Mercurial > hg > camir-aes2014

annotate toolboxes/FullBNT-1.0.7/bnt/potentials/@upot/upot_to_opt_policy.m @ 0:e9a9cd732c1e tip

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
first hg version after svn
author wolffd
date Tue, 10 Feb 2015 15:05:51 +0000
parents
children
rev   line source
wolffd@0 1 function [policy, EU] = upot_to_opt_policy(pot)
wolffd@0 2 % UPOT_TO_OPT_POLICY Compute an optimal deterministic policy given a utility potential
wolffd@0 3 % [policy, EU] = upot_to_opt_policy(pot)
wolffd@0 4 %
wolffd@0 5 % policy(a,b, ..., z) = P(do z | a, b, ..), which will be a delta function
wolffd@0 6 % EU is the contraction of this potential, i.e., P .* U
wolffd@0 7
wolffd@0 8 sz = pot.sizes; % mysize(pot.p);
wolffd@0 9 if isempty(sz)
wolffd@0 10 EU = pot.u;
wolffd@0 11 policy = [];
wolffd@0 12 return;
wolffd@0 13 end
wolffd@0 14
wolffd@0 15 parent_size = prod(sz(1:end-1));
wolffd@0 16 self_size = sz(end);
wolffd@0 17 C = pot.p .* pot.u; % contraction
wolffd@0 18 C = reshape(C, parent_size, self_size);
wolffd@0 19 policy = zeros(parent_size, self_size);
wolffd@0 20 for i=1:parent_size
wolffd@0 21 act = argmax(C(i,:));
wolffd@0 22 policy(i, act) = 1;
wolffd@0 23 end
wolffd@0 24 policy = myreshape(policy, sz);
wolffd@0 25 EU = sum(C(:));