csle_agents.constants package

Submodules

csle_agents.constants.constants module

Constants for csle-agents

class csle_agents.constants.constants.BAYESIAN_OPTIMIZATION

Bases: object

String constants related to Bayesian Optimization

L = 'L'

N = 'N'

OBJECTIVE_TYPE = 'objective_type'

PARAMETER_BOUNDS = 'parameter_bounds'

PARAMS = 'params'

POLICY_TYPE = 'policy_type'

STOP_DISTRIBUTION_ATTACKER = 'stop_distribution_attacker'

STOP_DISTRIBUTION_DEFENDER = 'stop_distribution_defender'

TARGET = 'target'

THETA1 = 'theta1'

THETAS = 'thetas'

THRESHOLDS = 'thresholds'

UCB = 'ucb'

UCB_KAPPA = 'ucb_kappa'

UCB_XI = 'ucb_xi'

UTILITY_FUNCTION = 'utility_function'

class csle_agents.constants.constants.BAYESIAN_OPTIMIZATION_EMUKIT

Bases: object

String constants related to Bayesian Optimization Emukit

ACQUISITION_FUNCTION_TYPE = 'acquisition_function_type'

ACQUISITION_OPTIMIZER_TYPE = 'acquisition_optimizer_type'

BETA = 'beta'

EVALUATION_BUDGET = 'evaluation_budget'

INPUT_SPACE_DIM = 'input_space_dim'

KERNEL_TYPE = 'kernel_type'

LENGTHSCALE_RBF_KERNEL = 'lengthscale_rbf_kernel'

OBJECTIVE_TYPE = 'objective_type'

OBS_LIKELIHOOD_VARIANCE = 'obs_likelihood_variance'

PARAMS = 'params'

POLICY_TYPE = 'policy_type'

VARIANCE_RBF_KERNEL = 'variance_rbf_kernel'

X_init = 'X_init'

Y_init = 'Y_init'

class csle_agents.constants.constants.C51_CLEAN

Bases: object

String constants related to the C51-clean agent

ANNEAL_LR = 'anneal_lr'

BUFFER_SIZE = 'buffer_size'

CLIP_VLOSS = 'clip_vloss'

CUDA = 'cuda'

END_EXPLORATION_RATE = 'end_exploration_rate'

EXP_FRAC = 'exploration_fraction'

LEARNING_STARTS = 'learning_starts'

MINIBATCH_SIZE = 'minibatch_size'

NORM_ADV = 'norm_adv'

NUM_ENVS = 'num_envs'

NUM_STEPS = 'num_steps'

N_ATOMS = 'n_atoms'

REWARD_SCALER = 'reward_scaler'

SAVE_MODEL = 'save_model'

START_EXPLORATION_RATE = 'start_exploration_rate'

STEPS_BETWEEN_UPDATES = 'steps_between_updates'

TAU = 'tau'

TRAIN_FREQ = 'train_frequency'

T_N_FREQ = 'target_network_frequency'

V_MAX = 'v_max'

V_MIN = 'v_min'

class csle_agents.constants.constants.CMA_ES_OPTIMIZATION

Bases: object

String constants related to CMA-ES Optimization

L = 'L'

N = 'N'

OBJECTIVE_TYPE = 'objective_type'

PARAMETER_BOUNDS = 'parameter_bounds'

PARAMS = 'params'

POLICY_TYPE = 'policy_type'

STOP_DISTRIBUTION_ATTACKER = 'stop_distribution_attacker'

STOP_DISTRIBUTION_DEFENDER = 'stop_distribution_defender'

TARGET = 'target'

THETA1 = 'theta1'

THETAS = 'thetas'

THRESHOLDS = 'thresholds'

UCB = 'ucb'

UCB_KAPPA = 'ucb_kappa'

UCB_XI = 'ucb_xi'

UTILITY_FUNCTION = 'utility_function'

class csle_agents.constants.constants.COMMON

Bases: object

Common string constants among all agents

ADAM = 'Adam'

AVERAGE_ATTACKER_RETURN = 'average_attacker_return'

AVERAGE_DEFENDER_RETURN = 'average_defender_return'

AVERAGE_HEURISTIC_RETURN = 'average_heuristic_return'

AVERAGE_RANDOM_RETURN = 'average_random_return'

AVERAGE_RETURN = 'average_return'

AVERAGE_TIME_HORIZON = 'average_time_horizon'

AVERAGE_UPPER_BOUND_RETURN = 'average_upper_bound_return'

BASELINE_PREFIX = 'baseline_'

BATCH_SIZE = 'batch_size'

CONFIDENCE_INTERVAL = 'confidence_interval'

EVALUATE_WITH_DISCOUNT = 'evaluate_with_discount'

EVAL_BATCH_SIZE = 'eval_batch_size'

EVAL_EVERY = 'eval_every'

EVAL_PREFIX = 'eval_'

EXPLOITABILITY = 'exploitability'

GAMMA = 'gamma'

L = 'L'

LEARNING_RATE = 'learning_rate'

LEARNING_RATE_DECAY_RATE = 'learning_rate_decay_rate'

LEARNING_RATE_EXP_DECAY = 'learning_rate_exp_decay'

MAX_ENV_STEPS = 'max_env_steps'

NUM_CACHED_SIMULATION_TRACES = 3

NUM_NODES = 'num_nodes'

NUM_PARALLEL_ENVS = 'num_parallel_envs'

NUM_TRAINING_TIMESTEPS = 'num_training_timesteps'

OBSERVATION = 'o'

OPTIMIZER = 'optimizer'

POLICY_LOSSES = 'policy_losses'

RUNNING_AVERAGE = 'running_average'

RUNNING_AVERAGE_ATTACKER_RETURN = 'running_average_attacker_return'

RUNNING_AVERAGE_DEFENDER_RETURN = 'running_average_defender_return'

RUNNING_AVERAGE_EXPLOITABILITY = 'running_average_exploitability'

RUNNING_AVERAGE_INTRUSION_LENGTH = 'running_average_intrusion_length'

RUNNING_AVERAGE_INTRUSION_START = 'running_average_intrusion_start'

RUNNING_AVERAGE_RETURN = 'running_average_return'

RUNNING_AVERAGE_START_POINT_CORRECT = 'running_average_start_point_correct'

RUNNING_AVERAGE_TIME_HORIZON = 'running_average_time_horizon'

RUNNING_AVERAGE_WEIGHTED_INTRUSION_PREDICTION_DISTANCE = 'running_average_weighted_intrusion_prediction_distance'

RUNTIME = 'runtime'

SAVE_EVERY = 'save_every'

SGD = 'SGD'

START_POINT_CORRECT = 'start_point_correct'

STATE = 's'

STOPPING_ENVS = ['csle-stopping-game-v1', 'csle-stopping-game-mdp-attacker-v1', 'csle-stopping-game-pomdp-defender-v1']

WEIGHTED_INTRUSION_PREDICTION_DISTANCE = 'weighted_intrusion_prediction_distance'

class csle_agents.constants.constants.CROSS_ENTROPY

Bases: object

String constants related to the cross-entropy

K = 'K'

L = 'L'

LAMB = 'lamb'

N = 'N'

OBJECTIVE_TYPE = 'objective_type'

POLICY_TYPE = 'policy_type'

STOP_DISTRIBUTION_ATTACKER = 'stop_distribution_attacker'

STOP_DISTRIBUTION_DEFENDER = 'stop_distribution_defender'

THETA1 = 'theta1'

THETAS = 'thetas'

THRESHOLDS = 'thresholds'

class csle_agents.constants.constants.DIFFERENTIAL_EVOLUTION

Bases: object

String constants related to differential evolution

BOUNDS = 'bounds'

L = 'L'

MUTATE = 'mutate'

N = 'N'

OBJECTIVE_TYPE = 'objective_type'

POLICY_TYPE = 'policy_type'

POPULATION_SIZE = 'population_size'

RECOMBINATION = 'recombination'

STOP_DISTRIBUTION_ATTACKER = 'stop_distribution_attacker'

STOP_DISTRIBUTION_DEFENDER = 'stop_distribution_defender'

THETA1 = 'theta1'

THETAS = 'thetas'

THRESHOLDS = 'thresholds'

class csle_agents.constants.constants.DQN

Bases: object

String constants related to DQN

BUFFER_SIZE = 'buffer_size'

DQN_BATCH_SIZE = 'dqn_batch_size'

EXPLORATION_FINAL_EPS = 'exploration_final_eps'

EXPLORATION_FRACTION = 'exploration_fracion'

EXPLORATION_INITIAL_EPS = 'exploration_initial_eps'

GRADIENT_STEPS = 'gradient_steps'

LEARNING_STARTS = 'learning_starts'

MAX_GRAD_NORM = 'max_grad_norm'

MLP_POLICY = 'MlpPolicy'

N_EPISODES_ROLLOUT = 'n_episodes_rollout'

TARGET_UPDATE_INTERVAL = 'target_update_interval'

TRAIN_FREQ = 'train_freq'

class csle_agents.constants.constants.DQN_CLEAN

Bases: object

String constants related to the DQN-Clean agent

ANNEAL_LR = 'anneal_lr'

BUFFER_SIZE = 'buffer_size'

CLIP_RANGE = 'clip_range'

CLIP_RANGE_VF = 'clip_range_vf'

CLIP_VLOSS = 'clip_vloss'

CUDA = 'cuda'

ENT_COEF = 'ent_coef'

EXP_FRAC = 'exploration_fraction'

GAE_LAMBDA = 'gae_lambda'

LEARNING_STARTS = 'learning_starts'

MAX_GRAD_NORM = 'max_grad_norm'

MINIBATCH_SIZE = 'minibatch_size'

MLP_POLICY = 'MlpPolicy'

NORM_ADV = 'norm_adv'

NUM_ENVS = 'num_envs'

NUM_MINIBATCHES = 'num_minibatches'

NUM_STEPS = 'num_steps'

REWARD_SCALER = 'reward_scaler'

SAVE_MODEL = 'save_model'

STEPS_BETWEEN_UPDATES = 'steps_between_updates'

TARGET_KL = 'target_kl'

TAU = 'tau'

TRAIN_FREQ = 'train_frequency'

T_N_FREQ = 'target_network_frequency'

UPDATE_EPOCHS = 'update_epochs'

VF_COEF = 'vf_coef'

class csle_agents.constants.constants.DYNASEC

Bases: object

String constants related to the DynaSec algorithm

CLIENTS_ARRIVAL_RATE = 'clients_arrival_rate'

EMULATION_MONITOR_SLEEP_TIME = 'emulation_monitor_sleep_time'

EMULATION_TRACES_TO_SAVE_W_DATA_COLLECTION_JOB = 'emulation_traces_to_save_with_data_collection_job'

INTRUSION_ALERTS_MEAN = 'intrusion_alerts_mean'

INTRUSION_ALERTS_MEAN_BASELINE = 'intrusion_alerts_mean_baseline'

INTRUSION_START_P = 'intrusion_start_p'

NO_INTRUSION_ALERTS_MEAN = 'no_intrusion_alerts_mean'

NO_INTRUSION_ALERTS_MEAN_BASELINE = 'no_intrusion_alerts_mean_baseline'

NUM_CLIENTS = 'num_clients'

REPLAY_WINDOW_SIZE = 'replay_window_size'

SLEEP_TIME = 'sleep_time'

STATIC_ATTACKER_TYPE = 'static_attacker_type'

TRAINING_EPOCHS = 'training_epochs'

WARMUP_EPISODES = 'warmup_episodes'

class csle_agents.constants.constants.ENV_METRICS

Bases: object

String constants representing environment metrics

ATTACKER_ACTION = 'a2'

AVERAGE_HEURISTIC_RETURN = 'average_heuristic_return'

AVERAGE_RANDOM_RETURN = 'average_random_return'

AVERAGE_UPPER_BOUND_RETURN = 'average_upper_bound_return'

DEFENDER_ACTION = 'a1'

OBSERVATION = 'o'

RETURN = 'R'

STATE = 's'

TIME_HORIZON = 'T'

TIME_STEP = 't'

class csle_agents.constants.constants.FICTITIOUS_PLAY

Bases: object

String constants related to fictitious play

N = 'N'

PAYOFF_MATRIX = 'payoff_matrix'

PLAYER_1_PRIOR = 'player_1_prior'

PLAYER_2_PRIOR = 'player_2_prior'

class csle_agents.constants.constants.HSVI

Bases: object

String constants related to HSVI

ACTION_SPACE = 'action_space'

EPSILON = 'epsilon'

INITIAL_BELIEF = 'initial_belief'

INITIAL_BELIEF_VALUES = 'initial_belief_values'

LB_SIZE = 'lower_bound_size'

LB_SIZES = 'lower_bound_sizes'

NUMBER_OF_SIMULATIONS = 'NUMBER_OF_SIMULATIONS'

OBSERVATION_SPACE = 'observation_space'

OBSERVATION_TENSOR = 'observation_tensor'

PRUNE_FREQUENCY = 'prune_frequency'

REWARD_TENSOR = 'reward_tensor'

SIMULATE_HORIZON = 'SIMULATE_HORIZON'

SIMULATION_FREQUENCY = 'simulation_frequency'

STATE_SPACE = 'state_space'

TRANSITION_TENSOR = 'transition_tensor'

UB_SIZE = 'upper_bound_size'

UB_SIZES = 'upper_bound_sizes'

USE_LP = 'use_lp'

WIDTH = 'width'

WIDTHS = 'widths'

class csle_agents.constants.constants.HSVI_OS_POSG

Bases: object

String constants related to the HSVI algorithm for OS-POSGs

ACTION_SPACE_PLAYER_1 = 'action_space_player_1'

ACTION_SPACE_PLAYER_2 = 'action_space_player_2'

EPSILON = 'epsilon'

EXCESSES = 'excesses'

INITIAL_BELIEF = 'initial_belief'

N = 'N'

OBSERVATION_FUNCTION = 'observation_function'

OBSERVATION_SPACE = 'observation_space'

PRUNE_FREQUENCY = 'prune_frequency'

REWARD_TENSOR = 'reward_tensor'

STATE_SPACE = 'state_space'

TRANSITION_TENSOR = 'transition_tensor'

WIDTHS = 'widths'

class csle_agents.constants.constants.KIEFER_WOLFOWITZ

Bases: object

String constants related to Kiefer-Wolfowitz

DELTA = 'delta'

GRADIENT_BATCH_SIZE = 'gradient_batch_size'

INITIAL_ALPHA = 'initial_alpha'

L = 'L'

N = 'N'

POLICY_TYPE = 'policy_type'

STOP_DISTRIBUTION_ATTACKER = 'stop_distribution_attacker'

STOP_DISTRIBUTION_DEFENDER = 'stop_distribution_defender'

THETA1 = 'theta1'

THETAS = 'thetas'

THRESHOLDS = 'thresholds'

class csle_agents.constants.constants.LOCAL_DFSP

Bases: object

String constants related to the local DFSP algorithm

AVERAGE_BEST_RESPONSE_ATTACKER_RETURN = 'average_best_response_attacker_return'

AVERAGE_BEST_RESPONSE_DEFENDER_RETURN = 'average_best_response_defender_return'

BEST_RESPONSE_EVALUATION_ITERATIONS = 'best_response_evaluation_iterations'

EQUILIBRIUM_STRATEGIES_EVALUATION_ITERATIONS = 'equilibrium_strategies_evaluation_iterations'

N_2 = 'N_2'

RUNNING_AVERAGE_BEST_RESPONSE_ATTACKER_RETURN = 'running_average_best_response_attacker_return'

RUNNING_AVERAGE_BEST_RESPONSE_DEFENDER_RETURN = 'running_average_best_response_defender_return'

class csle_agents.constants.constants.LP_FOR_CMDPs

Bases: object

String constants related to linear programming for CMDPs

ACTIONS = 'actions'

CONSTRAINT_COST_TENSORS = 'constraint_cost_tensors'

CONSTRAINT_COST_THRESHOLDS = 'constraint_cost_thresholds'

COST_TENSOR = 'cost_tensor'

STATES = 'states'

TRANSITION_TENSOR = 'transition_tensor'

class csle_agents.constants.constants.LP_FOR_NF_GAMES

Bases: object

String constants related to linear programming for normal-form games

ACTION_SPACE_PLAYER_1 = 'action_space_player_1'

ACTION_SPACE_PLAYER_2 = 'action_space_player_2'

N = 'N'

PAYOFF_MATRIX = 'payoff_matrix'

class csle_agents.constants.constants.MCS

Bases: object

String constants related to Multilevel Coordinate Search

EPSILON = 'epsilon'

GAMMA = 'gamma'

IINIT = 'iinit'

L = 'L'

LOCAL = 'local'

M = 'm'

NF = 'nf'

OBJECTIVE_TYPE = 'objective_type'

POLICY_TYPE = 'policy_type'

PRT = 'prt'

SMAX = 'smax'

STEP = 'step'

STEP1 = 'step1'

STOP = 'stop'

STOPPING_ACTIONS = 'stopping_actions'

STOP_DISTRIBUTION_DEFENDER = 'stop_distribution_defender'

THETAS = 'thetas'

THRESHOLDS = 'thresholds'

U = 'u'

V = 'v'

class csle_agents.constants.constants.NELDER_MEAD

Bases: object

String constants related to Nelder-Mead

COOLING_FACTOR = 'cooling_factor'

DELTA = 'delta'

IMPROVE_BREAK = 'improve_break'

IMPROVE_THRESHOLD = 'improve_threshold'

INITIAL_TEMPERATURE = 'initial_temperature'

L = 'L'

N = 'N'

OBJECTIVE_TYPE = 'objective_type'

POLICY_TYPE = 'policy_type'

STEP = 'step'

STOP_DISTRIBUTION_ATTACKER = 'stop_distribution_attacker'

STOP_DISTRIBUTION_DEFENDER = 'stop_distribution_defender'

THETA1 = 'theta1'

THETAS = 'thetas'

THRESHOLDS = 'thresholds'

UCB = 'ucb'

UTILITY_FUNCTION = 'utility_function'

class csle_agents.constants.constants.PARTICLE_SWARM

Bases: object

String constants related to simulated annealing

B_LOW = 'b_lo'

B_UP = 'b_up'

COGNITIVE_COEFFICIENT = 'Phi_p'

DELTA = 'delta'

INERTIA_WEIGHT = 'w'

L = 'L'

N = 'N'

OBJECTIVE_TYPE = 'objective_type'

POLICY_TYPE = 'policy_type'

S = 'S'

SOCIAL_COEFFICIENT = 'Phi_g'

STOP_DISTRIBUTION_ATTACKER = 'stop_distribution_attacker'

STOP_DISTRIBUTION_DEFENDER = 'stop_distribution_defender'

THETA1 = 'theta1'

THETAS = 'thetas'

THRESHOLDS = 'thresholds'

class csle_agents.constants.constants.PI

Bases: object

String constants related to PI

INITIAL_POLICY = 'initial_policy'

N = 'N'

NUM_ACTIONS = 'num_actions'

NUM_STATES = 'num_states'

REWARD_TENSOR = 'reward_tensor'

TRANSITION_TENSOR = 'transition_tensor'

class csle_agents.constants.constants.POMCP

Bases: object

String constants related to POMCP

A = 'A'

ACQUISITION_FUNCTION_TYPE = 'acquisition_function_type'

C = 'c'

C2 = 'c2'

DEFAULT_NODE_VALUE = 'default_node_value'

EVAL_ENV_CONFIG = 'eval_env_config'

EVAL_ENV_NAME = 'eval_env_name'

GAMMA = 'gamma'

INITIAL_PARTICLES = 'initial_particles'

LOG_STEP_FREQUENCY = 'log_step_frequency'

MAX_NEGATIVE_SAMPLES = 'max_negative_samples'

MAX_PARTICLES = 'max_particles'

MAX_PLANNING_DEPTH = 'max_planning_depth'

MAX_ROLLOUT_DEPTH = 'max_rollout_depth'

N = 'N'

NUM_EVALS_PER_PROCESS = 'num_evals_per_process'

NUM_PARALLEL_PROCESSES = 'num_parallel_processes'

O = 'O'

OBJECTIVE_TYPE = 'objective_type'

PARALLEL_ROLLOUT = 'parallel_rollout'

PLANNING_TIME = 'planning_time'

PRIOR_CONFIDENCE = 'prior_confidence'

PRIOR_WEIGHT = 'prior_weight'

PRUNE_ACTION_SPACE = 'prune_action_space'

PRUNE_SIZE = 'prune_size'

REINVIGORATED_PARTICLES_RATIO = 'reinvigorated_particles_ratio'

REINVIGORATION = 'reinvigoration'

ROLLOUT_POLICY = 'rollout_policy'

S = 'S'

USE_ROLLOUT_POLICY = 'use_rollout_policy'

VALUE_FUNCTION = 'value_function'

VERBOSE = 'verbose'

class csle_agents.constants.constants.PPG_CLEAN

Bases: object

String constants related to PPG clean

ADV_NORM_FULLBATCH = 'adv_norm_fullbatch'

ANNEAL_LR = 'anneal_lr'

AUX_BATCH_ROLLOUTS = 'aux_batch_rollouts'

BATCH_SIZE = 'batch_size'

BETA_CLONE = 'beta_clone'

CLIP_COEF = 'clip_coef'

CLIP_VLOSS = 'clip_vloss'

ENT_COEF = 'ent_coef'

E_AUXILIARY = 'e_auxiliary'

E_POLICY = 'e_policy'

GAE_LAMBDA = 'gae_lambda'

GAMMA = 'gamma'

LEARNING_RATE = 'learning_rate'

MAX_GRAD_NORM = 'max_grad_norm'

MINIBATCH_SIZE = 'minibatch_size'

NUM_AUX_GRAD_ACCUM = 'num_aux_grad_accum'

NUM_AUX_ROLLOUTS = 'num_aux_rollouts'

NUM_ITERATIONS = 'num_iterations'

NUM_MINIBATCHES = 'num_minibatches'

NUM_PHASES = 'num_phases'

NUM_STEPS = 'num_steps'

N_ITERATION = 'n_iteration'

TARGET_KL = 'target_kl'

TOTAL_STEPS = 'total_steps'

VF_COEF = 'vf_coef'

V_VALUE = 'v_value'

class csle_agents.constants.constants.PPO

Bases: object

String constants related to PPO

CLIP_RANGE = 'clip_range'

CLIP_RANGE_VF = 'clip_range_vf'

ENT_COEF = 'ent_coef'

GAE_LAMBDA = 'gae_lambda'

MAX_GRAD_NORM = 'max_grad_norm'

MLP_POLICY = 'MlpPolicy'

NUM_GRADIENT_STEPS = 'num_gradient_steps'

STEPS_BETWEEN_UPDATES = 'steps_between_updates'

TARGET_KL = 'target_kl'

VF_COEF = 'vf_coef'

class csle_agents.constants.constants.PPO_CLEAN

Bases: object

String constants related to PPO clean

ANNEAL_LR = 'anneal_lr'

CLIP_RANGE = 'clip_range'

CLIP_RANGE_VF = 'clip_range_vf'

CLIP_VLOSS = 'clip_vloss'

CUDA = 'cuda'

ENT_COEF = 'ent_coef'

GAE_LAMBDA = 'gae_lambda'

MAX_GRAD_NORM = 'max_grad_norm'

MINIBATCH_SIZE = 'minibatch_size'

MLP_POLICY = 'MlpPolicy'

NORM_ADV = 'norm_adv'

NUM_ENVS = 'num_envs'

NUM_MINIBATCHES = 'num_minibatches'

NUM_STEPS = 'num_steps'

REWARD_SCALER = 'reward_scaler'

STEPS_BETWEEN_UPDATES = 'steps_between_updates'

TARGET_KL = 'target_kl'

UPDATE_EPOCHS = 'update_epochs'

VF_COEF = 'vf_coef'

class csle_agents.constants.constants.Q_LEARNING

Bases: object

String constants related to Q-learning

A = 'A'

EPSILON = 'epsilon'

EPSILON_DECAY_RATE = 'epsilon_decay_rate'

INITIAL_STATE_VALUES = 'initial_state_values'

N = 'N'

S = 'S'

class csle_agents.constants.constants.RANDOM_SEARCH

Bases: object

String constants related to random search

DELTA = 'delta'

L = 'L'

N = 'N'

OBJECTIVE_TYPE = 'objective_type'

POLICY_TYPE = 'policy_type'

STOP_DISTRIBUTION_ATTACKER = 'stop_distribution_attacker'

STOP_DISTRIBUTION_DEFENDER = 'stop_distribution_defender'

THETA1 = 'theta1'

THETAS = 'thetas'

THRESHOLDS = 'thresholds'

class csle_agents.constants.constants.REINFORCE

Bases: object

String constants related to REINFORCE

CLIP_GRADIENT = 'clip_gradient'

GRADIENT_BATCH_SIZE = 'gradient_batch_size'

N = 'N'

class csle_agents.constants.constants.SARSA

Bases: object

String constants related to Sarsa

A = 'A'

EPSILON = 'epsilon'

INITIAL_STATE_VALUES = 'initial_state_values'

N = 'N'

S = 'S'

class csle_agents.constants.constants.SHAPLEY_ITERATION

Bases: object

String constants related to Shapley Iteration

ACTION_SPACE_PLAYER_1 = 'action_space_player_1'

ACTION_SPACE_PLAYER_2 = 'action_space_player_2'

DELTA = 'delta'

N = 'N'

REWARD_TENSOR = 'reward_tensor'

STATE_SPACE = 'state_space'

TRANSITION_TENSOR = 'transition_tensor'

class csle_agents.constants.constants.SIMULATED_ANNEALING

Bases: object

String constants related to simulated annealing

COOLING_FACTOR = 'cooling_factor'

DELTA = 'delta'

INITIAL_TEMPERATURE = 'initial_temperature'

L = 'L'

N = 'N'

OBJECTIVE_TYPE = 'objective_type'

POLICY_TYPE = 'policy_type'

STOP_DISTRIBUTION_ATTACKER = 'stop_distribution_attacker'

STOP_DISTRIBUTION_DEFENDER = 'stop_distribution_defender'

THETA1 = 'theta1'

THETAS = 'thetas'

THRESHOLDS = 'thresholds'

class csle_agents.constants.constants.SONDIK_VI

Bases: object

String constants related to Sondik’s VI

ACTION_SPACE = 'action_space'

INITIAL_BELIEF = 'initial_belief'

INITIAL_BELIEF_VALUES = 'initial_belief_values'

NUM_ALPHA_VECTORS = 'num_alpha_vectors'

OBSERVATION_SPACE = 'observation_space'

OBSERVATION_TENSOR = 'observation_tensor'

PLANNING_HORIZON = 'planning_horizon'

REWARD_TENSOR = 'reward_tensor'

STATE_SPACE = 'state_space'

TRANSITION_TENSOR = 'transition_tensor'

USE_PRUNING = 'use_pruning'

class csle_agents.constants.constants.T_FP

Bases: object

String constants related to T-FP

ATTACKER_THRESHOLDS = 'attacker_thresholds'

AVERAGE_BEST_RESPONSE_ATTACKER_RETURN = 'average_best_response_attacker_return'

AVERAGE_BEST_RESPONSE_DEFENDER_RETURN = 'average_best_response_defender_return'

BEST_RESPONSE_EVALUATION_ITERATIONS = 'best_response_evaluation_iterations'

DEFENDER_THRESHOLDS = 'defender_thresholds'

EQUILIBRIUM_STRATEGIES_EVALUATION_ITERATIONS = 'equilibrium_strategies_evaluation_iterations'

N_2 = 'N_2'

RUNNING_AVERAGE_BEST_RESPONSE_ATTACKER_RETURN = 'running_average_best_response_attacker_return'

RUNNING_AVERAGE_BEST_RESPONSE_DEFENDER_RETURN = 'running_average_best_response_defender_return'

THETA1_ATTACKER = 'theta1_attacker'

THETA1_DEFENDER = 'theta1_defender'

class csle_agents.constants.constants.VI

Bases: object

String constants related to VI

DELTA = 'delta'

NUM_ACTIONS = 'num_actions'

NUM_STATES = 'num_states'

REWARD_TENSOR = 'reward_tensor'

THETA = 'theta'

TRANSITION_TENSOR = 'transition_tensor'

Module contents