In the current version of LBF an agent gets only one value as a reward after each step. If agents had several resources they could collect in the environment could we change the reward from a value to a vector, where each component of that vector represents the reward on that current objective (i.e. the amount of resources collected after executing a certain action) that an agent receives after each step?
In the current version of LBF an agent gets only one value as a reward after each step. If agents had several resources they could collect in the environment could we change the reward from a value to a vector, where each component of that vector represents the reward on that current objective (i.e. the amount of resources collected after executing a certain action) that an agent receives after each step?