def env_step(self, action):
action = action.intArray
if len(action) != 3:
print action, len(action)
assert len(action) == self.simulationParameterObj.nbrReaches, "Expected " + str(
self.simulationParameterObj.nbrReaches) + " integer action."
if not InvasiveUtility.is_action_allowable(action, self.state):
theObs = Observation()
InvasiveUtility.is_action_allowable(action, self.state)
#map(int, results)
theObs.intArray = [-1]
returnRO = Reward_observation_terminal()
returnRO.r = self.Bad_Action_Penalty
returnRO.o = theObs
return returnRO
cost_state_unit = InvasiveUtility.get_unit_invaded_reaches(self.state,
self.simulationParameterObj.habitatSize) * self.actionParameterObj.costPerReach
stateCost = cost_state_unit + InvasiveUtility.get_invaded_reaches(
self.state) * self.actionParameterObj.costPerTree
stateCost = stateCost + InvasiveUtility.get_empty_slots(self.state) * self.actionParameterObj.emptyCost
costAction = InvasiveUtility.get_budget_cost_actions(action, self.state, self.actionParameterObj)
if costAction > self.actionParameterObj.budget:
theObs = Observation()
InvasiveUtility.is_action_allowable(action, self.state)
#map(int, results)
theObs.intArray = [-1]
returnRO = Reward_observation_terminal()
returnRO.r = self.Bad_Action_Penalty
returnRO.o = theObs
return returnRO
nextState = simulateNextState(self.state, action, self.simulationParameterObj,
self.actionParameterObj, self.dispertionTable, self.germinationObj)
self.state = nextState
theObs = Observation()
theObs.intArray = self.state
returnRO = Reward_observation_terminal()
returnRO.r = -1 * (costAction + stateCost)
returnRO.o = theObs
return returnRO
def env_step(self, action):
action = action.intArray
assert len(action) == self.simulationParameterObj.nbrReaches, "Expected " + str(
self.simulationParameterObj.nbrReaches) + " integer action."
if not InvasiveUtility.is_action_allowable(action, self.state):
theObs = Observation()
InvasiveUtility.is_action_allowable(action, self.state)
#map(int, results)
theObs.intArray = [-1]
returnRO = Reward_observation_terminal()
returnRO.r = self.Bad_Action_Penalty
returnRO.o = theObs
return returnRO
cost_state_unit = InvasiveUtility.get_unit_invaded_reaches(self.state,
self.simulationParameterObj.habitatSize) * self.actionParameterObj.costPerReach
stateCost = cost_state_unit + InvasiveUtility.get_invaded_reaches(
self.state) * self.actionParameterObj.costPerTree
stateCost = stateCost + InvasiveUtility.get_empty_slots(self.state) * self.actionParameterObj.emptyCost
costAction = InvasiveUtility.get_budget_cost_actions(action, self.state, self.actionParameterObj)
if costAction > self.actionParameterObj.budget:
theObs = Observation()
InvasiveUtility.is_action_allowable(action, self.state)
#map(int, results)
theObs.intArray = [-1]
returnRO = Reward_observation_terminal()
returnRO.r = self.Bad_Action_Penalty
returnRO.o = theObs
return returnRO
nextState = simulateNextState(self.state, action, self.simulationParameterObj,
self.actionParameterObj, self.dispertionTable, self.germinationObj)
self.state = nextState
theObs = Observation()
theObs.intArray = self.state
returnRO = Reward_observation_terminal()
returnRO.r = -1 * (costAction + stateCost)
returnRO.o = theObs
return returnRO
