| 查看: 649 | 回复: 1 | ||
| 【奖励】 本帖被评价1次，作者conanwj增加金币 0.8 个 | ||
[资源]
【分享】Advances in Multi-Objective Nature Inspired Computing.Springer.2010
|
||
|
免责声明 本资源来自于互联网，仅供网络测试之用，请务必在下载后24小时内删除！所有资源不涉及任何商业用途。发帖人不承担由下载使用者引发的一切法律责任及连带责任！ 著作权归原作者或出版社所有。未经发贴人conanwj许可，严禁任何人以任何形式转贴本文，违者必究！ 如果本帖侵犯您的著作权，请与conanwj联系，收到通知后我们将立即删除此帖！ Authors(Editors): Carlos A. Coello Coello Clarisse Dhaenens Laetitia Jourdan (Eds.) Publisher: Springer Pub Date: 2010 Pages: 200 ISBN: ISBN 978-3-642-11217-1 e-ISBN 978-3-642-11218-8 DOI 10.1007/978-3-642-11218-8 Studies in Computational Intelligence ISSN 1860-949X Library of Congress Control Number: 2009941079 Preface Combinatorial optimization comprises a large class of problems having applications in a variety of domains. A combinatorial optimization problem may be defined by a finite set of discrete solutions D and an objective function f that associates to each solution a value (most of the time a real value) that represents its quality. Hence, a combinatorial optimization problem consists in optimizing (minimizing or maximizing) a given criterion under a set of constraints that allows to delimit the set of feasible solutions. The wide variety of problems in combinatorial optimization is due to its numerous applications. Indeed, combinatorial optimization problems may be found in productionmanagement, in telecommunications network design, in bio-informatics, in knowledge discovery, and in scheduling, among many other tasks. Solving a combinatorial optimization problem requires the study of three main points: ? The definition of the set of feasible solutions. ? The determination of the objective function to optimize. ? The choice of the optimization method. The two first points deal with the modelling of the problem, whereas the third one deals with its resolution. In order to determine the set of feasible solutions, it is necessary to express the set of constraints of the problem. This requires a very good knowledge of the problem under study and of its application domain. For example, linear programming may be used for this sake. The choice of the objective function also requires a good knowledge of the problem. The definition of the objective function should be done very carefully, because, it is useless to develop a very good optimization method if the objective function is not properly defined. Finally, the choice of the optimization method will often depend on the complexity of the problem. Indeed, according to its complexity, it may or may not be possible to solve the problem optimally. In case of problems of the classP, a polynomial algorithm has been found for it, and such algorithm can be used to solve the problem. In case of problems of the classN P, two ways are possible. If the size of the problem is small, an exact algorithm that allows us to find the optimal solution may be used (e.g., Branch& Bound or dynamic programming). Unfortunately, these algorithms are based on enumerative procedures and may not be used on large size problems (even if, in fact, the size is not the only limiting criterion). In this case, it is necessary to use heuristic methods in order to find good solutions in a reasonable time. Among these heuristic methods, metaheuristics offer generic resolution schemes that can potentially be adapted to any type of optimization problem. Hence the modelling phase of the problem is very important as it will, for example, allow to recognize a problem of the class P from anN P-hard problem. In particular, the definition of the objective function is crucial but may be difficult to realize, especially for real-world problems. Most real problems are multi-objective by nature, because several criteria have to be simultaneously considered. Combinatorial optimization problems are not an exception, and multi-objective instances of them have been studied during several years. The first studies of multi-objective optimization problems transformed them into a succession of single-objective optimization problems. This involved the use of approaches such as lexicographic ordering (which optimizes one objective at a time, considering first the most important, as defined by the user) and linear aggregating functions (which use a weighted sum of the objectives, in which the weights indicate the importance of each of them, as defined by the user). Over the years, other types of approaches were proposed, aiming to provide compromise solutions without the need of incorporating explicit preferences from the user. Today, many multi-objective metaheuristics incorporate mechanisms to select and store solutions that represent the best possible trade-offs among all the objectives considered, without any need to rank or to add all the objectives. The solution of a multi-objective optimization problem involves two phases: 1. Search for the best possible compromises: At this stage, any search algorithm can be adopted, and normally, no preference information is adopted. The aim is to produce as many compromise solutions as possible, and to have them as spread as possible, such that a wide range of possible trade-offs can be obtained. 2. Selection of a single solution: Once we have produced a number of compromise solutions, the decision maker has to select one for the task at hand. This phase involves a process called multi-criteria decision making, whose discussion is beyond the scope of this book. The purpose of this book is to collect contributions that deal with the use of nature inspired metaheuristics for solving multi-objective combinatorial optimization problems. Such a collection intends to provide an overview of the state-of-the-art developments in this field, with the aim of motivating more researchers in operations research, engineering, and computer science, to do research in this area. This volume consists of eight chapters including an introduction (Chapter 1) that provides some basic concepts of combinatorial optimization and multi-objective op timization that aim to facilitate the understanding of the rest of the book. This chapter provides a short discussion on algorithms, incorporation of user’s preferences, performance measures and performance assessment, and the use of statistical tools (including the use of public-domain software) to assess the quality of the results obtained by a multi-objective metaheuristic. The rest of the chapters were contributed by leading researchers in the field. Next, we provide a brief description of each of them. Horoba and Neumann present in Chapter 2 a study of diversity mechanisms that influence the approximation ability of multi-objective evolutionary algorithms. The role of each diversity mechanism in situations in which they become crucial is also exemplified aiming to gain a more in-depth understanding of their importance. Durillo, Nebro, Garc′?a-Nieto and Alba present in Chapter 3 a study of different mechanisms to update the velocity of a multi-objective particle swarm optimizer. Four velocity update mechanisms that aim to improve performance are analyzed. A comprehensive study adopting 21 test problems, five multi-objective particle swarm optimization variants and three performance indicators is undertaken by the authors to validate their hypothesis. The results indicate that the velocity update mechanism does indeed affect the performance of multi-objective particle swarm optimizers. Chapter 4, by C′amara, Ortega and de Toro, deals with dynamic multi-objective optimization problems. The authors analyze the importance of this area, analyze some of the test problems and performance measures previously proposed within this area, and introduce new proposals themselves. They also explore the potential of parallelism in this type of problems. Liefooghe, Jourdan, Legrand, Humeau and Talbi present in Chapter 5 a software framework that allows a flexible and easy design of metaheuristics for multiobjective optimization. A rich number of components already available in this software platform allows the immediate use of a variety of multi-objective metaheuristics as well as several performance measures and associated tools for the statistical validation of results. In Chapter 6, Lust and Teghem provide a study of the multi-objective traveling salesman problem, including a literature survey and a new method to solve it. The proposed approach combines the use of a special initialization procedure that generates an initial approximation of the compromise solutions and a local search procedure that improves such initial approximation. The proposed approach is found to be superior to other proposals previously reported in the specialized literature for biobjective instances. Paquete and St¨utzle present in Chapter 7 an empirical study of the performance of multi-objective local search approaches. Three components are analyzed: the initialization strategy, the neighborhood structure and the archive bounding technique adopted. The biobjective traveling salesman problem is adopted as a case study in this work. The main outcome of this study was the identification of certain patterns of algorithm behavior and the establishment of dependence relations between certain algorithmic components and instance features. Finally, Chapter 8, by Nolz, Doerner, Gutjahr and Hartl, introduces a hybrid approach based on genetic algorithms, variable neighborhood search and path relink ing, which is used to solve a multi-objective optimization problem that arises from a post-natural-disaster situation. This application is modeled as a covering tour problem and real-world data are adopted to validate the proposed approach. We hope that these chapters will constitute a valuable reference for those wishing to do research on the use of nature inspired metaheuristics for solving multiobjective combinatorial optimization problems, since that has been the main goal of this book. Finally, we wish to thank all the authors for their high-quality contributions and for their help during the peer-reviewing process.We also wish to thank Dr. Matthieu Basseur, Dr. Jean-Charles Boisson and Dr. Nicolas Jozefowiez for their kind support during the preparation of the book. Our sincere thanks to Prof. Janusz Kacprzyk for accepting to include this volume in the Studies in Computational Intelligence series from Springer. We also thank Dr. Thomas Ditzinger, from Springer-Verlag in Germany, who always provided prompt responses to all our queries during the preparation of this volume. Carlos A. Coello Coello thanks Gregorio Flores for his valuable help, to the financial support provided by CONACyT project 103570, to CINVESTAV-IPN for providing all the facilities to prepare the final version of this book, and to his family for their continuous support. Mexico City, Mexico Villeneuve d’Ascq, France Villeneuve d’Ascq, France October 2009 Carlos A. Coello Coello Clarisse Dhaenens Laetitia Jourdan Editors 本资源共4个可选网络硬盘链接，18.21 MB。 -------------------------------------------------------------------------------------------------------- https://www.easy-share.com/1910523092/Advances in Multi-Objective Nature Inspired Computing.Carlos A. Coello Coello eds.9783642112171.p200.Springer.2010.rar https://rapidshare.com/files/394 ... .9783642112171.p200 https://www.divshare.com/download/11576241-617 https://www.sendspace.com/file/7ib0np -------------------------------------------------------------------------------------------------------- [ Last edited by conanwj on 2010-10-10 at 20:57 ] |
回复此楼» 猜你喜欢
一志愿东华大学化学070300，求调剂
已经有8人回复
-
277材料科学与工程080500求调剂
已经有7人回复
-
317求调剂
已经有18人回复
-
293求调剂
已经有5人回复
-
280分求调剂 一志愿085802
已经有7人回复
-
0854电子信息求调剂
已经有3人回复
-
263求调剂
已经有4人回复
-
石河子大学（211、双一流）硕博研究生长期招生公告
已经有3人回复
-
一志愿西安交通大学材料工程专业 282分求调剂
已经有13人回复
-
北科281学硕材料求调剂
已经有6人回复
» 本主题相关价值贴推荐，对您同样有帮助:
- Nature杂志投稿指南
已经有23人回复
- 《Recent Advances in Nanofabrication Techniques and Applications》2011
已经有178人回复
- 《Advances in Nanocomposite Technology》2011
已经有66人回复
- 《Nature》网站推荐南京大学张辰宇教授重大发现——第七种营养成分
已经有31人回复
- 2011年新书《Advances in Ceramics 》
已经有131人回复
- 有人投过ADVANCES IN ENGINEERING SOFTWARE这个期刊没？
已经有6人回复
- ISIF Journal of Advances in Information Fusion是否被SCI检索？
已经有4人回复
- 【求助】In-Situ Chemical Oxidation指的是什么？
已经有8人回复
- 巨慢 Advances in Polymer Technology
已经有7人回复
简单回复
2019-09-02 09:15
回复
五星好评 
