all changes for all parts

bustersAgents.py

@@ -144,3 +144,24 @@ class GreedyBustersAgent(BustersAgent):
             [beliefs for i, beliefs in enumerate(self.ghostBeliefs)
              if livingGhosts[i+1]]
         "*** YOUR CODE HERE ***"
+
+        max_level = []
+        
+        for belief in livingGhostPositionDistributions:
+            max_level.append(belief.argMax())
+
+        goalProbability, goalCoordinate = 0, None
+        for index, coordinate in enumerate(max_level):
+            # checking goal
+            if livingGhostPositionDistributions[index][coordinate] >= goalProbability:
+                goalCoordinate = coordinate
+                goalProbability = livingGhostPositionDistributions[index][coordinate]
+
+        prs = []
+
+        # checking all acitons
+        for action in legal:
+            nextLocation = Actions.getSuccessor(pacmanPosition, action)
+            prs.append((self.distancer.getDistance(nextLocation, goalCoordinate), action))
+            
+        return min(prs)[1]

inference.py

@@ -25,6 +25,7 @@ class DiscreteDistribution(dict):
     A DiscreteDistribution models belief distributions and weight distributions
     over a finite set of discrete keys.
     """
+
     def __getitem__(self, key):
         self.setdefault(key, 0)
         return dict.__getitem__(self, key)
@@ -75,7 +76,14 @@ class DiscreteDistribution(dict):
         {}
         """
         "*** YOUR CODE HERE ***"
-        raiseNotDefined()
+        total_calc = float(self.total())
+
+        # need to stop and return func since landed on ghost
+        if total_calc == 0:
+            return
+
+        for key in self.keys():
+            self[key] = self[key] / total_calc
 
     def sample(self):
         """
@@ -99,7 +107,23 @@ class DiscreteDistribution(dict):
         0.0
         """
         "*** YOUR CODE HERE ***"
-        raiseNotDefined()
+        #  already normalized?
+        if self.total() != 1:
+            self.normalize()
+
+        items_sorted = sorted(self.items())
+
+        dist = [i[1] for i in items_sorted]
+        values = [i[0] for i in items_sorted]
+
+        # incorporating the random as in spec
+        choice = random.random()
+
+        total, i = dist[0], 0
+        while choice > total:
+            i += 1
+            total += dist[i]
+        return values[i]
 
 
 class InferenceModule:
@@ -129,13 +153,14 @@ class InferenceModule:
             jail = self.getJailPosition(index)
             gameState = self.setGhostPositions(gameState, pos)
         pacmanPosition = gameState.getPacmanPosition()
-        ghostPosition = gameState.getGhostPosition(index + 1)  # The position you set
+        ghostPosition = gameState.getGhostPosition(
+            index + 1)  # The position you set
         dist = DiscreteDistribution()
         if pacmanPosition == ghostPosition:  # The ghost has been caught!
             dist[jail] = 1.0
             return dist
-        pacmanSuccessorStates = game.Actions.getLegalNeighbors(pacmanPosition, \
+        pacmanSuccessorStates = game.Actions.getLegalNeighbors(pacmanPosition,
-                gameState.getWalls())  # Positions Pacman can move to
+                                                               gameState.getWalls())  # Positions Pacman can move to
         if ghostPosition in pacmanSuccessorStates:  # Ghost could get caught
             mult = 1.0 / float(len(pacmanSuccessorStates))
             dist[jail] = mult
@@ -143,11 +168,13 @@ class InferenceModule:
             mult = 0.0
         actionDist = agent.getDistribution(gameState)
         for action, prob in actionDist.items():
-            successorPosition = game.Actions.getSuccessor(ghostPosition, action)
+            successorPosition = game.Actions.getSuccessor(
+                ghostPosition, action)
             if successorPosition in pacmanSuccessorStates:  # Ghost could get caught
                 denom = float(len(actionDist))
                 dist[jail] += prob * (1.0 / denom) * (1.0 - mult)
-                dist[successorPosition] = prob * ((denom - 1.0) / denom) * (1.0 - mult)
+                dist[successorPosition] = prob * \
+                    ((denom - 1.0) / denom) * (1.0 - mult)
             else:
                 dist[successorPosition] = prob * (1.0 - mult)
         return dist
@@ -169,7 +196,20 @@ class InferenceModule:
         Return the probability P(noisyDistance | pacmanPosition, ghostPosition).
         """
         "*** YOUR CODE HERE ***"
-        raiseNotDefined()
+        # check if ghost in jail currently
+        if jailPosition == ghostPosition:
+            # remember need floats now
+            if noisyDistance == None:
+                return 1.0
+            else:
+                return 0.0
+
+        if noisyDistance == None:
+            return 0.0
+
+        actual = manhattanDistance(pacmanPosition, ghostPosition)
+
+        return busters.getObservationProbability(noisyDistance, actual)
 
     def setGhostPosition(self, gameState, ghostPosition, index):
         """
@@ -195,7 +235,8 @@ class InferenceModule:
         """
         for index, pos in enumerate(ghostPositions):
             conf = game.Configuration(pos, game.Directions.STOP)
-            gameState.data.agentStates[index + 1] = game.AgentState(conf, False)
+            gameState.data.agentStates[index +
+                                       1] = game.AgentState(conf, False)
         return gameState
 
     def observe(self, gameState):
@@ -212,7 +253,8 @@ class InferenceModule:
         """
         Initialize beliefs to a uniform distribution over all legal positions.
         """
-        self.legalPositions = [p for p in gameState.getWalls().asList(False) if p[1] > 1]
+        self.legalPositions = [
+            p for p in gameState.getWalls().asList(False) if p[1] > 1]
         self.allPositions = self.legalPositions + [self.getJailPosition()]
         self.initializeUniformly(gameState)
 
@@ -251,6 +293,7 @@ class ExactInference(InferenceModule):
     The exact dynamic inference module should use forward algorithm updates to
     compute the exact belief function at each time step.
     """
+
     def initializeUniformly(self, gameState):
         """
         Begin with a uniform distribution over legal ghost positions (i.e., not
@@ -277,9 +320,20 @@ class ExactInference(InferenceModule):
         position is known.
         """
         "*** YOUR CODE HERE ***"
-        raiseNotDefined()
+        distribution = DiscreteDistribution()
+        #  from input vars
+        pacmanPosition = gameState.getPacmanPosition()
 
-        self.beliefs.normalize()
+        jailPosition = self.getJailPosition()
+
+        for pos in self.allPositions:
+            prob = self.getObservationProb(
+                observation, pacmanPosition, pos, jailPosition)
+            distribution[pos] = prob * self.beliefs[pos]
+
+        distribution.normalize()
+
+        self.beliefs = distribution
 
     def elapseTime(self, gameState):
         """
@@ -291,7 +345,20 @@ class ExactInference(InferenceModule):
         current position is known.
         """
         "*** YOUR CODE HERE ***"
-        raiseNotDefined()
+        import util
+
+        distribution = DiscreteDistribution()
+
+        # iterate all over pos
+        for old_pos in self.allPositions:
+            new_pos = self.getPositionDistribution(gameState, old_pos)
+            # update keys
+            old_prob = self.beliefs[old_pos]
+
+            for newPos in new_pos.keys():
+                distribution[newPos] += old_prob * new_pos[newPos]
+
+        self.beliefs = distribution
 
     def getBeliefDistribution(self):
         return self.beliefs
@@ -301,6 +368,7 @@ class ParticleFilter(InferenceModule):
     """
     A particle filter for approximately tracking a single ghost.
     """
+
     def __init__(self, ghostAgent, numParticles=300):
         InferenceModule.__init__(self, ghostAgent)
         self.setNumParticles(numParticles)
@@ -318,7 +386,14 @@ class ParticleFilter(InferenceModule):
         """
         self.particles = []
         "*** YOUR CODE HERE ***"
-        raiseNotDefined()
+        count = 0
+        self.particles = []
+        # going through all particles
+        while count < self.numParticles:
+            for position in self.legalPositions:
+                if count < self.numParticles:
+                    self.particles.append(position)
+                    count += 1
 
     def observeUpdate(self, observation, gameState):
         """
@@ -333,7 +408,25 @@ class ParticleFilter(InferenceModule):
         the DiscreteDistribution may be useful.
         """
         "*** YOUR CODE HERE ***"
-        raiseNotDefined()
+        # get from given input
+        pacmanPosition, jailPosition = gameState.getPacmanPosition(), self.getJailPosition()
+        distribution = DiscreteDistribution()
+
+        for par in self.particles:
+            prob = self.getObservationProb(
+                observation, pacmanPosition, par, jailPosition)
+            distribution[par] += prob
+
+        # check if not vialbe
+        if distribution.total() == 0:
+            self.initializeUniformly(gameState)
+        else:
+            distribution.normalize()
+            self.beliefs = distribution
+
+            for x in range(self.numParticles):
+                new_sample = distribution.sample()
+                self.particles[x] = new_sample
 
     def elapseTime(self, gameState):
         """
@@ -341,7 +434,20 @@ class ParticleFilter(InferenceModule):
         gameState.
         """
         "*** YOUR CODE HERE ***"
-        raiseNotDefined()
+        import util
+
+        cache = {}
+        # maybe use counter?
+        for i in range(self.numParticles):
+            particle = self.particles[i]
+
+            # check cases in spec
+            if particle in cache:
+                self.particles[i] = cache[particle].sample()
+            else:
+                dist = self.getPositionDistribution(gameState, particle)
+                cache[particle] = dist
+                self.particles[i] = dist.sample()
 
     def getBeliefDistribution(self):
         """
@@ -352,7 +458,14 @@ class ParticleFilter(InferenceModule):
         This function should return a normalized distribution.
         """
         "*** YOUR CODE HERE ***"
-        raiseNotDefined()
+        import util
+        distribution = util.Counter()
+
+        for element in self.particles:
+            distribution[element] += 1
+
+        distribution.normalize()
+        return distribution
 
 
 class JointParticleFilter(ParticleFilter):
@@ -360,6 +473,7 @@ class JointParticleFilter(ParticleFilter):
     JointParticleFilter tracks a joint distribution over tuples of all ghost
     positions.
     """
+
     def __init__(self, numParticles=600):
         self.setNumParticles(numParticles)
 
@@ -380,7 +494,16 @@ class JointParticleFilter(ParticleFilter):
         """
         self.particles = []
         "*** YOUR CODE HERE ***"
-        raiseNotDefined()
+        perm = list(itertools.product(self.legalPositions, repeat = self.numGhosts))
+
+        random.shuffle(perm)
+        size, n = len(perm), self.numParticles
+        
+        while size < n:
+            n -= size
+            self.particles += perm
+            
+        self.particles = self.particles + perm[:n]
 
     def addGhostAgent(self, agent):
         """
@@ -413,20 +536,48 @@ class JointParticleFilter(ParticleFilter):
         the DiscreteDistribution may be useful.
         """
         "*** YOUR CODE HERE ***"
-        raiseNotDefined()
+        pacmanPosition, distribution = gameState.getPacmanPosition(), DiscreteDistribution()
+        
+        for pros in self.particles:
+            curr_p = 1
+            for i in range(self.numGhosts):
+                noisy_dist = observation[i]
+                curr_p *= self.getObservationProb(noisy_dist, pacmanPosition, pros[i], self.getJailPosition(i))
+            distribution[pros] += curr_p
+
+        self.beliefs = distribution
+        # check for total/norm
+        if self.beliefs.total() == 0:
+            self.initializeUniformly(gameState)
+        else:
+            self.beliefs.normalize()
+            # ok through all particles and assign
+            for x in range(self.numParticles):
+                newPos = self.beliefs.sample()
+                self.particles[x] = newPos
 
     def elapseTime(self, gameState):
         """
         Sample each particle's next state based on its current state and the
         gameState.
         """
-        newParticles = []
+        newParticles, cache = [], {}
         for oldParticle in self.particles:
             newParticle = list(oldParticle)  # A list of ghost positions
 
             # now loop through and update each entry in newParticle...
             "*** YOUR CODE HERE ***"
-            raiseNotDefined()
+            prevPos = list(oldParticle)
+            # through all ghosts
+            for i in range(self.numGhosts):
+                # check if seen before
+                if (oldParticle, i) in cache:
+                    newParticle[i] = cache[(oldParticle, i)].sample()
+                else:
+                    newPosDist = self.getPositionDistribution(gameState, prevPos, i, self.ghostAgents[i])
+                    cache[(oldParticle, i)] = newPosDist
+                    # assign new ones to each particle
+                    newParticle[i] = newPosDist.sample()
 
             """*** END YOUR CODE HERE ***"""
             newParticles.append(tuple(newParticle))
@@ -442,6 +593,7 @@ class MarginalInference(InferenceModule):
     A wrapper around the JointInference module that returns marginal beliefs
     about ghosts.
     """
+
     def initializeUniformly(self, gameState):
         """
         Set the belief state to an initial, prior value.