Select what applies to setup a reinforcement learning based mathematical model or description that characterizes how a robot can be taught to walk.

it is a physics based model.

the robot model will produce an action to be taken based on the state input. it models a reward as per action that maximizes chances of success or of achieving goal.

the robot stands up$,$ the learning model rewards; the robot falls down, the model penalizes.

it does not involve an optimization method using an "objective function". the robot model will produce a reward based on the action taken.