Could you please explain the reason behind a line of code

[zhoujingzhe]

q = np.sum(np.multiply(z_concat, np.array(self.z)), axis=1)
what is the point in this code when you design which action is optimal?

[olavbm]

AFAIK, this line of code convolves the output distribution(z_concat) with the distribution-atoms(self.z).
Resulting in the Q-value for an action in a state.
As this is done in vectorized numpy code, the resulting variable, q, holds all q-values for all actions in all states in the batch.