp = i if crawl==KP else highest_gain[0]
         ki = i if crawl==KI else highest_gain[1]

script/research/lotterysim/secondary_discrete_auto_crawler.py

@@ -3,13 +3,13 @@ from core.lottery import DarkfiTable
 from core.utils import *
 from core.darkie import Darkie
 from tqdm import tqdm
-from core.strategy import SigmoidStrategy
+from core.strategy import *
 import os
 
 AVG_LEN = 5
 
 KP_STEP=0.01
-KP_SEARCH= -0.01
+KP_SEARCH=-0.01
 
 KI_STEP=0.01
 KI_SEARCH=-0.036
@@ -17,11 +17,10 @@ KI_SEARCH=-0.036
 KD_STEP=0.01
 KD_SEARCH=0.0384
 
-EPSILON=0.0001
-RUNNING_TIME=10000
+RUNNING_TIME=1000
 NODES = 1000
 
-highest_acc = 0
+highest_acc = 0.2
 
 KP='kp'
 KI='ki'
@@ -45,12 +44,12 @@ rand_running_time = args.rand_running_time
 debug = args.debug
 
 def experiment(controller_type=CONTROLLER_TYPE_DISCRETE, kp=0, ki=0, kd=0, distribution=[], hp=True):
-    dt = DarkfiTable(ERC20DRK, RUNNING_TIME, controller_type, kp=kp, ki=ki, kd=kd)
     RND_NODES = random.randint(5, NODES) if randomize_nodes else NODES
+    dt = DarkfiTable(sum([distribution[i] for i in range(RND_NODES)]), RUNNING_TIME, controller_type, kp=kp, ki=ki, kd=kd)
     for idx in range(0,RND_NODES):
-        darkie = Darkie(distribution[idx], strategy=SigmoidStrategy(EPOCH_LENGTH))
+        darkie = Darkie(distribution[idx], strategy=random_strategy(EPOCH_LENGTH))
         dt.add_darkie(darkie)
-    acc, apy, reward, stake_ratio, apr = dt.background(rand_running_time, hp)
+    acc, cc_acc, apy, reward, stake_ratio, apr = dt.background(rand_running_time, hp)
     return acc
 
 def multi_trial_exp(kp, ki, kd, distribution = [], hp=True):

script/research/lotterysim/secondary_discrete_auto_crawler_pi.py

@@ -3,7 +3,7 @@ from core.lottery import DarkfiTable
 from core.utils import *
 from core.darkie import Darkie
 from tqdm import tqdm
-from core.strategy import SigmoidStrategy
+from core.strategy import *
 import os
 
 AVG_LEN = 5
@@ -18,7 +18,7 @@ EPSILON=0.0001
 RUNNING_TIME=1000
 NODES = 1000
 
-highest_acc = 0
+highest_acc = 0.2
 
 KP='kp'
 KI='ki'
@@ -39,13 +39,13 @@ randomize_nodes = args.randomize_nodes
 rand_running_time = args.rand_running_time
 debug = args.debug
 
-def experiment(controller_type=CONTROLLER_TYPE_DISCRETE, kp=0, ki=0, distribution=[], hp=True):
-    dt = DarkfiTable(ERC20DRK, RUNNING_TIME, controller_type, kp=kp, ki=ki, kd=0)
+def experiment(controller_type, kp, ki, distribution=[], hp=True):
     RND_NODES = random.randint(5, NODES) if randomize_nodes else NODES
+    dt = DarkfiTable(sum([distribution[i] for i in range(RND_NODES)]), RUNNING_TIME, controller_type, kp=kp, ki=ki)
     for idx in range(0,RND_NODES):
-        darkie = Darkie(distribution[idx], strategy=SigmoidStrategy(EPOCH_LENGTH))
+        darkie = Darkie(distribution[idx], strategy=random_strategy(EPOCH_LENGTH))
         dt.add_darkie(darkie)
-    acc, apy, reward, stake_ratio, apr = dt.background(rand_running_time, hp)
+    acc, cc_acc,  apy, reward, stake_ratio, apr = dt.background(rand_running_time, hp)
     return acc
 
 def multi_trial_exp(kp, ki, distribution = [], hp=True):