update

src/mcts.jl

@@ -349,7 +349,7 @@ function selectChildNode(node::MCTSNode)::MCTSNode
   # loop thought node children dictionary to find the highest progress value
   for (k, childNode) in node.children
     if childNode.statevalue > highestProgressValue
-      highestProgressValue = childNode.statevalue
+      highestProgressValue = childNode.statevalue + childNode.reward
       nodekey = childNode.nodekey
     end
   end
@@ -402,7 +402,9 @@ isroot(node::MCTSNode)::Bool = node.nodekey == "root" ? true : false
   - `n::Integer`
     how many times action will be sampled from decisionMaker 
   - `w::Float64`
-    exploration weight
+    exploration weight. Value is usually between 1 to 2. 
+    Value 1.0 makes MCTS balance between exploration and exploitation like 50%-50%
+    Value 2.0 makes MCTS aggressively search the tree
   
 # Return
   - `plan::Vector{Dict}`
@@ -441,7 +443,7 @@ function runMCTS(
 
     expand(a, node, decisionMaker, progressValueEstimator, isterminal, n=n)
     
-    leaf_node = selectChildNode(node)
+    leaf_node = UCTselect(node, w)
     simTrajectoryReward = simulate(a, leaf_node, decisionMaker, progressValueEstimator,
                       isterminal, maxDepth, n=n)
     backpropagate(leaf_node, simTrajectoryReward)

test/test_1.jl

@@ -79,7 +79,17 @@ result = GeneralUtils.sendMqttMsg(outgoingMsg)
 outgoingMsg = Dict(
   :msgMeta=> msgMeta,
   :payload=> Dict(
-    :text=> "<<ok>>",
+    :text=> "<<OK, I'll take it.>>",
+  )
+)
+result = GeneralUtils.sendMqttMsg(outgoingMsg)
+
+
+
+outgoingMsg = Dict(
+  :msgMeta=> msgMeta,
+  :payload=> Dict(
+    :text=> "Dry please.",
   )
 )
 result = GeneralUtils.sendMqttMsg(outgoingMsg)
@@ -93,4 +103,3 @@ result = GeneralUtils.sendMqttMsg(outgoingMsg)
 
 
 
-