def __init__(self, nrows = 8, ncols = 8):
self.nrows = nrows
self.ncols = ncols
self.nstates = nrows * ncols
self.nactions = 4
self.left_edge = []
self.right_edge = []
self.top_edge = []
self.bottom_edge = []
self.gamma = 0.9
for x in range(self.nstates):
# note that edges are not disjoint, so we cannot use elif
if x % self.ncols == 0:
self.left_edge.append(x)
if 0 <= x < self.ncols:
self.top_edge.append(x)
if x % self.ncols == self.ncols - 1:
self.right_edge.append(x)
if (self.nrows - 1) * self.ncols <= x <= self.nstates:
self.bottom_edge.append(x)
SparseMDP.__init__(self, nstates = self.nrows * self.ncols, nactions = 4)
def __init__(self, size = 10, rsize = 10):
self.size = 10
self.goal_index = 0
self.state_names = [(x,y) for x in range(self.size) for y in range(self.size)] # x is the agent location, y is the goal location
self.states = range(len(self.state_names))
self.actions = [0,1] # left or right
self.nstates = self.size ** 2
self.nactions = 2
self.endstates = []
for (i,s) in enumerate(self.state_names):
if s[0] == s[1]:
self.endstates.append(i)
SparseMDP.__init__(self, nstates = self.nstates, nactions = self.nactions)
def __init__(self, size=10, rsize=10):
self.size = 10
self.goal_index = 0
self.state_names = [
(x, y) for x in range(self.size) for y in range(self.size)
] # x is the agent location, y is the goal location
self.states = range(len(self.state_names))
self.actions = [0, 1] # left or right
self.nstates = self.size**2
self.nactions = 2
self.endstates = []
for (i, s) in enumerate(self.state_names):
if s[0] == s[1]:
self.endstates.append(i)
SparseMDP.__init__(self, nstates=self.nstates, nactions=self.nactions)
def __init__(self, nrows = 5, ncols = 5, actions = None, walls=[(1,1),(1,2),(1,3),(2,1),(2,2),(2,3),(3,1),(3,2),(3,3)], endstates = [0]):
self.nrows = nrows
self.ncols = ncols
self.walls = walls
grid = [self.coords(i) for i in range(self.nrows * self.ncols)]
grid = [s for s in grid if not s in self.walls]
self.states = dict([(i,s) for (i,s) in enumerate(grid)])
self.rstates = dict([(s,i) for (i,s) in enumerate(grid)]) # reverse lookup by grid coords
if actions is None:
actions = range(8)
self.allowed_actions = actions
self.nstates = len(self.states)
self.nactions = len(actions)
self.endstates = endstates
SparseMDP.__init__(self, nstates = self.nstates, nactions = self.nactions)
def __init__(self, nrows=5, ncols=5):
self.nrows = nrows
self.ncols = ncols
grid = [self.coords(i) for i in range(self.nrows * self.ncols)]
grid = [(s,t) for s in grid for t in grid]
self.states = dict([(i,s) for (i,s) in enumerate(grid)])
self.rstates = dict([(s,i) for (i,s) in enumerate(grid)])
self.nstates = len(self.states)
self.current = 0
self.nactions = 8
self.endstates = []
self.observations = []
for i in range(self.nstates):
o = self.states[i]
self.observations.append([o[0][0],o[0][1],o[1][0],o[1][1]])
self.observations = np.array(self.observations)
for (s,t) in self.rstates:
if s[0] == t[0] and s[1] == t[1]:
self.endstates.append(self.rstates[(s,t)])
SparseMDP.__init__(self, nstates = self.nstates, nactions = self.nactions)