Python与SimPy：第四章模拟编程实践

需积分: 9 114 浏览量更新于2024-07-19 1 收藏 398KB PDF 举报

"《Python模拟编程：基于SimPy的应用》" 第四章 "Python模拟编程实战" 本章节深入探讨如何利用Python及其SimPy模拟库来实现第三章中所介绍的一些概念实例。SimPy是一个针对Python编写的面向对象、基于过程的离散事件仿真库，适用于模拟各种场景，如顾客、消息传递和交通工具等活跃元素，以及服务器、收银台和隧道等具有有限容量的被动组件。它支持模型设计中的进程和资源管理，并提供监控变量来收集模拟运行时的统计数据。 SimPy的核心在于其简单易用的接口，它使得开发者能够通过创建对象和定义交互来构建模拟模型。该库采用了面向对象的编程方式，用户可以创建自定义的进程类（Process）来代表活动实体的行为，例如排队的客户或处理请求的服务。资源类（Resource）则用于限制并发访问，确保系统中的竞争和阻塞行为得到模拟。随机数生成是通过Python的标准random模块实现的，这意味着SimPy无需依赖其他外部库，兼容性良好，不仅能在Python环境中运行，还能在Java（Jython）和.NET环境下运行，增加了其应用的灵活性。在本章中，读者将学习如何使用SimPy的基本结构和功能，例如定义事件、调度器、定时器以及如何设置并观察模拟环境中的状态变化。同时，作者会引导读者思考实验设计和分析问题，这些都是后续章节的重要内容。通过跟随第四章的指导，读者不仅能掌握SimPy的使用技巧，还将对如何利用模拟技术进行系统设计和性能评估有更深入的理解。在实际操作中，你将学会如何编写模拟脚本，模拟一个简单的交通系统或者排队服务场景，然后逐步引入复杂性，如多线程、随机事件和优化策略。这章旨在帮助你建立起对SimPy核心理念的扎实基础，为后续在解决实际问题时运用高级模拟技术和数据挖掘技巧打下坚实的基础。

8 CHAPTER 4. SIMULATION PROGRAMMING WITH PYTHON

Within the Car object, the yield Sim.hold, self, timeParking line is

used to represent the car remaining in the parking lot. To look at this line more

closely

1. yield: a Python keyword that returns a generator. In this case, it calls

the function that follows in a computationally eﬃcient manner.

2. Sim.hold: The hold function within Simpy.Simulation. This causes a

process to wait.

3. self: A reference to the current object (Car). The ﬁrst parameter passed

to the Sim.hold command. In this case it means the currently created Car

should wait for a period of time. If the yield hold line was preceded by

a yield request, the resource required by yield request would be in-

cluded in the yield hold, i.e. both the current process and any resources

the process is using would wait for the speciﬁed period of time.

4. timeParking: The time that the Car should wait. This is the second

parameter passed to the Sim.hold command.

Typically, yield, request, hold, release are used in sequence. For ex-

ample, if the number of parking spaces were limited, the act of a car taking a

parking space for a period of time would be represented as follows within the

Car.visit() function.

yield Sim.request, self, parkingspace

yield Sim.hold, self, timeParking

yield Sim.release, self, parkingspace

In this case, since the purpose of the simulation is to determine demand for

parking spaces, we assume there are unlimited spaces and we count the number

of cars in the parking lot at any point in time.

In SimPy, resources such as parking spaces are represented by Resources.

There are three types of resources.

• A Resource is something whose units are required by a Process. When

a Process is done with a unit of the Resource it releases the unit for

another Process to use. A Process that requires a unit of Resource when

all units are busy with other Processes can join a queue and wait for the

next unit to be available.

• A Level is homogeneous undiﬀerentiated material that can be produced

or consumed by a Process. An example of a level is gasoline in a tank at

a gas station.

• A Store models the production or consumption of speciﬁc items of any

type. An example would be a storeroom that holds a variety of surgical

supplies,

4.2. SIMULATING THE M (T )/M/∞ QUEUE 9

class G:

maxTime = 24.0 # hours

arrivalrate = 100 # per hour

parkingtime = 2.0 # hours

parkedcars = 0

seedVal = 9999

Figure 4.3: Global declarations.

Global declarations are in Fig. 4.3. While Python allows these to be in the

global namespace, it is preferred that these be put into a class created for the

purpose as shown here. In particular, this is generally used as a central location

for parameters of the model.

The main simulation class is shown in Fig. 4.4. The simulation class Parkingsim

is a subclass of Sim.Simulation. While it can have other functions, the central

function is the run(self) function. Note that the ﬁrst argument of a member

function of a class is self, indicating that the function has access to the other

functions and variables of the class. The run function takes one argument, the

random number seed. Sim.initialize() sets all Monitors and the simulation

clock. Then processes and resources are created. In this simulation there is one

process, the generation of arriving cars in Arrival. Note that at this point,

there are no Car’s, as the Arrival process will create the cars. This simulation

does not have any resources; but if, for example, the parking lot had a limited

capacity of 20 spaces, it could have been created here by:

self.parkinglot = Sim.Resource(capacity = 20, name=’Parking lot’,

unitName=’Space’, monitored=True, sim=self)

This creates the parking lot with 20 spaces. By default, resources are a

FIFO, non-priority queue with no preemption. In addition to the capacity of

the Resource (number of servers), there are two queues (lists) associated with

the resource. First is the activeQ, which is the list of process objects currently

using the resource. Second is the waitQ, which is the number of processes that

have requested but have not yet received a unit of the resource. If when creating

the resource the option monitored=True is set, then a Monitor is created for

each queue, and statistics can be collected on resource use and the wait queue for

further analysis. The last option is sim=self, which declares that the Process

or Resource is part of the simulation deﬁned in the current class derived from

Sim.Simulation. If the simulation was deﬁned in the global namespace (not

inside of a class), then this option would not be needed.

After Processes and Resources are created, any additional Monitors are cre-

ated here. Note that Processes, Resources, and Monitors are created using

self. constructs. This classiﬁes the object as part of the class, and not only

local to the run() function. Other classes and methods that are part of this

simulation (declared using sim = self) can refer to any object created here

using self.sim.objectname. This includes Monitors as well as variables that

need to be updated from anywhere in the simulation. For example, the vari-

able self.parkedcars can be updated from the Car class in Fig. 4.2 using

剩余44页未读，继续阅读

zzs213

粉丝: 0
资源: 4

Python与SimPy：第四章模拟编程实践

用于托管 Discord Overlay 的 DirectX 11 窗口.zip

【路径规划】吉萨金子塔建造算法栅格地图机器人路径规划【含Matlab仿真 2835期】.zip

【任务分配】蒙特卡洛算法无人机任务分配【含Matlab仿真 3016期】.zip

排序

用于挂接 DirectX API 调用的库.zip

高效PDCA工作法：从计划到调整的管理四步骤指南

CnPack 密码算法库

【路径规划】A_Star算法多机器人牛耕式分区路径规划【含Matlab仿真 2793期】.zip

在Windows上通过解压包形式安装MySQL 8.0.40

【路径规划】白鲸算法栅格地图机器人最短路径规划【含Matlab仿真 2945期】.zip

最新资源