我正在尝试使用Simpy模拟城市网格周围移动汽车的一些行为。但是,我在概念上围绕何时使用像
这样的东西时遇到了一些麻烦 yield self.env.timeout(delay)
或yield env.process(self.someMethod())
而只是调用方法self.someMethod()
。
在非常理论的层面上,我理解yield
语句和生成器,了解它们如何应用于迭代,但不太确定它与Simpy
的关系。
Simpy
教程仍然非常密集。
例如:
class Car(object):
def __init__(self, env, somestuff):
self.env = env
self.somestuff = somestuff
self.action = env.process(self.startEngine()) # why is this needed? why not just call startEngine()?
def startEngine(self):
#start engine here
yield self.env.timeout(5) # wait 5 seconds before starting engine
# why is this needed? Why not just use sleep?
env = simpy.Environment()
somestuff = "blah"
car = Car(env, somestuff)
env.run()
答案 0 :(得分:2)
看起来你并没有完全理解generator / async 功能呢。我在下面评论你的代码并希望它有所帮助 你要明白发生了什么:
import simpy
class Car(object):
def __init__(self, env, somestuff):
self.env = env
self.somestuff = somestuff
# self.startEngine() would just create a Python generator
# object that does nothing. We must call "next(generator)"
# to run the gen. function's code until the first "yield"
# statement.
#
# If we pass the generator to "env.process()", SimPy will
# add it to its event queue actually run the generator.
self.action = env.process(self.startEngine())
def startEngine(self):
# "env.timeout()" returns a TimeOut event. If you don't use
# "yield", "startEngine()" returns directly after creating
# the event.
#
# If you yield the event, "startEngine()" will wait until
# the event has actually happend after 5 simulation steps.
#
# The difference to time.sleep(5) is, that this function
# would block until 5 seconds of real time has passed.
# If you instead "yield event", the yielding process will
# not block the whole thread but gets suspend by our event
# loop and resumed once the event has happend.
yield self.env.timeout(5)
env = simpy.Environment()
somestuff = "blah"
car = Car(env, somestuff)
env.run()