Loading ...
Sorry, an error occurred while loading the content.

787Re: [aima-talk] Re: Performance Measure

Expand Messages
  • Chuck Wolber
    Oct 2 12:20 AM
      On Tue, 2 Oct 2007, Ravi Mohan wrote:

      > > Which is correct? I emailed Ravi Mohan about that, but never received
      > > a response.
      > I had replied to you as soon as I saw your email (The reply is
      > duplicated below, Please note the date. Maybe your filters plonked it
      > into your spam folder? )

      Weird. You're right, I did receive it, at least my mail server logs say
      so, but it wasn't trapped in my spam folder nor do I recall seeing it in
      my inbox.

      > > Since the Vaccum cleaner environment referred to in question 2.7
      > > refers to the assumptions made on page 36, I am confused at how a
      > > reward of ten points was associated with cleaning a location?
      > See Section 3.2 (Example Problems) for where the path cost comes from. I
      > am not sure where I got the performance measure of +10, but it looks
      > reasonable to me.
      > Feel free to alter it to whatever makes sense :-). Any value that is nx
      > for n > 1 should work.

      Thanks, that explains a lot.

      I'm still curious as to what the value of setting the performance penalty
      to -1 where the reward is +1. It seems as if they'll simply cancel each
      other out. Or perhaps I am over thinking it and missing the fact that
      question 2.9 speaks from the standpoint of an agent that cannot possibly
      maximize its performance measure.


      "The idea that any one of us [presidential candidates] can bring about
      this change is a fantasy, it is not the truth! We need you to bring about
      the change on all these issues, we need you involved, we need you taking
      --John Edwards
    • Show all 5 messages in this topic