q_values/value_vector function

[MaximeBouton]

I think it would be useful to have a function returning the q values at a given state.
The signature could be q_values(policy, s) or value_vector(policy, s).

Any thoughts?

[mykelk]

Here is one perspective.

[MaximeBouton]

It seems like action_values(policy, s) would be a more appropriate name.

[zsunberg]

Seems reasonable to me.

…

On Mon, Jul 2, 2018 at 7:21 AM Maxime ***@***.***> wrote: It seems like action_values(policy, s) would be a more appropriate name. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#73 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AEC0a4J5F99Gf3Rg6ufnGdx_6vr3BOS5ks5uCixygaJpZM4U_STr> .

[zsunberg]

what order would these be in? consistent with iterator(actions(mdp, s)) or ordered_actions(mdp)?

[MaximeBouton]

I vote ordered_actions(mdp), it makes it clearer I think.

[zsunberg]

What if not all the actions are in `actions(mdp, s)`? are the entries for the missing actions undefined?

…

On Thu, Jul 26, 2018 at 9:48 PM Maxime ***@***.***> wrote: I vote ordered_actions(mdp), it makes it clearer I think. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#73 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AEC0azzZNJtsOsKj3WjuH99a7qYjWS_Tks5uKpu1gaJpZM4U_STr> .