OK, I have posted the final project handout. Let me know if you have questions.
We will not be meeting during the last week. But, I will be in my office during class time in case you want to talk about your final project.
In case I did not mention it in class, Test 2 will be based solely on the rest of the book chapters. That is, it does not cover the papers I presented in class.
I have posted PS 3 and also decided to eliminate PS 4. Instead, the first three problem sets will count for 20% of the grade.
I've had to cancel today's class. Sorry for the inconvenience.
I have updated the PS 2 handout to explain a bit better how the value function is to be calculated.
I note that none of you included your email in the PS 1 handin. If I don't have your email already, please send it to me so I can send you your grade and comments back. Thanks.
I have just posted the next problem set. Let me know if it needs further explanations.
When building an MDP, the probability of all the links from a node must add up to 1. We also want a slider that will let us change some characteristics of these links. One way to do this is by using one of the built-in probability distribution functions and linking its standard deviation to a slider. Then, to pick the probabilities for x links you start by picking one number for each link from that same probability distributions function. Then, you normalize all these numbers (that is, divide each one by the sum of all numbers) so that they add up to 1.
I have accumulated my notes from past years and put them together into something that is not entirely unlike a textbook. I am hoping that this class will help me find all the bugs in it so that we will end up with a first draft.
The book will be available as a PDF for students of the class. As you will see, it covers the fundamentals of game theory, Economics, and AI that are needed by all multiagent practitioners. It also incorporates my many netlogo models. These are necessary to demonstrate the often non-intuitive dynamics of multiagent interaction.