eprllib.Agents.Rewards

Reward functions in eprllib

The reward functions are called in the step() method in the environment iteratively for each agent present in the respective timestep.

This module contains the reward functions used in reinforcement learning and implemented in eprllib.

You can implement your own reward function by creating a new class that inherits from eprllib.Agents.Rewards.BaseReward and overriding the get_reward method.

The module includes the following classes:

• eprllib.Agents.Rewards.BaseReward : The base class for creating reward functions.

• eprllib.Agents.Rewards.RewardSpec : which is used to specify the configuration of reward functions.

These classes are used in the EnvironmentConfig class to specify the rewards that can be performed on the environment.

Modules

BaseReward	Reward Function
RewardSpec	Specification for agent reward functions This module defines the RewardSpec class, which is used to specify the configuration of reward functions for agents in reinforcement learning environments.