Hi Wei, I am still thinking on books, prerequisites, etc.
I think the following paper: Hardegree, G. M. (1976). The Conditional in Quantum Logic. In Suppes, P., editor, Logic and Probability in Quantum Mechanics, volume 78 of Synthese Library, pages 55-72. D. Reidel Publishing Company, Dordrecht-Holland. could be at the intersection of your interest in decision theory, and the comp derivation of physics. At least both could make some use of the Stalnaker - Lewis "other worlds" type of semantics. Unfortunately the Hardegree-Stalnaker-Lewis counterfactuals appears in my work through the use of the movie-graph argument (MGA), (cf Maudlin's argument we talk some time ago)---as opposed to the UDA more easily understandable by those open to the everything-list's "more is simpler" principle. Just to tell you that the Joyce book you refer to us is indeed interesting and could motivate for mathematical tools common in decision theory, philosophical logic, and theories related to the machine interview I am engaged in. Note that the "Stalnaker Hardegree Lewis semantics" belongs to the family of modal semantics described in the last chapters of Chellas' book (Minimal Model, Conditional). You could perhaps even bypass Kripke semantics, and read those last chapters in priority. So, if you study Chellas last chapter, that will help you both to deepen your appreciation of Joyce, and will help for the understanding of my thesis. Take a look on http://www.phil.mq.edu.au/isl/, the web page on Greg Restall book on "substructural logics". I have not yet read that book, but that book makes an unified treatment of many "weak" logics (in particular those you get by weakening the structural rules). It gives an intro to the "Goldblatt frames" for non distributive logic(like quantum logic), to the linear logics and the coherence conditions. Giving the importance of resource in causalist decision theory I would bet on the relevance of such logics. He makes also a little intro to the use of category in those logics, but not enough (imo!). Another defect is that he changes the notation in linear logic, which is confusing. Bruno PS I have found a way to explain with knot theory what "logic" is, as a branch of math, by comparing propositions with knots, proofs with continuous deformation, and semantics with knot's invariants. As I said before one of the difficulty for writing a paper is the misunderstanding between logicians and physicist ...