검색 상세

A Methodology to Develop Dependable Component Models for Modeling and Simulation of Critical Infrastructures

초록/요약

Critical Infrastructures (CIs) in a highly civilized society constantly provide social constituents with essential services that are required to keep everyday life. As a complex system, which consists of diverse subsystems (or components), CIs have an emergent behavior that is synthetically caused by the behaviors of their subsystems. Also, CIs are vulnerable to several threats, such as natural disasters, terrors, or accidents, as well as having dependence on other CIs. These characteristics of CIs make it impossible for us to experiment with actual CIs in understanding their emergent behaviors because the experiment requires the cost that we cannot economically and socially allow. Under this situation, Modeling and Simulation (M&S) is a useful method for understanding the emergent behaviors of CIs. M&S frees us from several constraints such as time, cost, space, risk, policy, etc. There have been many successful researches that use M&S techniques in order to understand the emergent behaviors of CIs. Despite their successes, many practitioners still do not regard M&S as an engineering discipline but as a form of art. The reason is that the quality of models, which are developed in compliance with the previous modeling methods, excessively depends on the M&S practitioners’ capabilities and experience. In fact, CIs and their services are intrinsically designed to be dependable. However, the previous modeling methods are indifferent to the dependability of CIs or do not provide a way to explicitly and systematically apply this property into modeling CIs. We believe that the dependability is one of the most important attributes for the practitioners to produce the quality models of CIs. As a cornerstone in the process to transfer M&S of CIs from an art to an engineering discipline, we propose a CI M&S methodology, called Dependability-enhanced Unified Modeling and Simulation (DUMAS) methodology, in order to produce quality models. Especially, the proposed methodology intuitively, explicitly and systematically deals with the dependability existing in CIs. Firstly, we propose a system to model & simulate CIs based on Discrete Event System Specification (DEVS) formalism and Belief-Desire-Intention (BDI) and Role-Oriented Command Hierarchy (ROCH) models. Next, we attempt to integrate several previous modeling methods, such as goal modeling, agent-based modeling, object-oriented modeling, and feature modeling, into Unified Modeling and Simulation (UMAS) methodology. Then, we propose DUMAS methodology to explicitly and systematically treat dependability upon UMAS methodology. In order to validate the feasibility of the development process of DUMAS methodology, we conduct a case study to apply DUMAS methodologies into the M&S of a smart grid, i.e. a kind of CI. Then we discuss how the development process of DUMAS methodology affects modeling artifacts of the smart grid. From these evidences observed in the case study, we conclude that the dependability of CIs is a key attribute in developing quality models.

more

목차

Chapter I. Introduction 1
Chapter II. Related Work 5
A. The Characteristics of Critical Infrastructures 5
B. Modeling and Simulation of Critical Infrastructures 7
C. Methodologies and Models for Software Engineering 9
Chapter III. A System to Model and Simulate Critical Infrastructures 12
A. Theoretical Models and Architecture of CI M&S System 14
B. DEVS and its implementation 15
C. BDI and ROCH models and their implementation 19
Chapter IV. Unified Modeling and Simulation Methodology 30
A. Application Domain 31
B. Defining a Simulation Purpose 33
C. Designing Scenario 35
1. Constructing CIs 36
i. Reusing CIs 37
ii. Deploying Components 37
2. Planning Events 37
D. Simulating CIs 38
E. Developing a Component Model 39
1. Early Requirements Analysis 40
2. Late Requirements Analysis 43
3. Architecture Design 45
4. Detail Design 50
5. Implementation 52
F. Testing Smart Grid, VPP and WindTurbine 60
Chapter V. Dependability-enhanced Unified Modeling and Simulation Methodology 65
A. Singularity: Where Dependability should be Considered 67
B. Derivation of Dependability Requirements 68
C. Development Process of Dependable Component Model 74
Chapter VI. Case Study 76
A. Review of UMAS and DUMAS methodologies 76
B. Design of Case Study 79
C. Conduct of Case Study 84
Chapter VII. Discussion 92
Chapter VIII. Conclusion 94
References 97
Appendix A. SD Model for Component Models in UMAS Methodology 103
Appendix B. SR Models for Component Models in UMAS methodology 104
Appendix C. SD Model for Component Models in DUMAS Methodology 108
Appendix D. SR Models for Component Models in DUMAS methodology 109

more
반출 Meta View 목록