Comparative Analysis of Multi-Objective and Evolutionary Reinforcement Learning Algorithms in Electric Vehicle Charging Management - Extended Appendix

Appendix: Simulation

PV Power Curve

PV Power Curve: Hourly average power of the PV system over a recorded one-year period.

Appendix: Hyperparameter Optimization Results

Specific parameters for all agents are held constant, while others are subject to optimization. The values for these fixed parameters have been determined through previous manual random testing, ensuring a robust baseline for comparison. For the Multi-Objective Evolutionary Reinforcement Learning(MOEvoRL) models, parameters have been configured with a population size of 200 and a sigmoid activation function. The Multi-Objective Deep Deterministic Policy Gradient (MODDPG) models have a predefined network architecture consisting of three layers, each with 1024 units, both for the Actor and Critic networks. Adam optimizer is used for MODDPG optimization, with Relu as the activation function and Tanh for the output activation.

Parameter	Distribution	Min	Max
sbx_prob	uniform	0.6	0.95
mut_prob	uniform	0.1	0.3
sbx_eta	int_uniform	3	30
mut_eta	int_uniform	3	30
net_arch	categorical	[32, 32], [32, 32, 32], [32, 32, 32, 32]	-
batch_size	categorical	8, 16, 32	-
env_iterations	categorical	2, 5, 10	-
pop_size	fixed	200	-
activation_function	fixed	sigmoid	-

MOEvoRL-NSGA-II and MOEvoRL-SPEA2 Hyperparameter Search Space: The hyperparameter search space for the MOEvoRL-NSGA-II and MOEvoRL-SPEA2 agents.

Parameter	Distribution	Min	Max
conn_add_prob	uniform	0.6	0.9
conn_delete_prob	uniform	0.1	0.4
node_add_prob	uniform	0.6	0.9
node_delete_prob	uniform	0.1	0.4
survival_threshold	uniform	0.1	0.3
aggregation_mutate_rate	uniform	0.1	0.3
weight_mutate_rate	uniform	0.6	0.9
bias_mutate_rate	uniform	0.6	0.9
batch_size	categorical	8, 16, 32	-
env_iterations	categorical	2, 5, 10	-
pop_size	fixed	200	-
activation_function	fixed	sigmoid	-

MOEvoRL-FF-NEAT and MOEvoRL-RNN-NEAT Hyperparameter Search Space: The hyperparameter search space for the MOEvoRL-FF-NEAT agents.

Parameter	Distribution	Min	Max
gamma	uniform	0.95	0.9999
tau	uniform	0	0.1
per_alpha	uniform	0.1	1.0
learning_rate	uniform	(1 \times 10^{-5})	(1 \times 10^{-4})
learning_starts	int_uniform	1000	10000
policy_frequency	int_uniform	1	20
buffer_size	int_uniform	(1 \times 10^5)	(2 \times 10^6)
batch_size	categorical	128, 256, 512	-
net_arch	fixed	[1024, 1024, 1024]	-
activation_function	fixed	relu	-
output_activation_function	fixed	tanh	-
optimizer	fixed	adam	-

MOODPG Hyperparameter Search Space: The hyperparameter search space for the MODDPG agents

MOEvoRL-NSGA-II Optimal Hyperparameters

Parameter	Value
sbx_prob	0.90
mut_prob	0.18
sbx_eta	23
mut_eta	22
net_arch	[32, 32, 32]
batch_size	32
env_iterations	10
pop_size	200
activation_function	sigmoid

MOEvoRL-NSGA-II CS05-Med-42: Best hyperparameter parameter configuration.

Parameter	Value
sbx_prob	0.75
mut_prob	0.26
sbx_eta	23
mut_eta	12
net_arch	[32, 32, 32]
batch_size	16
env_iterations	10
pop_size	200
activation_function	sigmoid

MOEvoRL-NSGA-II CS05-Med-AR-42: Best hyperparameter parameter configuration.

Parameter	Value
sbx_prob	0.83
mut_prob	0.29
sbx_eta	26
mut_eta	27
net_arch	[32, 32, 32]
batch_size	16
env_iterations	10
pop_size	200
activation_function	sigmoid

MOEvoRL-NSGA-II CS10-Med-42: Best hyperparameter parameter configuration.

Parameter	Value
sbx_prob	0.74
mut_prob	0.14
sbx_eta	19
mut_eta	15
net_arch	[32, 32, 32]
batch_size	8
env_iterations	10
pop_size	200
activation_function	sigmoid

MOEvoRL-NSGA-II CS10-Med-AR-42: Best hyperparameter parameter configuration.

Parameter	Value
sbx_prob	0.81
mut_prob	0.25
sbx_eta	10
mut_eta	25
net_arch	[32, 32, 32]
batch_size	32
env_iterations	10
pop_size	200
activation_function	sigmoid

MOEvoRL-NSGA-II CS15-Med-42: Best hyperparameter parameter configuration.

Parameter	Value
sbx_prob	0.89
mut_prob	0.28
sbx_eta	27
mut_eta	28
net_arch	[32, 32, 32]
batch_size	32
env_iterations	5
pop_size	200
activation_function	sigmoid

MOEvoRL-NSGA-II CS15-Med-AR-42: Best hyperparameter parameter configuration.

MOEvoRL-SPEA2 Optimal Hyperparameters

Parameter	Value
sbx_prob	0.88
mut_prob	0.21
sbx_eta	16
mut_eta	24
net_arch	[32, 32, 32]
batch_size	32
env_iterations	10
pop_size	200
activation_function	sigmoid

MOEvoRL-SPEA2 CS05-Med-42: Best hyperparameter parameter configuration.

Parameter	Value
sbx_prob	0.76
mut_prob	0.20
sbx_eta	16
mut_eta	12
net_arch	[32, 32, 32, 32]
batch_size	16
env_iterations	10
pop_size	200
activation_function	sigmoid

MOEvoRL-SPEA2 CS05-Med-AR-42: Best hyperparameter parameter configuration.

Parameter	Value
sbx_prob	0.82
mut_prob	0.28
sbx_eta	14
mut_eta	8
net_arch	[32, 32, 32]
batch_size	32
env_iterations	10
pop_size	200
activation_function	sigmoid

MOEvoRL-SPEA2 CS10-Med-42: Best hyperparameter parameter configuration.

Parameter	Value
sbx_prob	0.85
mut_prob	0.22
sbx_eta	18
mut_eta	13
net_arch	[32, 32, 32, 32]
batch_size	16
env_iterations	10
pop_size	200
activation_function	sigmoid

MOEvoRL-SPEA2 CS10-Med-AR-42: Best hyperparameter parameter configuration.

Parameter	Value
sbx_prob	0.92
mut_prob	0.29
sbx_eta	20
mut_eta	29
net_arch	[32, 32, 32]
batch_size	32
env_iterations	10
pop_size	200
activation_function	sigmoid

MOEvoRL-SPEA2 CS15-Med-42: Best hyperparameter parameter configuration.

Parameter	Value
sbx_prob	0.95
mut_prob	0.18
sbx_eta	15
mut_eta	12
net_arch	[32, 32, 32]
batch_size	32
env_iterations	5
pop_size	200
activation_function	sigmoid

MOEvoRL-SPEA2 CS15-Med-AR-42: Best hyperparameter parameter configuration.

MOEvoRL-FF-NEAT Optimal Hyperparameters

Parameter	Value
conn_add_prob	0.72
conn_delete_prob	0.25
node_add_prob	0.60
node_delete_prob	0.28
survival_threshold	0.20
aggregation_mutate_rate	0.22
weight_mutate_rate	0.89
bias_mutate_rate	0.64
batch_size	32
env_iterations	10
pop_size	200
activation_function	sigmoid

MOEvoRL-FF-NEAT CS05-Med-42: Best hyperparameter parameter configuration.

Parameter	Value
conn_add_prob	0.73
conn_delete_prob	0.39
node_add_prob	0.79
node_delete_prob	0.32
survival_threshold	0.12
aggregation_mutate_rate	0.20
weight_mutate_rate	0.71
bias_mutate_rate	0.79
batch_size	8
env_iterations	2
pop_size	200
activation_function	sigmoid

MOEvoRL-FF-NEAT CS05-Med-AR-42: Best hyperparameter parameter configuration.

Parameter	Value
conn_add_prob	0.79
conn_delete_prob	0.33
node_add_prob	0.70
node_delete_prob	0.28
survival_threshold	0.16
aggregation_mutate_rate	0.26
weight_mutate_rate	0.65
bias_mutate_rate	0.66
batch_size	16
env_iterations	10
pop_size	200
activation_function	sigmoid

MOEvoRL-FF-NEAT CS10-Med-42: Best hyperparameter parameter configuration.

Parameter	Value
conn_add_prob	0.82
conn_delete_prob	0.18
node_add_prob	0.67
node_delete_prob	0.31
survival_threshold	0.23
aggregation_mutate_rate	0.26
weight_mutate_rate	0.74
bias_mutate_rate	0.84
batch_size	16
env_iterations	10
pop_size	200
activation_function	sigmoid

MOEvoRL-FF-NEAT CS10-Med-AR-42: Best hyperparameter parameter configuration.

Parameter	Value
conn_add_prob	0.73
conn_delete_prob	0.13
node_add_prob	0.80
node_delete_prob	0.33
survival_threshold	0.18
aggregation_mutate_rate	0.16
weight_mutate_rate	0.63
bias_mutate_rate	0.81
batch_size	32
env_iterations	10
pop_size	200
activation_function	sigmoid

MOEvoRL-FF-NEAT CS15-Med-42: Best hyperparameter parameter configuration.

Parameter	Value
conn_add_prob	0.70
conn_delete_prob	0.36
node_add_prob	0.61
node_delete_prob	0.34
survival_threshold	0.10
aggregation_mutate_rate	0.24
weight_mutate_rate	0.88
bias_mutate_rate	0.71
batch_size	32
env_iterations	10
pop_size	200
activation_function	sigmoid

MOEvoRL-FF-NEAT CS15-Med-AR-42: Best hyperparameter parameter configuration.

MOEvoRL-RNN-NEAT Optimal Hyperparameters

Parameter	Value
conn_add_prob	0.72
conn_delete_prob	0.35
node_add_prob	0.66
node_delete_prob	0.16
survival_threshold	0.17
aggregation_mutate_rate	0.24
weight_mutate_rate	0.81
bias_mutate_rate	0.78
batch_size	32
env_iterations	5
pop_size	200
activation_function	sigmoid

MOEvoRL-RNN-NEAT CS05-Med-42: Best hyperparameter parameter configuration.

Parameter	Value
conn_add_prob	0.83
conn_delete_prob	0.39
node_add_prob	0.84
node_delete_prob	0.39
survival_threshold	0.27
aggregation_mutate_rate	0.23
weight_mutate_rate	0.87
bias_mutate_rate	0.85
batch_size	8
env_iterations	5
pop_size	200
activation_function	sigmoid

MOEvoRL-RNN-NEAT CS05-Med-AR-42: Best hyperparameter parameter configuration.

Parameter	Value
conn_add_prob	0.76
conn_delete_prob	0.24
node_add_prob	0.77
node_delete_prob	0.25
survival_threshold	0.21
aggregation_mutate_rate	0.29
weight_mutate_rate	0.85
bias_mutate_rate	0.66
batch_size	32
env_iterations	10
pop_size	200
activation_function	sigmoid

MOEvoRL-RNN-NEAT CS10-Med-42: Best hyperparameter parameter configuration.

Parameter	Value
conn_add_prob	0.75
conn_delete_prob	0.15
node_add_prob	0.75
node_delete_prob	0.20
survival_threshold	0.23
aggregation_mutate_rate	0.24
weight_mutate_rate	0.80
bias_mutate_rate	0.67
batch_size	32
env_iterations	10
pop_size	200
activation_function	sigmoid

MOEvoRL-RNN-NEAT CS10-Med-AR-42: Best hyperparameter parameter configuration.

Parameter	Value
conn_add_prob	0.78
conn_delete_prob	0.24
node_add_prob	0.69
node_delete_prob	0.13
survival_threshold	0.28
aggregation_mutate_rate	0.12
weight_mutate_rate	0.83
bias_mutate_rate	0.78
batch_size	8
env_iterations	2
pop_size	200
activation_function	sigmoid

MOEvoRL-RNN-NEAT CS15-Med-42: Best hyperparameter parameter configuration.

Parameter	Value
conn_add_prob	0.73
conn_delete_prob	0.35
node_add_prob	0.65
node_delete_prob	0.33
survival_threshold	0.18
aggregation_mutate_rate	0.29
weight_mutate_rate	0.71
bias_mutate_rate	0.67
batch_size	8
env_iterations	2
pop_size	200
activation_function	sigmoid

MOEvoRL-RNN-NEAT CS15-Med-AR-42: Best hyperparameter parameter configuration.

MODDPG Optimal Hyperparameters

Parameter	Value
gamma	0.99
tau	0.04
per_alpha	0.71
learning_rate	1.1 x 10^{-5}
learning_starts	7500
policy_frequency	13
buffer_size	$1 x 10^6$
batch_size	256
net_arch	[1024, 1024, 1024]
activation_function	relu
output_activation_function	tanh
optimizer	adam

MODDPG CS05-Med-42: Best hyperparameter parameter configuration.

Parameter	Value
gamma	0.99
tau	0.09
per_alpha	0.60
learning_rate	1.8 x 10^{-5}
learning_starts	1500
policy_frequency	6
buffer_size	$1 x 10^6$
batch_size	256
net_arch	[1024, 1024, 1024]
activation_function	relu
output_activation_function	tanh
optimizer	adam

MODDPG CS05-Med-AR-42: Best hyperparameter parameter configuration.

Parameter	Value
gamma	0.99
tau	0.02
per_alpha	0.65
learning_rate	6.2 x 10^{-5}
learning_starts	9100
policy_frequency	6
buffer_size	$1 x 10^6$
batch_size	512
net_arch	[1024, 1024, 1024]
activation_function	relu
output_activation_function	tanh
optimizer	adam

MODDPG CS10-Med-42: Best hyperparameter parameter configuration.

Parameter	Value
gamma	0.99
tau	0.04
per_alpha	0.80
learning_rate	8.6 x 10^{-5}
learning_starts	5600
policy_frequency	16
buffer_size	$1 x 10^6$
batch_size	128
net_arch	[1024, 1024, 1024]
activation_function	relu
output_activation_function	tanh
optimizer	adam

MODDPG CS10-Med-AR-42: Best hyperparameter parameter configuration.

Parameter	Value
gamma	0.99
tau	0.03
per_alpha	0.77
learning_rate	4.5 x 10^{-5}
learning_starts	2000
policy_frequency	19
buffer_size	$1 x 10^6$
batch_size	512
net_arch	[1024, 1024, 1024]
activation_function	relu
output_activation_function	tanh
optimizer	adam

MODDPG CS15-Med-42: Best hyperparameter parameter configuration.

Parameter	Value
gamma	0.99
tau	0.02
per_alpha	0.84
learning_rate	1.6 x 10^{-5}
learning_starts	1700
policy_frequency	1
buffer_size	$1.5 x 10^6$
batch_size	128
net_arch	[1024, 1024, 1024]
activation_function	relu
output_activation_function	tanh
optimizer	adam

Appendix: Anaylsis of Charging Events

CS05 Charging Events

Arrival and Charging Patterns for CS05-Med-42

Arrival and Charging Patterns for CS05-Low-71

Arrival and Charging Patterns for CS05-Med-71

Arrival and Charging Patterns for CS05-High-71

EV Model Distribution in CS05-Med-42

EV Model Distribution in CS05-Low-71

EV Model Distribution in CS05-Med-71

EV Model Distribution in CS05-High-71

CS10 Charging Events

Arrival and Charging Patterns for CS10-Med-42

Arrival and Charging Patterns for CS10-Low-71

Arrival and Charging Patterns for CS10-Med-71

Arrival and Charging Patterns for CS10-High-71

EV Model Distribution in CS10-Med-42

EV Model Distribution in CS10-Low-71

EV Model Distribution in CS10-Med-71

EV Model Distribution in CS10-High-71

CS15 Charging Events

Arrival and Charging Patterns for CS15-Med-42

Arrival and Charging Patterns for CS15-Low-71

Arrival and Charging Patterns for CS15-Med-71

Arrival and Charging Patterns for CS15-High-71

EV Model Distribution in CS15-Med-42

EV Model Distribution in CS15-Low-71

EV Model Distribution in CS15-Med-71

EV Model Distribution in CS15-High-71

Appendix: Comparative Analysis of Training Dynamics

CS05-Med-42 Training Performance

R0 Comparison CS05-Med-42: Diagram provides a comparative analysis of the convergence behavior of five models in scenario CS05-Med-42, focusing on the reward R0 throughout the training episodes.

R1 Comparison CS05-Med-42: Diagram provides a comparative analysis of the convergence behavior of five models in scenario CS05-Med-42, focusing on the reward R1 throughout the training episodes.

R2 Comparison CS05-Med-42: Diagram provides a comparative analysis of the convergence behavior of five models in scenario CS05-Med-42, focusing on the reward R2 throughout the training episodes.

Success Rate Comparison CS05-Med-42: Diagram provides a comparative analysis of the convergence behavior of five models in scenario CS05-Med-42, focusing on the Success Rate throughout the training episodes.

Hypervolume Comparison CS05-Med-42: Diagram provides a comparative analysis of the convergence behavior of five models in scenario CS05-Med-42, focusing on the HV throughout the training episodes.

CS05-Med-AR-42 Training Performance

R0 Comparison CS05-Med-AR-42: Diagram provides a comparative analysis of the convergence behavior of five models in scenario CS05-Med-AR-42, focusing on the reward R0 throughout the training episodes.

R1 Comparison CS05-Med-AR-42: Diagram provides a comparative analysis of the convergence behavior of five models in scenario CS05-Med-AR-42, focusing on the reward R1 throughout the training episodes.

R2 Comparison CS05-Med-AR-42: Diagram provides a comparative analysis of the convergence behavior of five models in scenario CS05-Med-AR-42, focusing on the reward R2 throughout the training episodes.

Hypervolume Comparison CS05-Med-AR-42: Diagram provides a comparative analysis of the convergence behavior of five models in scenario CS05-Med-AR-42, focusing on the HV throughout the training episodes.

CS10-Med-42 Training Performance

R0 Comparison CS10-Med-42: Diagram provides a comparative analysis of the convergence behavior of five models in scenario CS10-Med-42, focusing on the reward R0 throughout the training episodes.

R1 Comparison CS10-Med-42: Diagram provides a comparative analysis of the convergence behavior of five models in scenario CS10-Med-42, focusing on the reward R1 throughout the training episodes.

R2 Comparison CS10-Med-42: Diagram provides a comparative analysis of the convergence behavior of five models in scenario CS10-Med-42, focusing on the reward R2 throughout the training episodes.

Success Rate Comparison CS10-Med-42: Diagram provides a comparative analysis of the convergence behavior of five models in scenario CS10-Med-42, focusing on the Success Rate throughout the training episodes.

Hypervolume Comparison CS10-Med-42: Diagram provides a comparative analysis of the convergence behavior of five models in scenario CS10-Med-42, focusing on the HV throughout the training episodes.

CS10-Med-AR-42 Training Performance

R0 Comparison CS10-Med-AR-42: Diagram provides a comparative analysis of the convergence behavior of five models in scenario CS10-Med-AR-42, focusing on the reward R0 throughout the training episodes.

R1 Comparison CS10-Med-AR-42: Diagram provides a comparative analysis of the convergence behavior of five models in scenario CS10-Med-AR-42, focusing on the reward R1 throughout the training episodes.

R2 Comparison CS10-Med-AR-42: Diagram provides a comparative analysis of the convergence behavior of five models in scenario CS10-Med-AR-42, focusing on the reward R2 throughout the training episodes.

Hypervolume Comparison CS10-Med-AR-42: Diagram provides a comparative analysis of the convergence behavior of five models in scenario CS10-Med-AR-42, focusing on the HV throughout the training episodes.

CS15-Med-42 Training Performance

R0 Comparison CS15-Med-42: Diagram provides a comparative analysis of the convergence behavior of five models in scenario CS15-Med-42, focusing on the reward R0 throughout the training episodes.

R1 Comparison CS15-Med-42: Diagram provides a comparative analysis of the convergence behavior of five models in scenario CS15-Med-42, focusing on the reward R1 throughout the training episodes.

R2 Comparison CS15-Med-42: Diagram provides a comparative analysis of the convergence behavior of five models in scenario CS15-Med-42, focusing on the reward R2 throughout the training episodes.

Success Rate Comparison CS15-Med-42: Diagram provides a comparative analysis of the convergence behavior of five models in scenario CS15-Med-42, focusing on the Success Rate throughout the training episodes.

Hypervolume Comparison CS15-Med-42: Diagram provides a comparative analysis of the convergence behavior of five models in scenario CS15-Med-42, focusing on the HV throughout the training episodes.

CS15-Med-AR-42 Training Performance

R0 Comparison CS15-Med-AR-42: Diagram provides a comparative analysis of the convergence behavior of five models in scenario CS15-Med-AR-42, focusing on the reward R0 throughout the training episodes.

R1 Comparison CS15-Med-AR-42: Diagram provides a comparative analysis of the convergence behavior of five models in scenario CS15-Med-AR-42, focusing on the reward R1 throughout the training episodes.

R2 Comparison CS15-Med-AR-42: Diagram provides a comparative analysis of the convergence behavior of five models in scenario CS15-Med-AR-42, focusing on the reward R2 throughout the training episodes.

Hypervolume Comparison CS15-Med-AR-42: Diagram provides a comparative analysis of the convergence behavior of five models in scenario CS15-Med-AR-42, focusing on the HV throughout the training episodes.

Appendix: Comparative Analysis of Failures

CS05-Low-71 Failure Analysis

Failure Classification CS05-Low-71 MOEvoRL-NSGA-II: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-NSGA-II algorithm in CS05-Low-71.

Failure Analysis CS05-Low-71 MOEvoRL-NSGA-II: Comprehensive failure analysis of the MOEvoRL-NSGA-II algorithm in CS05-Low-71.

Failure Classification CS05-Low-71 MOEvoRL-SPEA2: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-SPEA2 algorithm in CS05-Low-71.

Failure Analysis CS05-Low-71 MOEvoRL-SPEA2: Comprehensive failure analysis of the MOEvoRL-SPEA2 algorithm in CS05-Low-71.

Failure Classification CS05-Low-71 MOEvoRL-FF-NEAT: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-FF-NEAT algorithm in CS05-Low-71.

Failure Analysis CS05-Low-71 MOEvoRL-FF-NEAT: Comprehensive failure analysis of the MOEvoRL-FF-NEAT algorithm in CS05-Low-71.

Failure Classification CS05-Low-71 MOEvoRL-RNN-NEAT: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-RNN-NEAT algorithm in CS05-Low-71.

Failure Analysis CS05-Low-71 MOEvoRL-RNN-NEAT: Comprehensive failure analysis of the MOEvoRL-RNN-NEAT algorithm in CS05-Low-71.

Failure Classification CS05-Low-71 MODDPG: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MODDPG algorithm in CS05-Low-71.

Failure Analysis CS05-Low-71 MODDPG: Comprehensive failure analysis of the MODDPG algorithm in CS05-Low-71.

CS05-Med-71 Failure Analysis

Failure Classification CS05-Med-71 MOEvoRL-NSGA-II: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-NSGA-II algorithm in CS05-Med-71.

Failure Analysis CS05-Med-71 MOEvoRL-NSGA-II: Comprehensive failure analysis of the MOEvoRL-NSGA-II algorithm in CS05-Med-71.

Failure Classification CS05-Med-71 MOEvoRL-SPEA2: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-SPEA2 algorithm in CS05-Med-71.

Failure Analysis CS05-Med-71 MOEvoRL-SPEA2: Comprehensive failure analysis of the MOEvoRL-SPEA2 algorithm in CS05-Med-71.

Failure Classification CS05-Med-71 MOEvoRL-FF-NEAT: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-FF-NEAT algorithm in CS05-Med-71.

Failure Analysis CS05-Med-71 MOEvoRL-FF-NEAT: Comprehensive failure analysis of the MOEvoRL-FF-NEAT algorithm in CS05-Med-71.

Failure Classification CS05-Med-71 MOEvoRL-RNN-NEAT: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-RNN-NEAT algorithm in CS05-Med-71.

Failure Analysis CS05-Med-71 MOEvoRL-RNN-NEAT: Comprehensive failure analysis of the MOEvoRL-RNN-NEAT algorithm in CS05-Med-71.

Failure Classification CS05-Med-71 MODDPG: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MODDPG algorithm in CS05-Med-71.

Failure Analysis CS05-Med-71 MODDPG: Comprehensive failure analysis of the MODDPG algorithm in CS05-Med-71.

CS05-High-71 Failure Analysis

Failure Classification CS05-High-71 MOEvoRL-NSGA-II: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-NSGA-II algorithm in CS05-High-71.

Failure Analysis CS05-High-71 MOEvoRL-NSGA-II: Comprehensive failure analysis of the MOEvoRL-NSGA-II algorithm in CS05-High-71.

Failure Classification CS05-High-71 MOEvoRL-SPEA2: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-SPEA2 algorithm in CS05-High-71.

Failure Analysis CS05-High-71 MOEvoRL-SPEA2: Comprehensive failure analysis of the MOEvoRL-SPEA2 algorithm in CS05-High-71.

Failure Classification CS05-High-71 MOEvoRL-FF-NEAT: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-FF-NEAT algorithm in CS05-High-71.

Failure Analysis CS05-High-71 MOEvoRL-FF-NEAT: Comprehensive failure analysis of the MOEvoRL-FF-NEAT algorithm in CS05-High-71.

Failure Classification CS05-High-71 MOEvoRL-RNN-NEAT: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-RNN-NEAT algorithm in CS05-High-71.

Failure Analysis CS05-High-71 MOEvoRL-RNN-NEAT: Comprehensive failure analysis of the MOEvoRL-RNN-NEAT algorithm in CS05-High-71.

Failure Classification CS05-High-71 MODDPG: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MODDPG algorithm in CS05-High-71.

Failure Analysis CS05-High-71 MODDPG: Comprehensive failure analysis of the MODDPG algorithm in CS05-High-71.

CS10-Low-71 Failure Analysis

Failure Classification CS10-Low-71 MOEvoRL-NSGA-II: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-NSGA-II algorithm in CS10-Low-71.

Failure Analysis CS10-Low-71 MOEvoRL-NSGA-II: Comprehensive failure analysis of the MOEvoRL-NSGA-II algorithm in CS10-Low-71.

Failure Classification CS10-Low-71 MOEvoRL-SPEA2: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-SPEA2 algorithm in CS10-Low-71.

Failure Analysis CS10-Low-71 MOEvoRL-SPEA2: Comprehensive failure analysis of the MOEvoRL-SPEA2 algorithm in CS10-Low-71.

Failure Classification CS10-Low-71 MOEvoRL-FF-NEAT: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-FF-NEAT algorithm in CS10-Low-71.

Failure Analysis CS10-Low-71 MOEvoRL-FF-NEAT: Comprehensive failure analysis of the MOEvoRL-FF-NEAT algorithm in CS10-Low-71.

Failure Classification CS10-Low-71 MOEvoRL-RNN-NEAT: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-RNN-NEAT algorithm in CS10-Low-71.

Failure Analysis CS10-Low-71 MOEvoRL-RNN-NEAT: Comprehensive failure analysis of the MOEvoRL-RNN-NEAT algorithm in CS10-Low-71.

Failure Classification CS10-Low-71 MODDPG: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MODDPG algorithm in CS10-Low-71.

Failure Analysis CS10-Low-71 MODDPG: Comprehensive failure analysis of the MODDPG algorithm in CS10-Low-71.

CS10-Med-71 Failure Analysis

Failure Classification CS10-Med-71 MOEvoRL-NSGA-II: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-NSGA-II algorithm in CS10-Med-71.

Failure Analysis CS10-Med-71 MOEvoRL-NSGA-II: Comprehensive failure analysis of the MOEvoRL-NSGA-II algorithm in CS10-Med-71.

Failure Classification CS10-Med-71 MOEvoRL-SPEA2: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-SPEA2 algorithm in CS10-Med-71.

Failure Analysis CS10-Med-71 MOEvoRL-SPEA2: Comprehensive failure analysis of the MOEvoRL-SPEA2 algorithm in CS10-Med-71.

Failure Classification CS10-Med-71 MOEvoRL-FF-NEAT: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-FF-NEAT algorithm in CS10-Med-71.

Failure Analysis CS10-Med-71 MOEvoRL-FF-NEAT: Comprehensive failure analysis of the MOEvoRL-FF-NEAT algorithm in CS10-Med-71.

Failure Classification CS10-Med-71 MOEvoRL-RNN-NEAT: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-RNN-NEAT algorithm in CS10-Med-71.

Failure Analysis CS10-Med-71 MOEvoRL-RNN-NEAT: Comprehensive failure analysis of the MOEvoRL-RNN-NEAT algorithm in CS10-Med-71.

Failure Classification CS10-Med-71 MODDPG: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MODDPG algorithm in CS10-Med-71.

Failure Analysis CS10-Med-71 MODDPG: Comprehensive failure analysis of the MODDPG algorithm in CS10-Med-71.

CS10-High-71 Failure Analysis

Failure Classification CS10-High-71 MOEvoRL-NSGA-II: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-NSGA-II algorithm in CS10-High-71.

Failure Analysis CS10-High-71 MOEvoRL-NSGA-II: Comprehensive failure analysis of the MOEvoRL-NSGA-II algorithm in CS10-High-71.

Failure Classification CS10-High-71 MOEvoRL-SPEA2: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-SPEA2 algorithm in CS10-High-71.

Failure Analysis CS10-High-71 MOEvoRL-SPEA2: Comprehensive failure analysis of the MOEvoRL-SPEA2 algorithm in CS10-High-71.

Failure Classification CS10-High-71 MOEvoRL-FF-NEAT: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-FF-NEAT algorithm in CS10-High-71.

Failure Analysis CS10-High-71 MOEvoRL-FF-NEAT: Comprehensive failure analysis of the MOEvoRL-FF-NEAT algorithm in CS10-High-71.

Failure Classification CS10-High-71 MOEvoRL-RNN-NEAT: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-RNN-NEAT algorithm in CS10-High-71.

Failure Analysis CS10-High-71 MOEvoRL-RNN-NEAT: Comprehensive failure analysis of the MOEvoRL-RNN-NEAT algorithm in CS10-High-71.

Failure Classification CS10-High-71 MODDPG: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MODDPG algorithm in CS10-High-71.

Failure Analysis CS10-High-71 MODDPG: Comprehensive failure analysis of the MODDPG algorithm in CS10-High-71.

CS15-Low-71 Failure Analysis

Failure Classification CS15-Low-71 MOEvoRL-NSGA-II: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-NSGA-II algorithm in CS15-Low-71.

Failure Analysis CS15-Low-71 MOEvoRL-NSGA-II: Comprehensive failure analysis of the MOEvoRL-NSGA-II algorithm in CS15-Low-71.

Failure Classification CS15-Low-71 MOEvoRL-SPEA2: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-SPEA2 algorithm in CS15-Low-71.

Failure Analysis CS15-Low-71 MOEvoRL-SPEA2: Comprehensive failure analysis of the MOEvoRL-SPEA2 algorithm in CS15-Low-71.

Failure Classification CS15-Low-71 MOEvoRL-FF-NEAT: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-FF-NEAT algorithm in CS15-Low-71.

Failure Analysis CS15-Low-71 MOEvoRL-FF-NEAT: Comprehensive failure analysis of the MOEvoRL-FF-NEAT algorithm in CS15-Low-71.

Failure Classification CS15-Low-71 MOEvoRL-RNN-NEAT: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-RNN-NEAT algorithm in CS15-Low-71.

Failure Analysis CS15-Low-71 MOEvoRL-RNN-NEAT: Comprehensive failure analysis of the MOEvoRL-RNN-NEAT algorithm in CS15-Low-71.

Failure Classification CS15-Low-71 MODDPG: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MODDPG algorithm in CS15-Low-71.

Failure Analysis CS15-Low-71 MODDPG: Comprehensive failure analysis of the MODDPG algorithm in CS15-Low-71.

CS15-Med-71 Failure Analysis

Failure Classification CS15-Med-71 MOEvoRL-NSGA-II: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-NSGA-II algorithm in CS15-Med-71.

Failure Analysis CS15-Med-71 MOEvoRL-NSGA-II: Comprehensive failure analysis of the MOEvoRL-NSGA-II algorithm in CS15-Med-71.

Failure Classification CS15-Med-71 MOEvoRL-SPEA2: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-SPEA2 algorithm in CS15-Med-71.

Failure Analysis CS15-Med-71 MOEvoRL-SPEA2: Comprehensive failure analysis of the MOEvoRL-SPEA2 algorithm in CS15-Med-71.

Failure Classification CS15-Med-71 MOEvoRL-FF-NEAT: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-FF-NEAT algorithm in CS15-Med-71.

Failure Analysis CS15-Med-71 MOEvoRL-FF-NEAT: Comprehensive failure analysis of the MOEvoRL-FF-NEAT algorithm in CS15-Med-71.

Failure Classification CS15-Med-71 MOEvoRL-RNN-NEAT: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-RNN-NEAT algorithm in CS15-Med-71.

Failure Analysis CS15-Med-71 MOEvoRL-RNN-NEAT: Comprehensive failure analysis of the MOEvoRL-RNN-NEAT algorithm in CS15-Med-71.

Failure Classification CS15-Med-71 MODDPG: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MODDPG algorithm in CS15-Med-71.

Failure Analysis CS15-Med-71 MODDPG: Comprehensive failure analysis of the MODDPG algorithm in CS15-Med-71.

CS15-High-71 Failure Analysis

Failure Classification CS15-High-71 MOEvoRL-NSGA-II: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-NSGA-II algorithm in CS15-High-71.

Failure Analysis CS15-High-71 MOEvoRL-NSGA-II: Comprehensive failure analysis of the MOEvoRL-NSGA-II algorithm in CS15-High-71.

Failure Classification CS15-High-71 MOEvoRL-SPEA2: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-SPEA2 algorithm in CS15-High-71.

Failure Analysis CS15-High-71 MOEvoRL-SPEA2: Comprehensive failure analysis of the MOEvoRL-SPEA2 algorithm in CS15-High-71.

Failure Classification CS15-High-71 MOEvoRL-FF-NEAT: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-FF-NEAT algorithm in CS15-High-71.

Failure Analysis CS15-High-71 MOEvoRL-FF-NEAT: Comprehensive failure analysis of the MOEvoRL-FF-NEAT algorithm in CS15-High-71.

Failure Classification CS15-High-71 MOEvoRL-RNN-NEAT: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MOEvoRL-RNN-NEAT algorithm in CS15-High-71.

Failure Analysis CS15-High-71 MOEvoRL-RNN-NEAT: Comprehensive failure analysis of the MOEvoRL-RNN-NEAT algorithm in CS15-High-71.

Failure Classification CS15-High-71 MODDPG: Bar chart depicting the distribution of the three failure categories Phase Imbalance L1-L2, L1-L3, L2-L3 and Connection Node Overload for the MODDPG algorithm in CS15-High-71.

Failure Analysis CS15-High-71 MODDPG: Comprehensive failure analysis of the MODDPG algorithm in CS15-High-71.

Appendix: Violin Plots

Due to the large number of violin plots, these are not inserted directly into the README file, but only linked.

CSxx-Med-42 Violin Plots

• CS05-Med-42
• CS10-Med-42
• CS15-Med-42

CSxx-Med-AR-42 Violin Plots

• CS05-Med-AR-42
• CS10-Med-AR-42
• CS15-Med-AR-42

CS05-Low/Med/High-71 Violin Plots

• CS05-Low-71
• CS05-Med-71
• CS05-High-71

CS10-Low/Med/High-71 Violin Plots

• CS10-Low-71
• CS10-Med-71
• CS10-High-71

CS15-Low/Med/High-71 Violin Plots

• CS15-Low-71
• CS15-Med-71
• CS15-High-71