Question

我试图在PyBrain中实现类似于Maze问题的东西。然而，它更像是一个带紧急出口的房间，您可以在其中一个房间内找到出口。要将其转换为计算机方法，可以使用双向图，其中权重显示房间之间的路径。

我试图实现一个新的环境，但我有点失去应该是什么。例如，基于抽象环境类我已经考虑过这个：

#!/usr/bin/python2.7

class RoomEnv(Environment):
    # number of action values acceptable by the environment
    # Two events: go forward and go back through the door (but, how we know what room is connect to another?)
    indim = 2
    # Maybe a matrix where 0 is no connection and 1 is a connection(?)
    #            A,B,C,D,E,F
    #indim = array([[0,0,0,0,0,0],  # A
                    [0,0,0,0,0,1],  # B
                    [0,0,0,0,0,0],  # C
                    [0,0,0,0,0,0],  # D
                    [0,0,0,0,0,1],  # E
                    [0,0,0,0,0,1],  # F
                  ])

    # the number of sensors is the number of the rooms
    outdim = 6

    def getSensors(self):
        # Initial state:
        # Could be any room, maybe something random(?)

    def performAction(self, action):
        # We should look at all the states possible to learn what are the best option to go to the outside state.
        # Maybe a for loop that goes through all the paths and use some weight to know where is the best option?

        print "Action performed: ", action

    def reset(self):
        #Most environments will implement this optional method that allows for reinitialization.

此致

Answer 1

在 pybrain 中，您可以将房间定义为数组，然后将结构传递给Maze以创建新环境。例如：

structure = array([[1, 1, 1, 1, 1, 1, 1, 1, 1],
                   [1, 0, 0, 1, 0, 0, 0, 0, 1],
                   [1, 0, 0, 1, 0, 0, 1, 0, 1],
                   [1, 0, 0, 1, 0, 0, 1, 0, 1],
                   [1, 0, 0, 1, 0, 1, 1, 0, 1],
                   [1, 0, 0, 0, 0, 0, 1, 0, 1],
                   [1, 1, 1, 1, 1, 1, 1, 0, 1],
                   [1, 0, 0, 0, 0, 0, 0, 0, 1],
                   [1, 1, 1, 1, 1, 1, 1, 1, 1]])

# defining the environment
environment = Maze(structure, (7, 7))

在上面的示例中，1s显示墙，0显示代理可以行走的网格。因此，您可以修改结构以制作自己的结构。

PyBrain强化学习 - 迷宫和图形

1 个答案: