Methods and apparatus for managing an environment within a domain

A computer implemented method for managing an environment within a domain, the environment being operable to perform a task, the method, performed by a distributed node, comprising:

• obtaining a representation of a current state of the environment;
• and
• for each of a plurality of possible actions that may be executed on the environment in its current state:
• using a Reinforcement Learning, RL, process to obtain predicted values of a reward function representing possible impacts of execution of the possible action on performance of the task by the environment;
• and
• obtaining a risk contour associated with the possible action;
• the method further comprising:
• selecting, from among the possible actions and on the basis of the obtained predicted values of the reward function and risk contours, an action for execution on the environment;
• and
• initiating execution of the selected action on the environment;
• wherein a risk contour associated with a possible action comprises:
• probabilities of the environment entering possible future states on execution of the action;
• confidence values associated with the probabilities;
• and
• for the possible future states, a representation of a domain level consequence of the environment entering the possible future state.