SimPy中Container资源level属性瞬态值异常及日志记录咨询

我正在使用SimPy进行实验，尝试监控Container资源的level属性。为便于说明，现举一个简单示例：生产者进程每3个时间步生产一个物品并放入容器，消费者进程每个时间步从容器中取出一个物品，容器初始level为3。

实现代码 #

import simpy

# define a "producer" that puts a new item in the container every 3 time units
def producer(env, container):
    while(True):
        yield env.timeout(3)
        print(f'Producer at t = {env.now}: Produced item. '
              f'Requesting to put it into container. '
              f'Current container level (before putting): {container.level}')
        yield container.put(1)
        print(f'Producer at t = {env.now}: Successfully put produced item into container. '
              f'New container level: {container.level}')

# define a "consumer" that attempts to take one item from the container at every time unit
def consumer(env, container):
    while(True):
        print(f'Consumer at t = {env.now}: Requesting item from container. '
              f'Current container level (before taking item): {container.level}')
        yield container.get(1)
        print(f'Consumer at t = {env.now}: Consumed item from container. '
              f'New container level: {container.level}')
        yield env.timeout(2)

# setup the simpy environment
env = simpy.Environment()
container = simpy.Container(env, init = 3)
producer_process = env.process(producer(env, container))
consumer_process = env.process(consumer(env, container))

# run the simulation
print(f'Initial container level: {container.level}')
env.run(until=7)

运行输出 #

Initial container level: 3
Consumer at t = 0: Requesting item from container. Current container level (before taking item): 3
Consumer at t = 0: Consumed item from container. New container level: 2
Consumer at t = 2: Requesting item from container. Current container level (before taking item): 2
Consumer at t = 2: Consumed item from container. New container level: 1
Producer at t = 3: Produced item. Requesting to put it into container. Current container level (before putting): 1
Producer at t = 3: Successfully put produced item into container. New container level: 2
Consumer at t = 4: Requesting item from container. Current container level (before taking item): 2
Consumer at t = 4: Consumed item from container. New container level: 1
Producer at t = 6: Produced item. Requesting to put it into container. Current container level (before putting): 1
Consumer at t = 6: Requesting item from container. Current container level (before taking item): 2
Producer at t = 6: Successfully put produced item into container. New container level: 1
Consumer at t = 6: Consumed item from container. New container level: 1

我发现结果存在异常，例如在时间步6时，容器level在新生产的物品放入前就从1变为2。现咨询：SimPy中Container资源的level属性是何时及如何更新的？如何获取或生成正确的瞬态level变化日志？

一、Container level属性的更新时机与机制 #

SimPy的Container是事件驱动的资源，其level属性的更新遵循以下规则：

• 即时更新：当put()或get()操作能立即满足条件（如put()时容器有剩余容量、get()时容器有足够物品），level会在操作执行瞬间同步更新，无需等待事件挂起。
• 同时间点事件的顺序影响：多个事件在同一时间点触发时，SimPy会按事件的注册顺序依次处理。若前一个事件修改了level，后续同时间点的事件读取到的就是更新后的值——这正是t=6时异常的原因：生产者的put(1)事件比消费者的get(1)事件先被注册，SimPy先执行put操作将level从1改为2，导致消费者打印前就读取到了更新后的level。

二、生成正确瞬态level变化日志的方法 #

方案1：自定义LoggedContainer子类 #

通过继承simipy.Container，重载put()和get()方法，在操作前后强制记录level，确保日志与状态变化严格同步：

import simpy

class LoggedContainer(simpy.Container):
    def put(self, amount):
        print(f'[LOG] t={self._env.now} | PUT before: level={self.level}, adding={amount}')
        result = super().put(amount)
        def log_after_put(event):
            print(f'[LOG] t={self._env.now} | PUT after: level={self.level}')
        result.callbacks.append(log_after_put)
        return result

    def get(self, amount):
        print(f'[LOG] t={self._env.now} | GET before: level={self.level}, taking={amount}')
        result = super().get(amount)
        def log_after_get(event):
            print(f'[LOG] t={self._env.now} | GET after: level={self.level}')
        result.callbacks.append(log_after_get)
        return result

# 原生产者/消费者逻辑不变，替换容器类
def producer(env, container):
    while(True):
        yield env.timeout(3)
        yield container.put(1)

def consumer(env, container):
    while(True):
        yield container.get(1)
        yield env.timeout(2)

env = simpy.Environment()
container = LoggedContainer(env, init=3)
env.process(producer(env, container))
env.process(consumer(env, container))
env.run(until=7)

方案2：利用事件回调机制 #

不修改容器类，在创建put()/get()事件时添加回调函数，在事件完成时记录level：

import simpy

def log_put_success(env, container, amount):
    print(f'[LOG] t={env.now} | PUT completed: level={container.level}')

def log_get_success(env, container, amount):
    print(f'[LOG] t={env.now} | GET completed: level={container.level}')

def producer(env, container):
    while(True):
        yield env.timeout(3)
        put_event = container.put(1)
        put_event.callbacks.append(lambda e: log_put_success(env, container, 1))
        yield put_event

def consumer(env, container):
    while(True):
        get_event = container.get(1)
        get_event.callbacks.append(lambda e: log_get_success(env, container, 1))
        yield get_event
        yield env.timeout(2)

env = simpy.Environment()
container = simpy.Container(env, init=3)
print(f'Initial container level: {container.level}')
env.process(producer(env, container))
env.process(consumer(env, container))
env.run(until=7)

三、t=6异常的具体解释 #

原代码中t=6的事件执行顺序为：

1. 生产者的timeout(3)触发，打印level=1，发起put(1)请求；
2. 消费者的timeout(2)触发，此时SimPy已处理完生产者的put操作，level变为2，消费者打印该值后发起get(1)请求；
3. 生产者的put操作完成，此时get操作已同步执行，level被改回1，生产者打印该值；
4. 消费者的get操作完成，打印level=1。

SimPy按事件创建顺序处理同时间点事件，导致消费者读取到了提前更新的level值。

内容的提问来源于stack exchange，提问作者kaffeesachse