Decision Making in Complex Systems (eBook)

The DeciMaS Agent-based Interdisciplinary Framework Approach

Marina V. Sokolova, Antonio Fernández Caballero (Autoren)

eBook Download: PDF

2012
184 Seiten
Springer Berlin (Verlag)
978-3-642-25544-1 (ISBN)

Infos

The study of complex systems attracts the attention of many researchers in diverse fields. Complex systems are characterized by a high number of entities and a high degree of interactions. One of the most important features is that they do not involve a central organizing authority, but the various elements that make up the systems are self-organized. Moreover, some complex systems possess an emergency priority: climate change and sustainable development research, studies of public health, ecosystem habitats, epidemiology, and medicine, among others.

Unfortunately, a great number of today’s overlapping approaches fail to meet the needs of decision makers when managing complex domains. Indeed, the design of complex systems often requires the integration of a number of artificial intelligence tools and techniques. The problem can be viewed in terms of goals, states, and actions, choosing the best action to move the system toward its desired state or behavior. This is why agent-based approaches are used to model complex systems.

The main objective of this book is to bring together existing methods for decision support systems creation within a coherent agent-based framework and to provide an interdisciplinary and flexible methodology for modeling complex and systemic domains.

 

Inhalt

Title 2
Preface 7
Contents 10
Decision Making in Complex Systems 15
Introduction 15
Modeling of Natural and Complex Phenomena 16
Complex Systems: Definition and Principal Characteristics 16
Decision Making and Decision Support Systems 19
Decision Making 19
The Evolution of Decision Support Systems 20
Decision Making in Complex Systems 23
Composite Decisions 25
Motivation 27
Frameworks for the Design of Decision Support Systems 27
The Need for an Interdisciplinary Approach 28
Multi-agent Paradigm for Complex Systems Modeling 29
A Review on Frameworks for Decision SupportSystems 33
Introduction 33
Decision Support Systems in Academy and Research 34
Agent-Based Frameworks for Decision Support Systems 42
Frameworks for Multi-agent Systems Planning 42
Software Tools for Multi-agent Systems Design and Implementation 50
Comparison of Agent-Based Frameworks 57
Design and Implementation of the DeciMaSFramework 60
Introduction 60
The DeciMaS Framework 61
Approach towards Ontology Creation 61
The General Structure of the System 63
Description of the Ontological Basis of the Multi-agent Architecture 67
The Domain of Interest Ontology 68
The MAS Architecture Ontology 68
The Tasks Ontology 70
The Agent Ontology 70
The Interactions Ontology 71
The Distributed Meta-Ontology 72
Data Mining Methods in the DeciMaS Framework 73
Methods for Information Fusion and Preprocessing 76
Methods for Knowledge Discovery 81
Methods for Decision Generation 99
A Case Study for the DeciMaS Framework 102
Introduction 102
Human Health Environmental Impact Assessment 103
Environment and Human Health 103
Environmental Impact Assessment 105
Design of the Agent-Based Decision Support System 105
Meta-ontology of the System 105
Logical Levels of the ADSS 111
The Principal Abstractions of the System 113
Implementation in Jack 134
Program Architecture 134
Defining Agents in JACK 135
Data and Results 152
Introduction 152
Data for Experiment 153
Data Retrieval and Fusion 155
Information Fusion and Preprocessing 157
Detection and Elimination of Artifacts 157
Filling of Missing Values 158
Smoothing Results 159
Normalization Results 160
Results of the Correlation Analysis 160
Decomposition Results 161
Knowledge Discovery Results 163
Regression Models 163
Neural Network Models 165
Models Obtained with the Group Method of Data Handling 168
Committee Machines 170
Environmental Impact Assessment Results 172
Decision Making 174
Discussion of the Experiment 178
Conclusions 181
Conclusions 181
Future Work 183
References 185

Produkt-Details

• Erscheint lt. Verlag: 13.1.2012
• Reihe/Serie: Intelligent Systems Reference Library
• Reihe/Serie: Intelligent Systems Reference Library
• Zusatzinfo: XVI, 184 p.
• Verlagsort: Berlin
• Sprache: englisch
• Themenwelt: Informatik ► Theorie / Studium ► Künstliche Intelligenz / Robotik
• Themenwelt: Technik ► Bauwesen
• Schlagworte: agent-based software engineering • Complexity • Complex Systems • Decision Making • Intelligent Systems
• ISBN-10: 3-642-25544-2 / 3642255442
• ISBN-13: 978-3-642-25544-1 / 9783642255441