1
$\begingroup$

Assume i have x states each defined by an n-dimensional feature vector. In addition i have a set of actions which can be taken in each state resulting in a state action score. What would be an efficient way to derive the optimal actions for each state, optimising the score. Can this be expressed in terms of a genetic algorithm?

Knowledge:

There is no knowledge about underlying structure of states, actions and scores. Information is gathered through experiments.

Example:

State of patient defined by biomarkers, action taken: infuse drug with properties xzy, score to maximize: -1*(patients painlevel + nausea)

$\endgroup$
2
  • $\begingroup$ Do you have a model that says what new state results from taking an action in a given state? $\endgroup$ – Tom Minka Dec 12 '14 at 15:38
  • $\begingroup$ Hi, welcome to the site. You need to add much more information about what kind of information you have and what are you trying to solve. For example, it is unclear whether there is any uncertainty/randomness involve. Also, it is unclear whether there is any correspondence between the features and scores (or any other helpful structure). $\endgroup$ – Juho Kokkala Dec 12 '14 at 15:39

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.