Genesis Simulation 4 Fitness Analysis

Fitness, cell count, and offspring counts trend positive over multiple generations and epochs. Currently no organism is reaching old age. Death type tends towards exhaustion over time.

Energy efficiency may not be a valuable metric. Since there are no selection pressures on energy efficiency other than starvation and exhaustion, it is unclear if there are trends emerging over generations. Energy waste was removed from the fitness function as it’s strongly correlated with energy efficiency, essentially doubling the energy efficiency impact on fitness.

Population        Death       Offspring       Energy Efficiency       Cells

---

import pandas as pd
import json
import seaborn as sns
import matplotlib.pyplot as plt
import math

total_epochs = 5
performance_path = '/Users/luke/dev/Genetics/data/Rebalanced-Sim4-Epoch-%d-performance.json'
environment_path = '/Users/luke/dev/Genetics/data/Rebalanced-Sim4-Epoch-%d-environment.json'
metrics = [
    { 
        'metric': 'performance', 
        'path': performance_path 
    },
    {
        'metric': 'environment', 
        'path': environment_path 
    }
]

def genetic_data_to_dataframe(path):
    raw_df = pd.read_json(path)
    return pd.json_normalize(raw_df['log'])

def load_simulation_data():
    simulations = {}
    for metric in metrics:
        if not metric['metric'] in simulations:
            simulations[metric['metric']] = {}
        for i in range(total_epochs):
            simulation_instance  = genetic_data_to_dataframe( metric['path'] % i )
            simulation_instance['epoch'] = i
            simulations[metric['metric']][i] = simulation_instance

    return simulations

def merge_epochs(df):
    cat_df = pd.DataFrame();
    for i in range(total_epochs):
        cat_df = pd.concat( [ df[i], cat_df] )
    return cat_df
    

sns.set_theme(style="darkgrid")

simulations = load_simulation_data()

Death ID Mappings

ID	Death
0	Unknown
1	Stagnation
2	Exhaustion
3	Old Age

Population Over Time

all_environment_data = merge_epochs( simulations['environment'] )
fig = sns.lineplot(x="tickCount", y="totalOrganisms", hue='epoch',
        data=all_environment_data ).set_title("Population Over Time")

Fitness Analysis

all_organism_data = merge_epochs( simulations['performance'] )

g = sns.relplot(
    data=all_organism_data,
    x="birthTick", y="fitness", hue="epoch"
)
g.set(xlabel="Birth", ylabel="Fitness", title="Organism Fitness")
g.despine(left=True, bottom=True)
g.fig.set_size_inches(15,15)

fitnessPerEpoch = pd.DataFrame(columns=['epoch', 'fitness' ])
for i in range(total_epochs):
    fitnessPerEpoch.loc[ -1 ] = [ i , simulations['performance'][i]['fitness'].mean() ]
    fitnessPerEpoch.index = fitnessPerEpoch.index + 1

g = sns.lineplot(x="epoch", y="fitness", data=fitnessPerEpoch).set_title("Mean Fitness per Epoch")                           

Death Analysis

palette_color = sns.color_palette('bright') 
  
# plotting data on chart 
for i in range(total_epochs):
    
    simulations['performance'][i]['causeOfDeath'].value_counts().plot.pie( autopct='%1.1f%%', shadow=True, figsize=(2,2))
    plt.title("Epoch " + str(i) + " death breakout.")
    plt.show()

causeOfDeathPerEpoch = pd.DataFrame(columns=['Epoch', 'Cause of Death' ])
for i in range(total_epochs):
    simulations['performance'][i]['deathRating'] = simulations['performance'][i]['causeOfDeath'] / 4
    causeOfDeathPerEpoch.loc[ -1 ] = [ i , simulations['performance'][i]['deathRating'].mean() ]
    causeOfDeathPerEpoch.index = causeOfDeathPerEpoch.index + 1
g = sns.lineplot(x="Epoch", y="Cause of Death",
        data=causeOfDeathPerEpoch).set_title("Morbid Rating ")

Offspring Analysis

childrenPerEpoch = pd.DataFrame(columns=['Epoch', 'Avg Children' ])
for i in range(total_epochs):
    childrenPerEpoch.loc[ -1 ] = [ i , math.log(simulations['performance'][i]['offspring'].mean()) ]
    childrenPerEpoch.index = childrenPerEpoch.index + 1
    
sns.histplot( data=all_organism_data, x='offspring', hue='epoch', binwidth=1,
            element="poly" ).set_title('Offspring Count per Epoch')    
plt.show()

g = sns.lineplot(x="Epoch", y="Avg Children",
        data=childrenPerEpoch).set_title("Normalized Average Children")

Energy Efficiency Analysis

energyEfficiencyPerEpoch = pd.DataFrame(columns=['Epoch', 'Energy Efficiency' ])

for i in range(total_epochs):
    simulations['performance'][i]['differential'] = simulations['performance'][i]['totalEnergyHarvested'] - simulations['performance'][i]['totalEnergyMetabolized']
    simulations['performance'][i]['normalized'] = 1 / simulations['performance'][i]['differential']
    energyEfficiencyPerEpoch.loc[ -1 ] = [ i , 1 / simulations['performance'][i]['differential'].mean() ]
    energyEfficiencyPerEpoch.index = energyEfficiencyPerEpoch.index + 1
    g = sns.displot( data=simulations['performance'][i], x="differential", kind="kde")
    g.figure.subplots_adjust(top=.9);
    g.figure.suptitle( 'Epoch ' + str(i) + ' Energy Efficiency Distribution')
    plt.show()
    



g = sns.lineplot(x="Epoch", y="Energy Efficiency",
        data=energyEfficiencyPerEpoch)
plt.show()

Cell Analysis

cellsPerEpoch = pd.DataFrame(columns=['Epoch', 'Avg Cells' ])
for i in range(total_epochs):
    cellsPerEpoch.loc[ -1 ] = [ i , math.log(simulations['performance'][i]['cells'].mean()) ]
    cellsPerEpoch.index = cellsPerEpoch.index + 1
    sns.histplot( data=simulations['performance'][i], x='cells', binwidth=1,
                element="poly").set_title('Epoch ' + str(i) + ' Cell Count') 
    plt.show()

    
sns.histplot( data=all_organism_data, x='cells', hue='epoch', binwidth=1,
            element="poly" ).set_title('Cell Count per Epoch Overlay') 
plt.show()
sns.lineplot(x="Epoch", y="Avg Cells",
        data=cellsPerEpoch).set_title("Normalized Organism Size")
plt.show()