Discrete-Event Simulation E-Book

Abdessalem Jerbi

• Herausgeber: John Wiley & Sons
• Kategorie: Fachliteratur
• Serie: ISTE Consignment
• Sprache: Englisch

Seitenzahl: 321

Veröffentlichungsjahr: 2024

Beschreibung

The use of discrete-event simulation in various fields, such as in industry, logistics and public health, has really taken off over the last few decades. The implementation of discrete-event simulation does however require an understanding, and perhaps even a mastery, of precise theoretical and methodological principles.

Discrete-Event Simulation presents the key concepts involved in any discrete-event simulation project, covering the most frequently used techniques for analysing data and results, the methodological and practical aspects of implementing discrete-event simulation, along with an introduction to the use of the “Arena” discrete-event simulation tool. This book combines the elements presented with applied examples, as well as numerous examples of simulation projects in various fields.

Leseprobe

Series EditorJean-Charles Pomerol

Discrete-Event Simulation

Concepts and Production in Arena

Abdessalem Jerbi

First published 2024 in Great Britain and the United States by ISTE Ltd and John Wiley & Sons, Inc.

Apart from any fair dealing for the purposes of research or private study, or criticism or review, as permitted under the Copyright, Designs and Patents Act 1988, this publication may only be reproduced, stored or transmitted, in any form or by any means, with the prior permission in writing of the publishers, or in the case of reprographic reproduction in accordance with the terms and licenses issued by the CLA. Enquiries concerning reproduction outside these terms should be sent to the publishers at the undermentioned address:

ISTE Ltd27-37 St George’s RoadLondon SW19 4EUUK

www.iste.co.uk

John Wiley & Sons, Inc.111 River StreetHoboken, NJ 07030USA

www.wiley.com

© ISTE Ltd 2024The rights of Abdessalem Jerbi to be identified as the author of this work have been asserted by him in accordance with the Copyright, Designs and Patents Act 1988.

Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s), contributor(s) or editor(s) and do not necessarily reflect the views of ISTE Group.

Library of Congress Control Number: 2024943359

British Library Cataloguing-in-Publication DataA CIP record for this book is available from the British LibraryISBN 978-1-78630-974-7

Preface

Simulation is a process which consists of designing a model of the (real) system studied, carrying out experiments on this model (and not calculations) and interpreting the observations provided by the progress of the model in order to formulate decisions relating to the system. The goal may be to understand the dynamic behavior of the system, to compare different configurations of this system, to evaluate different control strategies or even to evaluate and optimize the performance of the studied system.

Simulation can be used in different fields such as:

production flow systems for:

balancing assembly lines,

the design of transfer systems between production stations,

the sizing of workshop stocks,

the comparison of queue control policies;

logistics flows and transport systems for:

warehouse design and sizing,

the sizing of a truck fleet,

the study of procedures for controlling the flow of vehicles in circulation;

service systems such as:

the study of banking transactions,

restaurant management,

comparison of maintenance policies.

This work attempts to explain in detail the different stages of the discrete event simulation technique. Thanks to simple educational examples and problems inspired by practice, the reader will then be able to apply these theoretical notions of modeling and simulation using the Arena simulation software. This book also presents various discrete event simulation projects in several fields such as the manufacturing field, the public health field and the supply chain field.

August 2024

1Simulation

1.1. Introduction

Simulation aims to imitate the behavior of real systems on a computer equipped with appropriate software. It is generally used to analyze systems and make operational or resource policy decisions. Nowadays, simulation is a tool that is becoming increasingly popular, as computers and software become increasingly efficient (Chung 2003; Fleury et al. 2006; Altiok and Melamed 2007; Kelton et al. 2015; Polenghi et al. 2018).

The study of a system leads in most cases to modeling its functioning through the establishment of mathematical or logical relationships. It is possible to use mathematical methods if these relationships are simple enough. This solution is then called an analytical solution. However, systems in reality are most often too complex to be able to apply such an evaluation. We then resort to simulation in order to estimate the desired characteristics of the model.

Let us take the example of an industrial company which seeks to expand without being sure of the potential gain generated by this expansion. Indeed, it would not be profitable for the company to invest money for an extension, then later remove this extension, if the latter is deemed unprofitable. A simulation study could enlighten decision-makers on the consequences of this extension by simulating the operation of the factory before and after extension. We then achieve results without making physical changes in the factory.

Simulation is a process which therefore consists of designing a model of a real system, carrying out experiments on this model, interpreting the observations provided by the execution of the simulation of the model and formulating decisions relating to the system. The goal of this process can be to compare different configurations of the system studied or to evaluate different strategies for its control in order to optimize its performances.

The fields of application of simulation are numerous and varied. A non-exhaustive list of problems for which simulation has proven to be a useful and powerful tool includes:

production flow systems (Addi Ait

2000

):

machining operations: simulations of machining operations may include processes involving manually or computer-controlled factory equipment for machining, turning, bending, cutting, and welding,

assembly operations: assembly operations simulation can cover any type of assembly line or manufacturing operation that requires the assembly of multiple components into a single part,

logistics flows and transport systems:

material handling equipment: material handling simulations include analysis of cranes, forklifts and automated guided vehicles,

warehousing: warehousing simulations may involve manual or automated storage and retrieval of raw materials or finished products;

the production of services:

hospitals and medical clinics: models of hospitals and medical clinics can be simulated to determine the number of rooms, nurses and doctors for a particular location,

retail stores: retail stores may need to know how many checkouts to use,

food or entertainment facilities: entertainment facilities, such as multi-theater movie theater complexes, may be interested in the number of ticket sellers, ticket checkers or concession stand attendants to employ,

information technology: information technology models generally concern the number and type of networks or support resources to be made available,

customer ordering systems: customer ordering systems may need to know how many customer order representatives are needed on duty.

1.2. Advantages of simulation

1.2.1. Compressed time experimentation

As the model is simulated on a computer, experimental simulations can be carried out in compressed time. This is a major advantage because some processes can take months or even years to complete. Long system processing times can make robust analysis difficult to achieve. With a computer model, the operation of long processes can be simulated in seconds. Additionally, multiple repetitions of each simulation can easily be performed to increase the statistical reliability of the analysis (Kelton et al. 2015).

1.2.2. Reduced analytical requirements

Before the existence of digital simulation, we were forced to use other tools that were more analytically demanding. Even then, only simple systems involving probabilistic elements could be analyzed. More complex systems were strictly the domain of the mathematical researcher or operations research analyst. Furthermore, systems could only be analyzed with a static approach at a given point in time. On the other hand, the advent of simulation methodologies has made it possible to study systems dynamically in real time. Additionally, the development of simulation-specific software has eliminated many of the complicated basic calculations and programming requirements that might otherwise have been necessary. These reduced analytical requirements have made it possible to analyze many more different types of systems than before with a greater variety of experiments (Kelton et al. 2015; Polenghi et al. 2018).

1.2.3. Easy to demonstrate models

Most simulation software has the ability to dynamically animate the operation of the model. Animation is useful both for debugging the model and for verifying and demonstrating its operation. Animation-based debugging makes it easy to observe flaws in model logic. Using animation during a presentation can help establish the credibility of the model through the dynamic demonstration of how the system model handles different situations. Without the capability of animation, we would be limited to less effective textual and digital presentations (Kelton et al. 2015).

1.3. Disadvantages of simulation

Although simulation has many advantages, there are also some disadvantages. These drawbacks are not really associated directly with the modeling and analysis of a system, but rather with the expectations associated with simulation projects (Chung 2003; Fleury et al. 2006; Melamed and Rutgers 2007). These disadvantages are as follows:

Under no circumstances can simulation give accurate results when the input data are inaccurate: no matter how good a model is, if its input data are not accurate, we cannot reasonably expect to obtain accurate output data. Unfortunately, the data collection phase is considered the most difficult part of the simulation process.

Nevertheless, it is common that little time is allocated to this phase. In many cases of simulations, historical data of questionable quality have been accepted in order to save input data collection time. Too often, the exact nature or conditions under which these data are collected are unknown.

Simulation alone cannot solve problems: some managers may believe that carrying out a model simulation project is enough to solve the problem they are facing. However, simulation by itself cannot actually solve the problem. It can provide direction for potential solutions to resolve the problem and it is up to the manager to implement the proposed changes.

1.4. Concepts relating to simulation

1.4.1. System concept

Different definitions have been given to the word “system” in the field of simulation (Kelton et al. 2015). Here are some examples:

a system is a set of components linked together;

a system is an organized set of functional elements;

a system is a combination of parts that coordinate to achieve a result so as to form a set;

a system is a combination of people, machines, materials and information intended to satisfy a given objective.

All these definitions have in common a few key words which characterize the notion of system. Indeed, a system is characterized by ordered parts which compose it. Each of these parts has its own laws and a certain independence. These parts also have links or relationships between them. Furthermore, the whole thing, or the system, changes over time and is influenced by the environment in which it exists and which reacts on it. Finally, this set is most often subject to constraints and only exists to achieve a goal.

In summary, a system can be defined by the knowledge of its parts or its components, the laws specific to each component and the interactions which determine its purpose.

Example: in a production plant, the parts that make up the system could be the workforce and departments of purchasing and supply, inventory management, manufacturing and production scheduling, sales and administrative. Each of these services has its own operating laws and is partially independent of the others. But also, they interact with each other. Knowledge of these interactions and their application to the system will significantly influence the goal to be achieved by the system such as profitability, investment, profit, etc.

1.4.2. Model and modeling

A model is a representation of a system whose goal is to explain and/or predict certain aspects of its behavior. This representation is more or less faithful, because, on the one hand, the model must be sufficiently complete in order to be able to answer the various questions that can be asked about the system it represents and, on the other hand, it must not be too complex to be easily manipulated (Kelton et al. 2015).

When modeling a system, any modeler must ensure that the boundaries of their model are clearly defined through the internal variables, inputs and necessary outputs. In addition, they must judge the level of detail to incorporate into the model.

The two methods of progressive refinement and submodel accumulation are commonly used for modeling. Progressive refinement proceeds through iterations towards increasingly increasing levels of complexity. This evolutionary process makes it possible to take into account the inadequacies of the model at a given stage and to improve it at the next stage. The accumulation of submodels is used when the system to be studied is large. In this case, the system is divided into distinct subsystems in order to control the complexity of the system. All problems can manifest themselves in the integration stage, which leads to often delicate interface problems.

There are several classifications of simulation models. However, the most used consists of classifying them according to three dimensions, which are as follows:

Static or dynamic: static models have no interaction with time, which plays an essential role in dynamic models.

Continuous or discrete: the state of the continuous system can change continuously over time, such as the level of a reservoir as water enters and exits. In contrast, state changes in a discrete model can only occur at specific times. This is the case of a manufacturing system where parts arrive and leave at specific times and where machines stop and start again at specific times. It is possible that the same model has both continuous and discrete elements of change. This model is then called a mixed continuous-discrete model; for example, the case of a system for extruding PVC evacuation tubes and cutting them into units of four meters in length.

Deterministic or stochastic: models with constant inputs and variables are considered deterministic. On the other hand, stochastic models work with at least some random inputs.

Terminating or non-terminating: terminating and non-terminating systems are distinguished by:

initial starting conditions: terminating systems generally begin each time period without any influence from the previous period. Many systems that use client-type inputs are considered terminating-type systems. A bank, for example, does not let its customers stay overnight on its premises. This means that every new day or period of time, the system starts empty. The non-terminating system, on the other hand, can start with entries already present in the system from the previous period;