Utils

This notebook gathers multiple functions used as helpers in the library.

Helpers for the environments

 isBetween_c_Vec (a, b, c, r)

Checks whether point c is crossing the line formed with point a and b.

	Type	Details
a	tensor, shape = (1,2)	Previous position.
b	tensor, shape = (1,2)	Current position.
c	tensor, shape = (Nt,2)	Positions of all targets.
r	int/float	Target radius.
Returns	array of boolean values	True at the indices of found targets.

 coord_mod (coord1, coord2, mod)

Computes the distance difference between two coordinates, in a world with size ‘mod’ and periodic boundary conditions.

	Type	Details
coord1	value, np.array, tensor (can be shape=(n,1))	First coordinate.
coord2	np.array, tensor – shape=(1,1)	Second coordinate, substracted from coord1.
mod	int	World size.
Returns	float	Distance difference (with correct sign, not absolute value).

 isBetween_c_Vec_nAgents (a, b, c, r)

Checks whether point c is crossing the line formed with point a and b. Code to run several agents in parallel.

	Type	Details
a	array, shape = (n,2)	Previous position of all n agents.
b	array, shape = (n,2)	Current position of all n agents.
c	array, shape = (Nt,2)	Positions of all targets.
r	float	Target radius
Returns	array of boolean values, shape = (Nt, n)	True at the indices of found targets.

 get_encounters (agent_previous_pos, agent_pos, target_positions, L, r)

Considering the agent walks, it checks whether the agent finds a target while walking the current step. Code to run several agents in parallel.

	Type	Details
agent_previous_pos	array, shape = (n,2)	Position of the n agents before taking the step.
agent_pos	array, shape = (n,2)	Position of the n agents.
target_positions	array, shape = (Nt,2)	Positions of the targets.
L	int	World size.
r	float	Radius of the targets.
Returns	array of boolean values, shape = (Nt, n)	True at the indices of found targets.

 get_config (config, config_path='configurations/learning/')

Get the configuration file of the given experiment and config name (e.g. exp_0).

	Type	Default	Details
config	str		Config name (e.g. exp_0)
config_path	str	configurations/learning/	path to configurations
Returns	dict		Dictionary with the parameters of the loaded configuration.

 get_policy (results_path, agent, episode)

Gets the policy of an agent at a given episode.

	Type	Details
results_path	str	Path of the folder from which to extract the data.
agent	int	Agent index.
episode	int	Episode.
Returns	list	Policy.

 get_performance (results_path, agent_list, episode_list)

Extract data with the efficiencies obtained in the postlearning analysis.

	Type	Details
results_path	str	Path of the folder from which to extract the data.
agent_list	list	List with the agent indices.
episode_list	list	List with the episodes.
Returns	np.array, shape=(len(agent_list), len(episode_list))	Average performances obtained by the agents in the postlearning analysis.

 get_opt (path, df)

Get the highest efficiency obtained by the benchmark models and the corresponding parameters.

	Type	Details
path	str	Path from which to get the data.
df	panda dataframe	Dataframe with the results from the optimization with Tune.
Returns	list	Efficiency of each walk.