speedups in value and action

JuliaPOMDP · May 21, 2017 · 4a9bb82 · 4a9bb82
1 parent 34fcbc0
commit 4a9bb82
Showing 1 changed file with 21 additions and 26 deletions.
diff --git a/src/vanilla.jl b/src/vanilla.jl
@@ -66,41 +66,36 @@ alphas(policy::QMDPPolicy) = policy.alphas
 
 function action(policy::QMDPPolicy, b::DiscreteBelief)
     alphas = policy.alphas
-    ihi = 0
-    vhi = -Inf
     (ns, na) = size(alphas)
     @assert length(b.b) == ns "Length of belief and alpha-vector size mismatch"
-    for ai = 1:na
-        util = dot(alphas[:,ai], b.b)
-        if util > vhi
-            vhi = util
-            ihi = ai
-        end
-    end
+
+    util = alphas'*b.b
+    ihi = indmax(util)
     return policy.action_map[ihi]
 end
 
-function action(policy::QMDPPolicy, b)
-    bv = Array(Float64, n_states(policy.pomdp))
-    for (i,s) in enumerate(ordered_states(policy.pomdp))
-        bv[i] = pdf(b, s)
-    end
-    return action(policy, DiscreteBelief(bv))
-end
 
 function value(policy::QMDPPolicy, b::DiscreteBelief)
     alphas = policy.alphas
-    vhi = -Inf
     (ns, na) = size(alphas)
     @assert length(b.b) == ns "Length of belief and alpha-vector size mismatch"
-    for ai = 1:na
-        util = 0.0
-        for si = 1:length(b.b)
-            util += b.b[si] * alphas[si,ai]
-        end
-        if util > vhi
-            vhi = util
-        end
+
+    util = alphas'*b.b
+    return maximum(util)
+end
+
+function value(policy::QMDPPolicy, b)
+    return action(policy, DiscreteBelief(belief_vector(policy, b)))
+end
+
+function action(policy::QMDPPolicy, b)
+    return action(policy, DiscreteBelief(belief_vector(policy, b)))
+end
+
+function belief_vector(policy::QMDPPolicy, b)
+    bv = Array(Float64, n_states(policy.pomdp))
+    for (i,s) in enumerate(ordered_states(policy.pomdp))
+        bv[i] = pdf(b, s)
     end
-    return vhi
+    return bv
 end